Lng Accidents and Risks

RESOURCE FILES

Managing LNG Risks From the campaign to prevent LNG terminals in Passamaquoddy Bay Documents and references used to explain LNG

Compiled by Arthur MacKay Bocabec, NB, Canada September, 2012

RESOURCE FILES

WHAT ARE RESOURCEFILES? Resource files are created from the contents of the working reference and publication files of Art MacKay and are made available for reference purposes. They contain documents, drawings, photographs and other resources accumulated over a 50 year period, including public domain materials as well as materials with copyrights held by Arthur MacKay and others. Since online resources come and go, they have been converted to PDFs and archived to preserve their contents. They can be accessed directly where the links are still active. Live links and copyright requirements are specified for each item if still available. Art MacKay can be contacted at art@bayof fundy.ca to clarify availability for further publication. Entire files composed of physical documents, books, photos, cds, etc. are available and sold separately.

St. Andrews Citizens Will be Burning!

© Art MacKay, 2009, St. Andrews, New Brunswick, Canada, January 10, 2008 The US Coast Guard’s recently released Waterway Suitability Report, prepared for the Federal Energy Regulatory Commission’s (FERC) Downeast LNG application, will have the residents of New Brunswick’s premier resort area burning again.

When, on the heels of Quoddy Bay LLC’s earlier proposal at Sipyik, Downeast LNG first introduced their plan to construct an LNG terminal and storage facility directly across the St. Croix River from St. Andrews at Robbinston, Maine, the citizens of that resort town and the greater Quoddy Region rose as one to denounce these developments and made their position clear with submissions to FERC and government officials in both Canada and the United States, Impressive, passionate, packed public meetings and protests were held. Subsequently, Canadian opponents to the LNG plans have been supported at every political level in Canada. Conservative Prime

Minister Harper and local Member of Parliament Hon. Greg Thompson have publicly and in closed session with President George Bush, expressed their firm position that tankers will not be allowed through Head Harbour Passage, the essential waterway that they consider to be internal Canadian waters and, in the long term, too valuable and hazardous a waterway to be used by supertankers. This was a similar position to that taken about 30 years ago when the Pittston Company of Greenwich, Conn. applied to turn Eastport and Moose Island into a gigantic oil refinery and tank farm, threatening fishing, tourism, marine life, and whales including the endangered north Atlantic right whale. In previous responses to Quoddy Bay LLC, the US Coast Guard stated that the participation of the Canadian government was paramount to their release of a similar report for Quoddy Bay LLC and it was withheld. This is not the case with the Downeast LNG proposal. In this case, while the report requires Canadian consultation by Downeast LNG, the USCC has chosen to assess Canadian waters without the approval of Canada, an interesting move that causes great concern in a contest that, more and more, seems to revolve around testing Canadian sovereignty and, not incidentally, Canadian resolve. As can be seen in the accompanying chart taken from the USCC Waterway Suitability Study, St. Andrews citizens can now see that they are within the hazard zone for this development. A similar study for Calais LNG will, undoubtedly, duplicate this scenario, but will move Zone 3 more deeply into the town as tankers move up the St. Croix River. The USCC defines these zones as: • Zone 1 (red) - 500 meter radius with resultant fire and severe thermal radiation hazards. By definition these are areas in which LNG shipments occur in relatively narrow harbors or channels, or ships pass under major bridges or over tunnels, or come in within 500 meters of major infrastructure such as military installations, commercial/business centers, or national icons. • Zone 2 (yellow with black line) - from 500 to 1600 meters with less severe thermal radiation hazards to public safety and property. These are areas of broader channel widths, larger open harbors, or over 500 meters from major critical infrastructure elements. • Zone 3 (yellow) – from 1600 to 3500 meters with potential pockets of flammable vapor. These are areas where LNG traffic and deliveries occur approximately 1.6 kilometers from major infrastructure or in large bays or open water. The thermal radiation risks to public safety and property are significantly reduced. While thermal risks may be reduced with distance, MIT Professor Emeritus James Fay , points out that the actual zone of impact, the area were fires are ignited and people suffer serious burns, may be greater than the distance used in the report. He states that, “for all credible spills, including terrorist attacks on the storage tank and LNG tanker, the danger zone for humans extends almost 4 miles from the terminal site” or about 2.5 kilometers, and life and property will be lost from so-called collatera impacts. This greater distance envelopes all of St. Andrews and the ability of fire departments may be non-existent since their facilities are within the real zone of impact and these professionals may well be immobilized by an event itself. the nearest assistance would be St. Stephen and St. George. As for bringing tankers through Head Harbour Passage, this is a red herring. Of course it can be done. It’s risk analysis and the real question is for how long will it take to have an accident and at what cost? Since LNG tankers can only enter and leave during the day, at slack tide (if that truly exists in some areas along the route), when the visibility is more than 2 miles, and the wind is less than 25 mph, then the number of days when access is available will be severely limited. In fact, these data are available and it is a wonder that they have not been required for the USCC report and company submissions to FERC. Imagine the financial impacts to Downeast LNG and its leader Dean Girdis during those delightful years like the one when fog held to the West Isles for more than 30 days and 30 nights. At $100,000 a day, layovers adds up! Passamaquoddy Bay and key fishing areas of Campobello Island

could well see numerous gigantic tankers stacked up waiting to move. Foget the hazardous passage and the old sow whirlpool, layovers will be substantial and local boats will be unable to pursue fishing, whale watching, and recreational activities during passage and while at anchor or at the terminal. The eco-economy of Quoddy will be effectively shut down. Since the arrival of tankers is “secret” due to fears of terrorism, local operators will have little time to respond and will be forced to the side by armed gunboats, as the are in Boston Harbour. What a delightful vision. Unless the new administration in the United States recognizes the folly of agitating their neighbours, best friends, and largest trading partner, the first LNG explosion may be coming soon, much sooner, than anticipated by the LNG promoters. A once interesting and functional “international community” that drews upon the abundant natural resources of this unique Quoddy ecosystem, has been split asunder by these LNG development proposals. In spite of the plethora of carpetbaggers who have wandered through Charlotte and Washington Counties, some folks have never stopped looking for the knight on a white steed; the saviour who will bring economic salvation to an area of perceived poverty. The sad truth is that the wealth required to provide a truly sustainable future for all of the citizens of Quoddy has always been here for those with eyes to see. They will forget their old laws; they will barter their country for baubels. Then will disease eat the life from their blood. (Hanisse’ono. The Evil One from Iroquois legend.) Art MacKay is a biologist, writer, and artists with over 40 years professional experience in the Bay of Fundy and northern Gulf of Maine. He is the author of many reports and articles about these ecosystems. ******************************************************************************************

Are we safe around this industry?

”Algeria

Blast Has Officials

Rethinking LNG Safety” Jan 19th, 2004,

30 dead - 74 injured

(LNG liquifaction trains)

Duke Energy Moss Bluff Texas gas storage site, Aug.20th, 2004 AP photo 3 mile radius evacuated, could not extinguish blaze for days

Belgian Gas Explosion Kills 14 Workers Had Reported Piercing Underground Pipeline

Associated Press Saturday, July 31, 2004

18 dead 200 injured

“Society has a way of ignoring the Low Probability-High Consequence accident. This is exactly what terrorists look for.” Richard Wilson, Harvard, giving a talk at the high security Los Alamos Labs in New Mexico after 9/11

How Does FERC Define Worst Case Accident Scenario? “A break in one LNG transfer arm that lasts for 10 minutes. Total spill is ~550,000 gallons. This represents less than 1.5% of the capacity of one LNG storage tank. “ Bill Powers, Border Power Plant Working Group, Calif.

James A Fay Professor Emeritus, Senior Lecturer, MIT • Acknowledged world wide expert on fluid mechanics and spread of hazardous materials on the ocean • Education, honors, textbooks too lengthy and numerous to list here. See: www.JamesFay.com

James Fay’s argument in a nutshell: The FERC regulations do not prevent people living here from getting 2nd degree burns in 30 seconds if a bad accident occurs. Fay’s limits of exposure by distance ( see circles on next slide) DO prevent these people from getting burned if a bad accident happens. If Fay’s limits were observed, the LNG tankers could not be allowed in here because of safety reasons.

Safe Distances from Terminal: Small circle for spill with fire from one hold of tanker Large red circle for spill from storage tank Large blue circle for safe distance from tanker spill for flammable vapor James Fay report

Safe Distance from tanker path for heat radiation from spill with fire James Fay

Green line is the tanker path

• The potential for retarding a pool fire is nonexistent • A gasoline fire is only 8% of the intensity of an LNG fire

James Fay “Its too hot to get near and you can’t put it out until it runs out of fuel”.

ABS Consulting Report for FERC- 2004 For unconfined spills on water: 1.It is impossible to predict with certainty what would happen 2.A pool fire is possible that would burn people and property a mile away 3.This could cause 2nd degree burns in 30 seconds and 3rd degree burns in 50 seconds 4.A flammable vapor cloud could travel several thousand feet 5.Until we understand more we should use the high end of the scale of prediction www.soundenergysolutions.com

The U.S. and FERC with respect to LNG -Reckless behavior -No public conscience -No vision -No planning -No responsibility -Unacceptable risk

The U.S. and FERC with respect to LNG -Reckless behavior -No public conscience -No vision -No planning -No responsibility -Unacceptable risk

Tim Riley, California lawyer who opposes LNG says it very well: A sound, safe energy policy requires solutions that make America stronger not weaker, make America more self sufficient not more dependent, make America safer not more vulnerable, make America fossil free not more polluted. Prepared by Joyce Morrell Campobello, Oct.2004

Public Safety Issues at the Proposed Pleasant Point LNG Terminal James A. Fay 77 Massachusetts Avenue, Rm. 3-258 Cambridge, MA 02139

August 5, 2004

1 Introduction Quoddy Bay L.L.C.1 has proposed to construct and operate a liquefied natural gas (LNG) import terminal on the Sipayik tribal land at Pleasant Point, near Eastport, ME. To reach this terminal, ocean-going LNG tankers must move through Canadian waters between Campobello and Deer islands (Canada) as well as U.S. and Canadian waters between Eastport and Deer Island. A tanker spill at any location along this route would have serious consequences for persons and property on the shore adjacent to the stricken vessel, whether that be on Campobello or Deer I. or Eastport and the Sipayik Reservation. Natural gas, a hydrocarbon fuel, is usually piped directly from a gas well to the end consumer, never being stored locally in large amounts. When cooled to liquid form, however, as much as 50,000 tons can be stored in insulated tanks on land or aboard ship. In this form it is especially hazardous if it escapes by accident from its container, spilling onto ground or water and turning very rapidly into gaseous form, whereupon it will mix with air and then burn if ignited. By its very nature, an LNG import terminal is a hazardous industrial facility which could experience accidental fires that might harm surrounding populations and property. To build and operate an LNG terminal at the Pleasant Point site, Quoddy Bay must obtain permission from the Federal Energy Regulatory Commission (FERC)2 , an independent agency that regulates interstate commerce in natural gas and electricity. Although primarily an economic regulator, FERC has asserted jurisdiction over the safety aspects of the LNG facilities it permits. FERC requires facility owners to meet certain technical standards in site selection and equipment design and operation before it awards the right to import LNG and to connect the facility to an interstate natural gas transmission line. FERC’s jurisdiction does not extend to safety aspects of marine tankers; they are regulated by the U.S. Coast Guard.3 FERC’s objective in safety regulation is to limit, but not necessarily prevent, harm to persons and property outside the confines of the terminal site, should there be an accidental release of LNG at the site. The principal harmful effects are two: vapor plumes or clouds that can be ignited outside the site boundaries and harmful thermal radiation from on-site fires that extends across the site borders. But FERC’s safety rules do not consider all credible spills on the site or any from the LNG tankers while in transit to the terminal or being unloaded. This report explains the safety requirements that will likely be applied by federal regulators to the proposed LNG terminal in Pleasant Point. It delineates the geographic extent of harmful effects that could be expected from LNG spills at the site, including those that are excluded from FERC and U.S. Coast Guard safety regulations.

2 FERC site selection criteria FERC rules4 require the LNG terminal owner to install extensive technological features that will limit the harmful consequences of an accidental spill of LNG to within the property line enclosing the terminal. The harmful effects are twofold: combustible mixtures of vapor and air, such as might be driven by the wind blowing over an evaporating pool of spilled LNG, and thermal radiation from a fire burning above a liquid spill on the site. The types of spills to be considered are also twofold: a 1 Quoddy Bay L.L.C. is owned by Smith Cogeneration (www.smithcogeneration.com). 2 Federal Energy Regulatory Commission (www.ferc.gov). 3 The safety of the natural gas pipeline connecting the terminal to the interstate transmission line is regulated by the Office of Pipeline Safety of the U.S. Department of Transportation, but the FERC permit for the LNG terminal confers on the terminal owner the right to seek seizure of private land to construct the connecting pipeline, if necessary. 4 Code of Federal Regulations, 49 CFR 193.

1

Figure 1: The primary and secondary containment tanks for a ”full containment” storage tank of the type to be used at the proposed Harpswell LNG terminal.

spill from transfer piping connecting the storage tanks and the regasification or unloading facilities, and the failure of the primary storage tank enclosure. Limiting these effects at a terminal requires the construction of impounding areas surrounding potential spill sources so as to collect the spilled liquid and slow its vaporization or burning rate. If the spills are sufficiently small, harmful effects will not extend beyond the site line. For transfer line spills, the LNG is collected in a central impounding area. For storage tank spills, the inner storage container is surrounded by a secondary containment tank of slightly larger size, as shown in Figure 2, which can contain all the LNG that might spill from the inner primary container. The potential for harmful effects to humans from a given spill decreases with distance from the spill site. The harmful effect of ignitable natural gas vapor is measured by the flammability distance, a distance down wind from the spill site at which the vapor has been so diluted by mixing with air that it cannot be ignited. Any ignition at a closer distance can propagate a flame, but that flame will not propagate beyond the flammability distance. If the latter distance lies within the site boundary, no flame can extend beyond that boundary. Thermal radiation from on-site LNG fires fed by an evaporating pool of spilled LNG can cause first, second or third degree burns to the skin of humans exposed to the radiation, depending upon the intensity of radiation. For a given fire, this intensity decreases with distance from the fire. The least intense thermal radiation that FERC rules allow humans outside the site boundary to be exposed to is 5 kilowatts per square meter, an amount that produces second degree burns after only thirty seconds exposure.5 The FERC requirements for the proposed Quoddy Bay terminal can be estimated from the Final Environmental Impact Statement for the Hackberry LNG project in Louisiana.6 This project, consisting of three storage tanks and two unloading piers, employs the technology likely to be used at the Pleasant Point facility. Values from this report of the flammability and thermal radiation dis5 More intense and thereby more damaging exposure is permitted depending upon land use characteristics at the site

boundary. 6 Final Environmental Impact Statement, Hackberry LNG Project, Cameron LNG, LLC. FERC/EIS-0156. Office of Energy Projects, Federal Energy Regulatory Commission, Washington, DC 20426. August 2003.

2

Table 1: Flammability and radiation distances for FERC-defined spills

Spill source Transfer piping Storage tank (primary)

Size (ton)

Flammability (ft)

5 kW/m2 Radiation (ft)

840

770

320

74,000

929

tances for a transfer line spill, and the thermal radiation distance for a primary containment spill, are listed in Table 1, together with the amounts of the respective spill volumes. It would appear that for these FERC-defined spills neither radiation nor flammability will exceed the FERC limits beyond the site boundary.

3 Risks that FERC ignores There are several important public safety risks that are not considered in the FERC regulations discussed above. 1. First of all, FERC allows damaging thermal radiation beyond the site boundary as long as its level is below 5 kilowatts per square meter. However, it is not until the thermal radiation intensity falls below 1.6 kilowatts per square meter that there is no damage to exposed humans. A safe radiation distance for fires would be that for which the thermal radiation level does not exceed 1.6 kilowatts per square meter. Distances at which the radiation exceeds this value would lie within a thermal radiation danger zone. 2. Secondly, FERC’s regulations ignore the greatest risks of all, that foreign or domestic terrorists could destroy the storage tank primary and secondary containment systems, or the LNG tanker cargo hold, allowing LNG to spill unhindered onto ground or water, where it would most likely burn. Because the lateral extent of such spills would be so much greater than those considered in the FERC regulations, it is to be expected that their harmful effects would exist very far beyond the site boundaries. To show how public safety can be adversely affected by credible spills that have been overlooked by FERC, we have extended Table 1 to include the effects listed above.7 This expanded assessment is listed in Table 2. Two additional spills are considered, those from the secondary storage tank containment system and a single hold of a marine tanker (last two rows of Table 2). For these and the previous spills of Table 1, the safe radiation distance defining the outer boundary of the thermal radiation danger zone, mentioned in item 1 above, has been calculated for all spills (last column of Table 2). Also, the flammability distance for the FERC primary containment failure accident is shown in the flammability column. 7 The methods used for this assessment are identical to those contained in ”Consequence assessment methods for incidents involving releases from liquefied natural gas carriers”, Report 131-04 GEMS 1288209, ABS Consulting, Inc., May 13, 2004, (available on FERC web site at www.ferc.gov/industries/gas/indus-act.asp) and its Attachment 1 of June 29, 2004, as listed on the FERC site at http://ferris.ferc.gov/idmws/search/fercgensearch.asp under docket AD04-6.

3

Table 2: Flammability and radiation distances for all credible spills

Spill source

Transfer piping

3.1

Size (ton)

Flammability (ft) Danger zone

1.6 kW/m2 Radiation (ft) Danger zone

840

770

1,230

Storage tank (primary)

74,000

1,490

Storage tank (secondary)

74,000

19,685

Tanker hold

5,250

19,360

7,870

Thermal danger zones

The thermal radiation danger zones for the largest credible spills listed in Table 2 are shown in Figure 2. All of these extend beyond the site boundaries, especially so for the tanker and secondary tank spill with fire. But even the FERC spills with fire from transfer piping and primary containment send damaging radiation beyond the site boundaries. Altogether, about 20 square miles of U.S. shore land in the Pleasant Point area and 3 square miles on Deer Island are at risk for damage to humans from on-site spills at the proposed LNG terminal.

3.2

Tanker danger zones

Spills from a fully loaded LNG tanker can occur not only at the unloading dock, as shown in Figure 2, but also at any point along the ship channel while approaching the terminal. Figure 3 shows the proposed path to be followed by an LNG tanker heading for the terminal. Thermal radiation danger zones for spills at four locations along the path are shown. At any location, about 2 square miles on the U.S. shoreline and an equal amount on the Canada shoreline (Campobello and Deer Islands) lie within the thermal danger zone.

3.3

Flammable vapor danger zones

The blue circle in Figure 2 depicts the flammability danger zone for a spill, without fire, from the tanker while located at the terminal pier. For any such spill, the flammable vapor plume or cloud would extend from the tanker in the downwind direction, encompassing an area of about a square mile. Winds from the northwest, and clockwise to the southeast, would send the vapor plume to U.S. land area from Eastport to the Passamaquoddy shoreline, while winds from the southwest, and clockwise to the northwest, would send the vapor over land areas of Deer Island. The spills described in Tables 1 and 2 do not include spills without fire from the secondary containment of the land storage tank. Because such a spill would be more than ten times the tanker spill in volume, the corresponding flammability distance would be considerably greater than the blue circle shown in Figure 2.

4

X

X

Figure 2: The thermal radiation and flammable vapor danger zones for spills listed in Table 2. Red circles are distances to radiation intensities of 1.6 kW/m2 for a spill with fire; larger for loss of secondary containment of land storage tank, smaller for spill from one hold of LNG tanker. Blue circle is flammable vapor distance for a tanker spill.

5

Figure 3: The path of a tanker approaching the proposed LNG terminal (green dashed line) and the radiation danger zones for a spill at four locations along this path.

6

4 Conclusions 1. The federal safety requirements for the proposed Pleasant Point LNG terminal will not prevent harm to humans outside the site boundary for the spill scenarios that FERC considers. 2. For all credible spills, including terrorist attacks on the storage tank and LNG tanker, the danger zone for humans extends almost 4 miles from the terminal site, encompassing 20 square miles of land in the Pleasant Point area. 3. For a tanker spill anywhere along the route leading to the LNG terminal, the thermal radiation danger zone for humans extends 1.5 miles from the tanker route, encompassing up to 4 square miles of land along U.S. and Canada shores in Eastport, Campobello Island and Deer Island, depending upon the spill location along the tanker track.

7

REFERENCE ARCHIVE

SANDIA REPORT SAND2008-3153 Unlimited Release Printed May 2008

Breach and Safety Analysis of Spills Over Water from Large Liquefied Natural Gas Carriers Anay Luketa, M. Michael Hightower, Steve Attaway

Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185 and Livermore, California 94550 Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under Contract DE-AC04-94AL85000. Approved for public release; further dissemination unlimited.

Issued by Sandia National Laboratories, operated for the United States Department of Energy by Sandia Corporation. NOTICE: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government, nor any agency thereof, nor any of their employees, nor any of their contractors, subcontractors, or their employees, make any warranty, express or implied, or assume any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represent that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government, any agency thereof, or any of their contractors or subcontractors. The views and opinions expressed herein do not necessarily state or reflect those of the United States Government, any agency thereof, or any of their contractors. Printed in the United States of America. This report has been reproduced directly from the best available copy. Available to DOE and DOE contractors from U.S. Department of Energy Office of Scientific and Technical Information P.O. Box 62 Oak Ridge, TN 37831 Telephone: Facsimile: E-Mail: Online ordering:

(865)576-8401 (865)576-5728 [email protected] http://www.osti.gov/bridge

Available to the public from U.S. Department of Commerce National Technical Information Service 5285 Port Royal Rd Springfield, VA 22161 Telephone: Facsimile: E-Mail: Online order:

(800)553-6847 (703)605-6900 [email protected] http://www.ntis.gov/help/ordermethods.asp?loc=7-4-0#online

SAND2008-3153 Unlimited Release Printed May 2008

Breach and Safety Analysis of Spills Over Water from Large Liquefied Natural Gas Carriers Anay Luketa Fire and Aerosol Sciences Department Mike Hightower Energy Systems Analysis Department Steve Attaway Mechanical Environments Department Sandia National Laboratories P.O. Box 5800 Albuquerque, New Mexico 87185 Abstract In 2004, at the request of the Department of Energy, Sandia National Laboratories (Sandia) prepared a report, “Guidance on the Risk and Safety Analysis of Large Liquefied Natural Gas (LNG) Spills Over Water”. That report provided a framework for assessing hazards and identifying approaches to minimize the consequences to people and property from an LNG spill over water. The report also presented the general scale of possible hazards from a spill from 125,000 m3 to 150,000 m3 class LNG carriers, at the time the most common LNG carrier capacity. Because of the increasing size and capacity of many new LNG carriers, the Department of Energy requested that Sandia assess the general scale of possible hazards for a breach and spill from newer LNG carriers with capacities up to 265,000 m3. Building on the research and analyses presented in the 2004 report, Sandia reassessed emerging accidental and intentional threats and then conducted detailed breach analyses for the new large LNG carrier designs. Based on the estimated breach sizes, breach locations, and LNG carrier configurations, we estimated LNG spill rates and volumes and conducted thermal hazard and vapor dispersion analyses. This report summarizes the different analyses conducted, the expected range of potential hazards from a large LNG carrier spill over water, and risk management approaches to minimize consequences to people and property from such a spill.

3

ACKNOWLEDGEMENTS The authors received technical, programmatic, and editorial support on this project from a number of individuals and organizations both inside and outside Sandia National Laboratories. We would particularly like to express our thanks for their support and guidance in the technical evaluations and development of this report. The authors would also like to thank Marcus Epperson and Benjamin Taylor at Sandia for providing exceptional support in ensuring successful computations on the Razor cluster. The U.S. Department of Energy was instrumental in providing coordination, management, and technical direction. Special thanks go to DOE personnel in the Office of Oil and Natural Gas, Office of Fossil Energy, for their help in supporting the modeling, analysis, and technical evaluations. To support the technical analysis required for this project, the authors worked with many organizations, including maritime agencies, LNG industry, and government agencies to collect background information on ship and LNG cargo tank designs and accident and threat scenarios needed to assess emerging large LNG carrier breach and spill safety hazard implications. The following individuals were especially helpful in supporting our efforts by providing information and data, coordinating industry and governmental agency interactions, and reviewing technical evaluations. Robert Corbin – Department of Energy John Cushing – US Coast Guard Ray Martin – US Coast Guard Ken Smith – US Coast Guard Pavagada Vasanth – US Coast Guard Terry Turpin – Federal Energy Regulatory Commission Chris Zerby – Federal Energy Regulatory Commission David Weimer – Det Norske Veritas Patricia Outtrim – Cheniere LNG

4

CONTENTS LIST OF ACRONYMS ...................................................................................................................6 1. EXECUTIVE SUMMARY ........................................................................................................7 2. BACKGROUND ........................................................................................................................9 2.1 Overview of LNG Carriers ..............................................................................................10 2.2 Size and Capacity of Emerging LNG Carrier Designs ....................................................11 3. THREAT AND BREACH ANALYSES ..................................................................................13 3.1 Analysis of Intentional Threat Scenarios for Large LNG Carriers..................................13 3.2 Analysis of Intentional Breaching of Large LNG Carriers..............................................13 4. LNG SPILL HAZARD EVALUATIONS AND RISK REDUCTION....................................16 4.1 Pool Fire Hazard Analyses...............................................................................................16 4.2 Vapor Dispersion Analyses..............................................................................................20 4.3 Hazard and Risk Reduction Considerations ....................................................................22 5. CONCLUSIONS.......................................................................................................................23 APPENDIX....................................................................................................................................25 A1. Surface Emissive Power..........................................................................................................25 A2. Fuel Volatilization Rate ..........................................................................................................25 A3. Flame Height...........................................................................................................................27 A4. Flame Tilt and Drag ................................................................................................................29 A5. Atmospheric Attenuation ........................................................................................................29

LIST OF FIGURES Figure 1. Typical Containment Systems for LNG Carriers ..........................................................10 Figure 2. Example of a 205,000 m3 Membrane Regasification LNG Carrier ..............................12 Figure 3: Example of Large Capacity LNG Carrier Structural Model .........................................14

LIST OF TABLES Table 1. Table 2. Table 3. Table 4. Table 5.

Emerging LNG Carrier Size and Capacity .....................................................................11 Thermal hazard distances from a pool fire for near-shore operations ............................19 Thermal hazard distances from a pool fire for offshore operations................................20 Distance to the LFL for vapor dispersion for near-shore operations..............................21 Distance to the LFL for vapor dispersion for offshore operations .................................21

5

LIST OF ACRONYMS CFD

Computational fluid dynamics

CTH

Hydrodynamics Code 3-D

DOE

Department of Energy

DWP

Deep Water Ports

GAO

Government Accountability Office

FERC

Federal Energy Regulatory Commission

kW/m2

kilowatts per square meter

km

kilometer – 1000 meter

Knts

Knots – 0.514 m/s

L/D

Height to diameter ratio

LFL

Lower flammability limit

LNG

Liquefied natural gas

LNGC

LNG Carrier

m

meters

m2

square meter (area)

m3

cubic meter (volume)

m/s

meters per second

psi

pounds per square inch

RLNGC

Regasification LNG Carrier

SRV

Storage and Regasification Vessel

USCG

United States Coast Guard

°C

degrees Celsius

°F

degrees Fahrenheit

°K

degrees Kelvin

6

1. EXECUTIVE SUMMARY The demand for natural gas in the U.S. could significantly increase the number and frequency of marine liquefied natural gas (LNG) imports [EIA, 2003]. Because of the increased demand for natural gas, many LNG import terminals around the world are being designed to handle and operate with larger capacity LNG carriers. While studies have been conducted to assess the consequences and risks of potential spills from the current size and capacity of LNG carriers, no hazard studies have been conducted for the emerging larger capacity LNG carriers. Most current LNG carriers transport 125,000 m3 - 145,000 m3 of LNG in either four or five cargo tanks. Many new LNG carriers are being designed to carry up to 265,000 m3 of LNG. The increasing importance of LNG imports suggests that consistent methods and approaches be used to identify the hazards and protect the public and property from a potential LNG spill. For that reason, the U.S. Department of Energy (DOE), Office of Fossil Energy, requested that Sandia National Laboratories (Sandia) assess and quantify the potential hazards and consequences of a large spill from these emerging larger capacity LNG carriers. The effort undertaken for these larger LNG carriers was similar to that presented in the 2004 Sandia report “Guidance on Risk and Safety Assessment of Large Liquefied Natural Gas (LNG) Spills Over Water” [Hightower, et.al., 2004]. For this new effort, DOE specifically requested that Sandia:


Reassess current threat considerations and recommendations from intelligence agencies for marine energy imports,




Evaluate the potential breaching sizes and LNG spill rates and volumes for the larger capacity LNG carriers for both near-shore and offshore operations, and




Assess the range of potential hazards from a spill from these larger capacity LNG carriers, and risk management considerations needed to improve public safety.

To support these efforts, Sandia worked with the U.S. DOE, the U.S. Coast Guard (USCG), LNG industry representatives, and government intelligence agencies to collect background information on ship and LNG cargo tank designs and the most recent credible breach scenarios. The information gathered was used to conduct detailed three-dimensional, dynamic, structural analyses of cargo tank breaching for various scenarios. These results were then used to calculate possible spill rates and volumes, and to model the associated hazards and consequences of the potential spills. While a discussion of the specific threats and expected consequences is beyond the scope of this report, we do discuss the range of breaches that were calculated for these events. A detailed summary of the structural modeling conducted to calculate the potential breaches from various intentional events is presented in an associated report [Luketa, et al., 2008]. The hazard results developed were based on a range of nominal, or most likely, spill conditions and are not site-specific. Site-specific hazard distances will change depending upon the location of the facility, number, size, and type of LNG carriers or regasification vessels used, as well as environmental conditions. Therefore, the hazard results presented are intended to convey the scale of possible hazard distances for a large spill over water from emerging large capacity LNG carriers. While the major hazards expected from an LNG spill for the intentional events 7

considered are thermal hazards from a fire, vapor dispersion distances for potential spills were also calculated. Dispersion is significantly influenced by environmental conditions and potential ignition sources, and the information presented should again be used for identifying the scale of hazards, not necessarily be used for defining hazard distances for a specific site. As noted in the 2004 Sandia LNG report, scenarios could include breaching of more than one LNG cargo tank during intentional events and was considered in these evaluations. Also, cascading damage to an adjacent LNG cargo tank from initial damage to one LNG cargo tank may be possible, based on current experimental data and modeling evaluations, and was considered. As discussed in the 2004 Sandia LNG report, while not considered the most likely LNG spill events, consideration of up to three tanks spilling at any one time is expected to provide a conservative analysis of possible cascading damage concerns and associated hazards. Near-shore Operations Based on these detailed analyses for emerging LNG tanker designs up to 265,000 m3, the range of breach sizes calculated for credible intentional scenarios appropriate for near-shore operations, where there is waterway surveillance, monitoring and control, ranged between 2 – 12 m2. Our analysis suggests that in these near-shore operations, the most likely or nominal intentional events would result in an LNG cargo tank breach of approximately 5 m2. For this size breach in the larger LNG carriers, the spill rates and spill volumes increase slightly and therefore the thermal hazard distances are approximately 7–8% greater than the previous results presented in the 2004 Sandia LNG study for similar event considerations. This is due to the greater amount of LNG above the waterline, or hydrostatic head, and the larger LNG volumes per cargo tank for the larger LNG carriers. Even with the increase in thermal hazard distances from pool fires for the larger ships, the most significant impacts to public safety and property are still within approximately 500 m of a spill, with lower public health and safety impacts at distances beyond approximately 1600 m for near-shore operations. Offshore Operations For offshore operations, where there is less control and surveillance of ship operations, credible intentional scenarios can be larger and the calculated breach sizes can range from 5 – 16 m2, with the most likely or nominal intentional breaching scenario resulting in an LNG cargo tank breach of approximately 12 m2. For offshore LNG facilities where consideration of a breach size of 12 m2 is appropriate, the most significant impacts to public safety and property are within approximately 700 m of a spill, with lower public health and safety impacts at distances beyond approximately 2000 m. Given the location of many proposed offshore facilities, these hazard distances suggest the potential for minimal impact to public safety or property from even a large spill from these larger capacity LNG carriers.

8

2. BACKGROUND The increasing demand for natural gas in the U.S. could significantly increase the number and frequency of marine LNG imports. Net imports of natural gas into the U.S. are expected to grow from 0.5 trillion cubic feet in 2006 to 2.9 trillion cubic feet in 2030 [EIA, 2008]. Currently there are five operational LNG marine terminals. Four to eight new LNG terminals are expected to be constructed in the next four to five years and more than 40 new terminal sites are under consideration and investigation. A factor in the siting of LNG receiving terminals is the proximity to market. Therefore, terminals are being considered in areas with high natural gas demands, which include locations on all three U.S. coasts. Most are being planned to handle one to two LNG tanker shipments per week. A fleet of over 250 specially designed LNG ships is currently being used to transport natural gas around the globe. Worldwide, there are over 20 LNG export (liquefaction) terminals and over 50 import (re-gasification) terminals. This commercial network handles approximately 120 million tons of LNG every year. LNG carriers often travel through areas of dense traffic. In 2000, for example, Tokyo Bay averaged one LNG cargo every 20 hours and one cargo per week entered Boston harbor. Estimates are that world wide LNG trade will increase 35% by 2020. The major areas for increased LNG imports are Europe, North America, and Asia [EIA, 2008]. As LNG imports have increased in the U.S., safety and security concerns have been raised. In response to these concerns, background information on LNG properties, siting processes, and safety and security operations have been developed [FERC, 2004; DOE, 2005; USCG, 2005; Parfomak, 2003 and 2007]. While many studies have been conducted to assess the consequences and risks of potential LNG spills for the current class of LNG carrier, none has been conducted for the newer and larger capacity LNG carriers. Many of the current LNG carriers are designed to carry approximately 125,000 m3 - 145,000 m3 of LNG in either four or five cargo tanks. Because of the increased demand for LNG, many LNG import terminals around the world are being designed to handle and operate with LNG carriers with capacities up to 265,000 m3. The increasing importance of LNG imports suggests that consistent methods and approaches be identified and implemented to help ensure protection of public safety and property from a potential LNG spill. For that reason, the U.S. Department of Energy (DOE), Office of Fossil Energy, requested that Sandia National Laboratories (Sandia) assess and quantify the potential hazards and consequences of a large spill from these larger LNG carrier designs. The effort was similar to what was presented in the 2004 Sandia report “Guidance on Risk and Safety Assessment of Large Liquefied Natural Gas (LNG) Spills Over Water”, for the current class of LNG carriers [Hightower et.al, 2004]. Specifically, DOE requested:


An assessment of the current threat recommendations from intelligence agencies for marine energy imports,




An evaluation of the potential breaching sizes for controlling intentional events against emerging large LNG carriers and potential spill rates and volumes, and




An assessment of the potential range of hazards from an LNG spill over water from these larger capacity LNG carrier designs and potential risk management needs and considerations.

9

To support this effort, Sandia worked with the U.S. DOE, the U.S. Coast Guard, LNG industry groups, and government intelligence agencies to collect background information on ship and LNG cargo tank designs and breach scenarios. The information gathered was used to model potential breach sizes, associated spill rates and volumes, and the extent and severity of hazards to the public and property. 2.1 Overview of LNG Carriers Specially designed ships are used to transport LNG to U.S. import terminals. Some of the special features of LNG ships include:
Construction of specialized materials and equipped with systems designed to safely store LNG at temperatures of -260°F (-162.2°C).
Constructed with double hulls. This construction method not only increases the integrity of the hull system but also provides additional protection for the cargo tanks in the event of accidents.
Coast Guard regulations and the "International Code for the Construction and Equipment of Ships Carrying Liquefied Gases in Bulk" (International Gas Carrier Code) require that LNG ships meet a Type IIG standard, which is an intermediate-level safety design standard for hazardous cargoes that includes requirements on double-hull designs and materials, subdivision, damage stability, and cargo tank location. In general, LNG ships are classified according to the type of system that contains the LNG, either a Moss system or a membrane/prismatic system shown in Figure 1.

(a)

(b)

Figure 1. Typical Containment Systems for LNG Carriers (a) Moss Spherical Design (b) Membrane/Prismatic Design The difference between the two designs is that the Moss system use spheres built from aluminum that contain the LNG and have a structural integrity independent of the ship. For the membrane systems, the LNG is contained within thin, stainless steel membranes directly supported by the hull structure.

10

2.2 Size and Capacity of Emerging LNG Carrier Designs Many new LNG carriers are being designed to carry as much as 265,000 m3 of LNG. The new 215,000 m3 membrane carriers are often referred to as Q-flex designs, and the 265,000 m3 membrane carriers are often referred to as Q-max designs. Table 1 provides an overview of the general size and dimensions of current and emerging LNG carriers for both membrane and Moss-type cargo tank configurations. Table 1. Emerging LNG Carrier Size and Capacity (Poten and Partners, 2006) CLASS Tanks

145,000 m

3

MEMBRANE DESIGNS 3 3 155,000 m 215,000 m

265,000 m

4

4

5

5

Length (m)

283

288

315

345

Width (m)

44

44

50

55

Draft (m)

11.4

11.5

12

12

CLASS Tanks

138,000 m

3

MOSS DESIGNS 3 3 145,000 m 200,000 m

255,000 m

5

4

5

5

Length (m)

287

290

315

345

Width (m)

46

49

50

55

Draft (m)

11

11.4

12

12.5

3

3

From the data presented in Table 1, a couple of key points should be noted. One is that the new larger LNG carrier designs are becoming longer and wider, not necessarily deeper. Because of channel depth limitations in many ports, the new ships are designed to have similar drafts as current LNG carriers. The overall heights are slightly greater with the tank height above the waterline about 20 m versus 15 m for current LNG carriers. Another point is that the volume of LNG per cargo tank is increasing from nominally 30,000 – 40,000 m3 for the current fleet of carriers to as much as 53,000 m3 for the larger LNG carriers. This means that spill rates and the spill volumes from the new large capacity LNG carriers could be larger. There are several variations of the new larger capacity LNG carriers being developed. For example, several new LNG carriers are being designed to include regasification capabilities. With the advent of flexible pipeline and unloading buoy systems, gasification of LNG on the LNG carrier can now be conducted offshore and the natural gas pumped through a flexible riser system down to a sea floor natural gas pipeline and then onto shore. This enables LNG unloading to occur many miles offshore. In some cases this can provide alternatives to on-shore import terminals, which are being considered by the U.S. Coast Guard in several LNG DeepWater Port (DWP) applications. The regasification configured carriers are commonly referred to as Regasification LNG Carriers (RLNGCs). Figure 2 provides a drawing of a planned 205,000 m3 LNG regasification carrier. The main difference between an RLNG carrier and an LNG carrier is that the front cargo storage tanks are often reduced in size to accommodate regasification equipment in the bow, which reduces overall LNG storage capacity. This change in the size of the forward cargo tank can be 11

seen in Figure 2. The regasification equipment and buoy docking system reduce the average LNG cargo capacity by about 10,000 m 3 for both the 215,000 m3 and the 265,000 m3 membrane carriers. While an RLNG carrier contains less LNG, the overall structural design, size, and dimensions are very similar to the emerging large capacity LNG carrier designs, especially the membrane carriers noted in Table 1. Therefore, the results presented in this report are applicable to many of the large regasification LNG carriers being considered and proposed for use at many locations.

Figure 2. Example of a 205,000 m3 Membrane Regasification LNG Carrier A second variation of emerging, large-capacity LNG carriers are vessels commonly called Storage and Regasification Vessels or SRV’s. These vessels often are designed to remain offshore and act as a floating LNG terminal. They are connected through a buoy and riser system, similar to the RLNGC system discussed above, to a sea floor natural gas pipeline that goes onto shore. The SRV’s store LNG supplied and transferred from smaller LNG carriers, regasify the LNG on-board, and then pump the natural gas through the buoy and flexible riser system down to a sea floor natural gas pipeline. While many proposed SRV designs are similar to the emerging large capacity LNG carrier designs that were evaluated, some SRV’s have unique designs, configurations, and operational characteristics developed for specific sites and needs. Therefore, while the results presented in this report may be applicable and representative of some SRV designs and configurations, they may not be applicable to others and site-specific assessments will be required to determine if the results presented in this report would be applicable to a specific SRV.

12

3. THREAT AND BREACH ANALYSES The LNG shipping industry has an exemplary safety record, with only eight accidents over the past 40 years. None of these accidents have led to a loss of life or a breach of the vessel’s cargo containment system. Even with this excellent safety record, consideration should be given to what might be a potential LNG cargo tank breach based on a possible accidental collision with another ship, grounding, or ramming. Based on the previous work on breach sizes for accidental events in the 2004 Sandia LNG study, it is clear that accidental events in near-shore LNG operations are smaller and much easier to mitigate through operational safety improvements than spills caused by intentional events. Therefore, for this report, DOE requested that Sandia focus on assessing the potential breach sizes, spills, and associated hazard distances for credible intentional events against emerging larger capacity LNG carriers carrying up to 265,000 m3 of LNG. 3.1 Analysis of Intentional Threat Scenarios for Large LNG Carriers For the 2004 Sandia LNG report, Sandia worked with intelligence groups and agencies and used historical data to establish a range of potential intentional LNG cargo tank breaches that could be considered credible and possible. This included evaluating information on insider and hijacking attacks on ships, as well as information on external attacks on ships. The level of knowledge, materials, and planning needed to create these types of intentional breaching events was also considered. For this report, Sandia again contacted intelligence agencies and reviewed recent threat information in order to identify the most current estimate by the intelligence agencies of the credible intentional threats to consider in modeling breaching events against the larger capacity LNG carriers. The threats identified and considered included attacks with hand held munitions, attacks with explosives by a team of hijackers or attackers, attacks by boats and airplanes with and without explosives, underwater mines and explosives, as well as consideration of more sophisticated techniques. Additionally, with the development of LNG deepwater ports, threats against both near-shore and offshore operations and facilities were considered. 3.2 Analysis of Intentional Breaching of Large LNG Carriers Based on the credible intentional threats identified, a series of scoping evaluations were conducted to identify the controlling threat scenarios that provide the highest spill rates and largest spill volumes and therefore create the largest hazard zones. For those scenarios and events identified, a series of very detailed, three-dimensional, shock physics-based analyses were conducted. Structural drawings of the large capacity LNG carriers obtained through LNG industry contacts were used to develop detailed three-dimensional models of the LNG carriers. The structural models included all major vessel structural elements including the inner and outer hull thicknesses and materials, all stiffeners and their dimensions and materials, and the frame and web dimensions, spacing, and materials. Figure 3 shows the type of structural elements included in the breach analyses.

13

Figure 3. Example of Large Capacity LNG Carrier Structural Elements Modeled Also included in the analytical model along with the inner and outer hull structural elements were the LNG cargo tanks, insulation, the LNG, and the sea water next to the outer hull. The breach analyses were conducted using a three-dimension shock physics code developed at Sandia called CTH. This computer code is capable of modeling multi-dimensional, multi-material, strong shock-wave controlled physics problems. This type of analysis approach is necessary to accurately model large-scale structural deformations and material responses under the very high strain rates that occur during many intentional threats such as high velocity penetration or explosion scenarios. The detailed three-dimensional breach analyses conducted required a massively parallel computing platform using 920 processors. Each analysis required approximately 2 to 3 weeks of computational time. A number of different threats and threat locations were analyzed based on the controlling threat scenarios identified from the scoping studies. A summary of the detailed structural models and specific analysis results for each of the threat scenarios evaluated is presented in an associated report [Luketa et al, 2008]. Near-shore Operations Breaching Analysis Summary Based on these detailed analyses for the emerging large capacity LNG tanker designs, the range of hole sizes calculated for credible intentional threats appropriate for near-shore operations, where there is waterway surveillance and control, ranged from 2 – 12 m2. Our analysis suggests that in these near-shore operations, the most likely or nominal intentional breaching scenarios would result in an LNG cargo tank breach of approximately 5 m2. The overall results obtained were not significantly different from the breach results identified in the 2004 Sandia LNG report. This is for two reasons. First, the controlling intentional threats since 2004 have not increased significantly, and the general design of the larger LNG carriers are similar to current LNG carrier designs. While there are differences in the general size, dimensions, and thicknesses of many of the structural elements in the larger capacity LNG carriers, these structural differences only have

14

a minor impact on the breach size of the inner hull, which controls the rate of the LNG spilling onto the water. Offshore Operations Breaching Analysis Summary For offshore operations, where there is less waterway control and surveillance of ship operations, credible intentional threats can be larger and the calculated breach sizes can range from 5 – 16 m2, with a most likely or nominal intentional breaching scenario resulting in an LNG cargo tank breach of approximately 12 m2. This range of breach sizes should be considered for facilities or operations about 5 or more miles offshore, where surveillance, control, and risk management of both intentional and accidental events can be much more difficult. Cascading Damage Spill Considerations As noted in the 2004 Sandia LNG report, threats could include breaching of more than one LNG cargo tank during intentional events, and these types of multiple events were considered and the impact of the hazard results discussed in the following chapter. Damage to an adjacent LNG cargo tank from the initial damage to one LNG cargo tank could be possible, based on current experimental data and modeling evaluations, and was considered in our analyses. As discussed in the 2004 Sandia LNG report, multiple tank spills are not considered the most likely or nominal LNG spill event, but should be a consideration in developing risk management and mitigation approaches to LNG spills and associated hazards. Consideration of up to three tanks spilling at any one time is expected to provide a conservative analysis of possible multi-tank damage concerns and associated hazards.

15

4. LNG SPILL HAZARD EVALUATIONS AND RISK REDUCTION The intent of the fire and dispersion analyses is to provide an understanding of the general scale of possible public safety hazards from larger capacity LNG carrier spills. It should be understood that this is not a site-specific analysis which takes into account environmental and surrounding conditions for a particular site. Thus, the results presented are not to be used prescriptively, that is, applied generally to any given site. For a given facility, an analysis which incorporates the particular environmental and facility conditions for that site should be performed as recommended in the 2004 Sandia report. The following discussions provide a general description of the models and assumptions used for the fire and dispersion hazard calculations presented in this report. A more detailed description of the models and approaches can be found in the 2004 Sandia LNG report. Note that the experimental data for several parameters used in these analyses vary considerably. Due to the complexity of the physics involved and the lack of experimental data for an LNG spill and subsequent fire or dispersion event for the large spills expected, a parametric analysis of the sensitivity of different values over the range of applicable experimental data was considered. The various factors that contribute to the variation in LNG hazard analyses has been previously discussed [Hightower, et al., 2004, Luketa, 2006]. As additional experimental data is obtained, the conservatism in the approach used in this report could be reduced. The parametric approach and associated analyses though are useful in providing the scale and range of possible public safety hazards from spills from larger capacity LNG carriers. The analysis results presented can be used by government officials to identify the scale of potential hazards for LNG marine import operations. 4.1 Pool Fire Hazard Analyses As discussed in the 2004 Sandia LNG report, a pool fire is the most likely outcome from the breach of an LNG tanker due to the high probability of immediate ignition of the LNG during the event. The extent of thermal damage to populations and structures from the radiated heat from a pool fire is a function of the total amount of energy received, which depends not only on the magnitude of the heat flux, but also on the area and the orientation of the receiving object relative to the fire, exposure duration, and material properties of the object. Thus, an assessment of the thermal hazards from a pool fire requires evaluating heat flux levels in terms of energy per unit time per unit area (or power per unit area) as a function of distance away from the fire and the fire’s duration and the exposed surface area of a receiving object and its properties. The consideration of these quantities will allow for assessment of the total energy received and hence the extent of thermal damage. For this analysis a solid flame model was used to predict thermal hazard distances at levels that would severely impact populations. A solid flame model represents the surface of the flame with a simple, usually cylindrical geometry. The thermal radiation is uniformly emitted from this surface and the average radiant surface emissive power is based upon empirical correlations with pool diameter. The geometric view factor is modeled, which is the fraction of radiant energy that is received by an object’s field of view. The attenuation of the thermal radiation by water vapor and carbon dioxide in the atmosphere is included in the analysis.

16

The disadvantage of solid flame models is their inability to model more complex flame shapes such as those arising from irregular shaped pools or object interaction with the flame. Therefore, solid flame models are most appropriately used for sites where pool formation is not restricted, such as in wide or open waterways, harbors, bays, or open water. For sites where there are nearby shorelines or structures that can alter the nature of the pool spread and fire, such as modifying pool geometry or through fire interaction with structures, these models have diminished capability to predict hazards. In contrast to an open waterway, the numerous structures comprising an urban environment can affect the distribution of thermal energy or radiated to people and structures. In some cases increased shielding can occur, thereby reducing the thermal energy received, or ‘hot spots’ from recirculation zones or reflecting surfaces can occur, thereby increasing the amount of thermal energy received. Additionally, the presence of obstacles can affect overall thermal hazards by providing additional fuel for latent fire propagation. For those cases where solid flame models are not appropriate for use, many computational fluid dynamics (CFD) based codes have the ability to model irregular pool geometries as well as fire and smoke propagation. However, to be used accurately, a CFD model should be validated for use in the specific application proposed. It should also be noted that thermal damage is only one aspect of assessing the hazards arising from a LNG pool fire, especially in an urban environment. Smoke propagation can become a visibility hazard or a hazard when drafted in through ventilation systems of buildings. Human behavior during fire evacuations can influence the number of casualties/injuries and hence risks to the public since large populations and complicated pathways for exit can restrict effective evacuation efforts. All these factors are considerations in understanding the overall risks from an LNG pool fire in an urban environment, and therefore risk analysis and risk management should be coordinated with local public safety and emergency response organizations to reduce overall risks to the public and property for specific sites. To determine the size of a pool fire, the amount of LNG draining over time from a breached tank, as well as the spreading of LNG on water must be calculated. The spilling and spreading of LNG onto water can be classified as a multi-phase, multi-component problem. In the event of a breach not only will there be LNG flowing out, but there can also be water entering the tank, the degree to which will depend upon the breach size and location. It is expected that any water entering the tank would be turned to ice and in the process would cause the LNG to vaporize. The amount of LNG spilled between the hulls will depend upon the breach size and location, as well as the framing design. The ability of the tank to maintain atmospheric pressure above the height of the LNG is also a consideration. Below we summarize our approach for analyzing each of these different elements of a spill and an associated sensitivity analysis to identify the scale and range of the potential hazards. Most simplified models for the draining of LNG from a tank apply the Bernoulli’s equation which neglects the effect of viscosity. Bernoulli’s equation is a good approximation for large ratios of tank cross sectional to orifice areas (~100 or greater) since viscous effects will be negligible. There are free surface CFD-based codes that can model, using simplifying assumptions, the spilling and spreading of LNG onto water. However the Bernoulli’s equation which was used for this analysis can provide a reasonable approximation for the rate of LNG

17

flowing out of a tank that is in accord with the intent of providing the general scale of the range of hazards from these events. Once spilled onto the water, the shape and size of a spreading LNG pool can be affected by several factors: wind, waves, currents, confinement, composition, rapid phase transitions, and object interaction. Despite these complexities, in order to obtain an estimate of pool size, a steady mass balance can be utilized in which the mass flux of LNG flowing into the pool is balanced by the mass flux being evaporated. The results presented in this analysis used such an approximation. The pool will grow and then eventually shrink and break up after reaching a maximum diameter. The results presented pertain to the maximum pool diameter during spreading assuming an average flow rate from the tank. It should be noted that the hydrostatic head and cargo tank volumes differ between the current and emerging larger LNG carriers. While the nominal LNG level in the cargo tanks extends approximately 15 m above the waterline for current LNG carriers, the LNG level for the larger LNG carriers is approximately 20 m above the waterline. Spill volumes for the larger LNG carriers are about 41,000 m3. These result in slightly higher spill rates and larger spill volumes, which result in estimated pool diameters and associated hazards for the larger LNG carriers that are slightly larger than for the current class of LNG carriers. As was done in the 2004 Sandia LNG report, nominal fire modeling parameters along with variations around the nominal case were used to calculate the thermal hazards. The justification for the range of values used can be found in the Appendix. Due to the non-site specific nature of the analysis, the affect of wind tilting the flame was not included. It should be noted that a minor modification to the calculation procedure from the 2004 Sandia LNG report has been made by way of incorporating an average among several flame height correlations instead of using a single flame height correlation. This approach was used because of the lack of large-scale data to identify the best correlation and results in about a 2% decrease in the average thermal hazard distance relative to past analyses. The flame height correlations considered and the approach is presented in the Appendix. Tables 2 and 3 provide the results for thermal hazards from a pool fire for near-shore breach events and for offshore breach events respectively. The analyses present hazard distances for heat flux levels of 37.5 kW/m2 and 5 kW/m2. The 37.5 kW/m2 value is a level at which process equipment is damaged after 10 minutes of exposure, and is currently considered to represent the extent of hazards to structures and equipment. The 5 kW/m2 value is a level at which seconddegree burns occur on bare skin after 30 seconds of exposure, and is currently considered to represent the extent of hazards to people in an open area. The time and length scales for cascading cryogenic or thermal damage to additional LNG cargo tanks is unknown at this time because rapid multi-tank failures involve very complex physical process that will be an area of ongoing research for some time to come. In order to address the potential for cascading damage and possible hazard distances, an analysis of the breach and spill from three LNG cargo tanks at one time was conducted. Each tank breach assumes a similar hole-size with simple orifice flow. The assumption is that all tanks could possibly fail, which would affect the fire duration, but in the short timescales that it takes to reach a maximum fire

18

size from a large spill, only three tanks were considered to be contributing to the maximum size of the pool fire. Near-shore Operations Pool Fire Hazard Analysis Results For near-shore operations, the intentional breach cases considered for the emerging larger capacity LNG carriers are presented in Table 2. The average thermal hazard distance for the nominal or most likely breach size of 5 m2 for a 41,000 m3 spill is about 450 m for the 37.5 kW/m2 level and 1400 m for the 5 kW/m2 level. For comparison, the nominal hazard distance results presented in the 2004 Sandia LNG report for the smaller size LNG carriers for similar breach sizes at the 5 kW/m2 level was about 1300 m. Thus, the increase in hydrostatic head and tank volumes for the larger capacity LNG carriers results in an approximately 7 – 8 % increase in the thermal hazard distances, and increased fire durations. The results though indicate that the thermal hazard distances for the 37.5 kW/m2 and 5 kW/m2 heat flux levels for the larger LNG ships for near-shore locations are still expected to be within the 500 m and 1600 m hazard zones suggested in the 2004 Sandia LNG report. In Tables 2 and 3, “τ” is the atmospheric transmissivity, which is discussed in the Appendix. Table 2. Thermal hazard distances from a pool fire for near-shore operations SURFACE EMISSIVE POWER (kW/m2)

HOLE SIZE (m2)

TANKS BREACHED

DISCHARGE COEFFICIENT

BURN RATE (m/s)

2

3

0.6

3 x 10-4

220

0.6

-4

220

DISTANCE TO 5 37.5 kW/m2 kW/m2 (m) (m)

POOL DIAMETER (m)

BURN TIME (min)

0.8

225

57

282

881

0.8

615

23

774

2197

τ

INTENTIONAL EVENTS

5

3

3 x 10

-4

5*

1

0.6

3 x 10

220

0.8

355

23

446

1344

5

1

0.3

3 x 10-4

220

0.8

251

46

315

975

0.6

-4

220

0.8

435

23

547

1487

-4

5

1

2 x 10

5

1

0.6

8 x 10

220

0.8

217

23

273

1042

5

1

0.6

3 x 10-4

220

0.5

355

23

305

1050

-4

5

1

0.6

3 x 10

175

0.8

355

23

373

1188

5

1

0.6

3 x 10-4

350

0.8

355

23

617

1683

0.6

-4

220

0.8

550

10

692

1981

12

1

3 x 10

*nominal case

Offshore Operations Pool Fire Hazard Analysis Results For offshore operations, generally 5 of more miles offshore, intentional threats can be larger and as noted in Section 3 result in a larger nominal or most likely breach size of 12 m2. The hazard distance results calculated for a 41,000 m3 spill and a range of possible breach sizes for offshore operations for the larger LNG carriers are shown in Table 3. The results suggest that for offshore operations and associated breach events, the thermal hazard distance at the 37.5 kW/m2 and 5 kW/m2 heat flux levels are approximately 700 m and 2000 m, respectively.

19

Table 3. Thermal hazard distances from a pool fire for offshore operations

HOLE SIZE (m2)

TANKS BREACHED

DISCHARGE COEFFICIENT

BURN RATE (m/s)

SURFACE EMISSIVE POWER (kW/m2)

τ

POOL DIAMETER (m)

BURN TIME (min)

DISTANCE TO 5 37.5 kW/m2 kW/m2 (m) (m)

INTENTIONAL EVENTS 5

3

0.6

3 x 10-4

220

0.8

615

23

774

2196

12

3

0.6

3 x 10-4

220

0.8

953

9.6

1090

3168

0.6

-4

220

0.8

550

9.6

692

1980

-4

12*

1

3 x 10

12

1

0.3

3 x 10

220

0.8

389

19

466

1429

12

1

0.6

2 x 10-4

220

0.8

674

9.6

786

2335

12

1

0.6

8 x 10-4

220

0.8

337

9.6

407

1261

-4

12

1

0.6

3 x 10

220

0.5

550

9.6

462

1539

12

1

0.6

3 x 10-4

175

0.8

550

9.6

553

1738

-4

12

1

0.6

3 x 10

350

0.8

550

9.6

864

2452

16

1

0.6

3 x 10-4

220

0.8

635

7.2

741

2202

*nominal case

4.2 Vapor Dispersion Analyses For the controlling intentional breach events identified in this study, which cause the biggest breaches and largest spills, there is a high expectation that the events will provide secondary ignition sources that will provide ignition of any spilled LNG. Additionally, if a dispersion event does occur, the vapor cloud could ignite from ignition sources on the ship itself. Thus, the probability of a natural gas cloud fully extending, especially in a near-shore populated area where many ignition sources exist, and then igniting is very low. The cloud will most likely ignite when it comes in contact with the first available ignition source. Since the possibility of a dispersion event though cannot be ruled out, dispersion calculations were performed to determine the distance to the lower flammability limit (LFL) for a vapor cloud from an un-ignited LNG spill from these emerging larger capacity LNG carriers. The LFL for natural gas can change slightly depending on the experimental conditions and measurement techniques used. For this analysis it is defined as a 5% concentration of methane in air by volume for ambient conditions [Liao, et al, 2005]. These calculations were performed using Vulcan [Nicolette, 1996, Holen, et al, 1990], a computational fluid dynamics (CFD) based code. It should also be realized that the hazard zone area is elongated in the downwind direction from the spill point, rather than spread over a uniform circle, for a dispersion event. Therefore, dispersion distances and associated hazards are significantly influenced by site-specific environmental and operational conditions. For the analyses and information presented in this report, nominal environmental and atmospheric conditions were assumed. Therefore, the information presented should be used to identify the scale of possible dispersion hazards from a potential spill from the larger capacity LNG carriers, not necessarily for defining hazard distances for a specific site. For a site where dispersion issues may be a concern, a site-specific dispersion calculation should be conducted using wind, topography, and environmental conditions for that location to assess potential impacts on public safety and property. Guidance

20

on performing vapor dispersion calculations using CFD codes have been discussed by Luketa, et al., 2007. Near-shore Operations Vapor Dispersion Analysis Results The information presented is based on calculations performed for stable atmospheric conditions with a wind speed of 2.33 m/s for the near-shore intentional events that nominally cause a 5 m2 breach of 1 tank spilling 41,000 m3 of LNG, or approximately 70% of its contents. Dispersion analyses are more appropriate for the nominal single tank spill events. Multiple tank spills and large vapor dispersions are more unlikely due to the multiple ignition sources available for these events and the fact that cascading multiple tank spill scenarios are often from fire damage, such that an ignition source for most cascading spill scenarios is present. For the given spill volume and head, an LNG pool will be created that lasts for about 1380 seconds. The vapor generation and vapor flow conditions from the LNG pool were calculated using a liquid density of 450 kg/m3 and a vapor density of 1.74 kg/m3. Two values for evaporative mass flux were evaluated that span the range of experimental values reported in the literature, which vary by an order of magnitude. Table 4 indicates that the distance to the LFL ranges from 2800 to 3300 m with an average of 3050 m. Table 4. Distance to the LFL for vapor dispersion for near-shore operations POOL DIAMETER (m)

HOLE SIZE (m2)

NUMBER OF TANKS

290

5

1

4.5 x 10

-4

2800

917

5

1

4.5 x 10

-5

3300

MASS FLUX (m/s)

DISTANCE TO LFL (m)*

*Assumes no Ignition source along path

Offshore Operations Vapor Dispersion Analysis Results The information presented is based on calculations performed for stable atmospheric conditions with a wind speed of 2.33 m/s for the offshore intentional threat scenarios that nominally cause a 12 m2 breach of 1 tank spilling 41,000 m3 of LNG. Again, dispersion analyses are more appropriate to consider for the nominal single tank events. The larger spill rate will create a larger LNG pool but will last for only about 576 seconds. Again, the vapor generation and flow conditions from the LNG pool were based on a liquid density of 450 kg/m3 and a vapor density of 1.74 kg/m3, and two values of evaporative mass flux were used that span the range of experimental values noted above. Table 5 indicates that the distance to the LFL ranges from 4000 to 5200 m with an average of 4600 m. As noted, this distance will change depending upon the offshore facility design, operations, environmental conditions, and the number of ships that might be involved in an event. The analyses presented should be used as a guide on the scale of potential hazards, but site-specific analyses may have to be considered for many offshore operations because of the variability in operational scenarios. Table 5. Distance to the LFL for vapor dispersion for offshore operations POOL DIAMETER (m)

HOLE SIZE (m2)

NUMBER OF TANKS

450

12

1

1420

12

1

MASS FLUX (m/s)

DISTANCE TO LFL (m)*

4.5 x 10

-4

4000

4.5 x 10

-5

5200

*Assumes no Ignition source along path

21

4.3 Hazard and Risk Reduction Considerations Risk prevention and mitigation techniques can be important tools in reducing both the potential for a spill and the hazards from a spill, especially in zones where the potential impact on public safety and property can be high. However, what might be applicable for cost-effective risk reduction in one location might not be appropriate at another. Therefore, coordination of risk prevention and management approaches with local and regional emergency response and public safety officials is important in providing a comprehensive, efficient, and cost-effective approach to protecting public safety and property at a specific site. Near-shore Operations The analyses presented suggest that for near-shore operations, a nominal intentional event would result in an LNG cargo tank breach of approximately 5 m2. For this size breach in the larger capacity LNG carriers, the spill rates and spill volumes increase slightly and therefore the thermal hazard distances are approximately 7–8% greater than the results presented in the 2004 Sandia LNG study for current LNG carrier designs. With this modest increase in thermal hazard distances, the most significant impacts to public safety and property for near-shore operations are still approximately 500 m of a spill, with lower public health and safety impacts at distances beyond approximately 1600 m. Also, potential vapor dispersion distances for near-shore operations are similar to those suggested in the 2004 Sandia report. As such, the risk mitigation and risk management approaches suggested in the 2004 report are still appropriate for use with the larger capacity ships. Proactive risk management approaches can reduce both the potential and the hazards of such events. The approaches could include:
Improvements in ship and terminal safety/security systems,
Modifications to improve effectiveness of LNG tanker escorts, vessel movement control zones, and safety operations near ports and terminals,
Improved surveillance and searches, and
Improved emergency response coordination and communications with first responders and public safety officials. Offshore Operations For offshore operations, where there might be less surveillance or control, credible intentional threats could be larger, with a nominal breach size of about 12 m2. From the analyses presented, the most significant impacts to public safety and property from an LNG spill and fire are within approximately 700 m of a spill, with lower public health and safety impacts at distances beyond approximately 2000 m. Vapor dispersion distances for a spill for these offshore operations for the larger capacity LNG carriers or regasification carriers could extend up to nominally 5000 m. Given the location of many of these proposed offshore facilities, the hazard distances suggest that there might be minimal impact to public safety or property from even a large spill. As such, risk management might best be directed at providing approaches, measures, or systems to ensure that the offshore facilities and operations are maintained sufficiently offshore such that they do not inadvertently or inappropriately impact near-shore public safety and property.

22

5. CONCLUSIONS Because of the increasing size and capacity of many new LNG carriers, the Department of Energy requested that Sandia assess the general scale of possible hazards for a breach and spill from newer LNG carriers with capacities ranging up to 265,000 m3. Building on the research and analyses presented in Sandia’s 2004 LNG report, we reassessed emerging accidental and intentional threats and then conducted detailed three-dimensional breach analyses for several new large capacity (up to 265,000 m3) LNG carrier designs. Based on the estimated breach sizes, breach locations, and LNG carrier configurations, we estimated LNG spill rates and volumes and conducted thermal hazard and vapor dispersion analyses. The results include analysis of the hazards of potential LNG spills at both near-shore and offshore facilities and operations, which should help improve the understanding of the range of hazards for different marine LNG import options. The results can be summarized as follows: Near-shore Operations • For the identified breach scenarios for near-shore LNG marine import operations, the calculated breach sizes to the inner hull range between 2 – 12 m2. Our analysis suggests that intentional breaching scenarios would result in a nominal tank breach of 5 m2. • The estimated thermal hazard distances from a pool fire for the larger capacity LNG carriers are approximately 7–8% greater than the distances presented in the 2004 Sandia LNG study for near-shore operations. This is due to the greater amount of LNG above the waterline, or hydrostatic head, for the larger capacity LNG carriers versus current LNG carrier designs. • Even with the increase in thermal hazard distances from pool fires for the larger ships, the most significant impacts to public safety and property are still within approximately 500 m of a spill, with lower public health and safety impacts at distances beyond approximately 1600 m. • Based on current threats, it is possible that more than one LNG cargo tank could be breached. This includes cascading failure to adjacent cargo tanks from the initial damage. This type of damage is possible and should be considered as a variation of the nominal case in site-specific evaluations. • While the most likely outcome of a large spill from an intentional event is expected to be a pool fire, a vapor dispersion analysis was conducted. The average distance to the vapor dispersion LFL from an LNG spill over water for a nominal 5 m2 breach would be about 3,050 m. This result was obtained from the range of 2800 m - 3300 m obtained when considering a range of mass flux values. • The likelihood of a natural gas cloud fully extending, especially in a near-shore urban area, and then igniting is very low. The cloud will most likely ignite from the first available ignition source and progress to a pool fire. For a dispersion event, the hazard zone area is elongated in the downwind direction from the spill point, rather than spread over a uniform circle. • Pool fire and vapor dispersion hazard distances are significantly influenced by site-specific environmental, topographical, and operational conditions. The results presented use nominal environmental and operational conditions and can be used to identify the general

23

•

scale of possible hazards from a potential spill, but should not be used to define hazard distances for a specific site. For near-shore operations, risk prevention and risk management should be considered as ways to reduce the hazards to public safety and property, especially for near-shore operations.

Offshore Operations • For the identified breach scenarios for offshore LNG marine import operations, the calculated breach sizes to the inner hull range between 5-16 m2. Our analysis suggests that intentional breaching scenarios would results in a nominal tank breach of 12 m2. • The most significant impacts to public safety and property from pool fires are within approximately 700 m of a spill, with lower public health and safety impacts at distances beyond approximately 2000 m. The 2004 Sandia LNG study did not conduct threat, breach, and hazard analyses for offshore facilities, such as LNG deepwater ports or other offshore facilities, since these facilities were not in operation at that time. • Based on current threats, it is possible that more than one LNG cargo tank could be breached. This includes cascading failure to adjacent cargo tanks from the initial damage. This type of damage is possible and should be considered as a variation of the nominal case in site-specific evaluations. • While the most likely outcome of a potential LNG spill would be a pool fire, a vapor dispersion analysis was conducted. The average distance to vapor dispersion LFL from an LNG spill over water for a nominal 12 m2 breach would be about 4,600 m. This result was obtained from the range of 4000 m - 5200 m obtained when considering a range of mass flux values. • As noted above, the hazard zone area for a vapor dispersion event is elongated in the downwind direction from the spill point, rather than spread over a uniform circle and will likely ignite when it encounters the first ignition source. For offshore operations, there may be fewer ignition sources relative to near-shore operations. • Pool fire and vapor dispersion hazard distances are significantly influenced by site-specific environmental and operational conditions. The results presented use nominal environmental and operational conditions and can be used to identify the general scale of possible hazards from a potential spill, but the wide variety of offshore facility designs, operations, and number of LNG ships and designs being considered suggest that the results presented should not be used to define hazard distances for a specific offshore facility. • For offshore operations, risk prevention and risk management may have a different focus than near-shore operations, since many spills and associated hazards might not impact the on-shore public and property. Overall, the results obtained from the more detailed analyses conducted and presented in this report for the emerging larger capacity LNG carriers are similar to the previous conclusions, recommendations, and guidance presented in the 2004 Sandia LNG report concerning the general scale of hazards to the public and property from a large LNG spill over water and approaches to reduce those risks and consequences.

24

APPENDIX The following provides a discussion of the parameter values used to predict the thermal hazard distances from a large-scale LNG pool fire on water. It is recommended that the range of parameter values provided be incorporated into site specific analyses that use a solid flame model. A1. Surface Emissive Power The surface emissive power has been shown to initially increase with increasing pool diameter as indicated by the Montoir experiments on land for pool fires up to 35 m [Nedelka,1989; British Gas, 1988; Tucker, 1988]. There is indication that the surface emissive power asymptotes to a maximum value somewhere between 257 – 273 kW/m2 when plotted as a function of pool diameter. The limit appeared to be reached near a pool diameter of 35 m, and thus the surface emissive power would not be expected to significantly increase for larger diameters. Beyond this maximum value, the surface emissive power would expect to decrease with increasing diameter due to greater smoke production. Smoke is made up of a mixture of gases, vapors, and particulate matter from a fire. Carbon particulates, or soot, is included as a particulate matter of smoke and is responsible for the luminosity of the fire. Smoke will absorb a significant portion of the radiation to result in much lower emission to the surroundings. Soot and smoke is a result of incomplete combustion which is affected by radiative losses and limited oxygen supply. In the Montoir experiments, smoke shielding was observed in the upper half of the 35 m diameter LNG fires, while the lower half was highly emissive and essentially smoke free. This behavior is observed with heavier hydrocarbon fuels, but with smoke shielding occurring much closer to the fuel surface in an equivalent sized fire. The emissive power of black smoke is approximately 20 kW/m2. Periodically the flame will break through the smoke, revealing areas of higher surface emissive power around 120 kW/m2. Thus, for heavier hydrocarbons the timeaveraged, area-weighted surface emissive power asymptotes to a value of about 40 kW/m2. Thus, it would be expected that LNG, at some pool diameter, would display similar behavior, but the diameter at which this occurs is unknown due to lack of data at very large scales and cannot be predicted analytically based upon existing data sets. Although it’s expected that the average surface emissive power will drop below 200 kW/m2 for pool diameters 100 m and greater, it is unknown by how much it will decrease. It is recommended that until additional data is obtained, due to safety considerations, a conservative value for surface emissive power should be used when applying a solid flame model by the range of values of ±50 kW/m2 around 220 kW/m2 based on existing data for LNG pool fires on water [Raj, 1979]. The maximum value of 350 kW/m2 obtained from narrow-angle radiometer measurements from the Montoir tests could also be included as a data point for uncertainty analysis. A2. Fuel Volatilization Rate The fuel volatilization rate, herein called the burn rate, will affect the size of pool, with higher burn rates resulting in smaller pools. Higher burn rates also increase flame height, hence there is

25

a trade-off in the effect that burn rate has on thermal hazard distances. While a decrease in pool area will tend to reduce hazard distances, the increase in flame height will tend to increase hazard distances. However, the overall affect will be to decrease thermal hazard distances for increasing burn rates. The only experiment able to obtain burn rate data for LNG pool fires on water are the tests funded by the USCG which reported calculated burn rates ranging from 4x10-4 to 11x10-4 m/s [Raj, 1979]. The calculations use the total quantity spilled divided by the approximate pool area and time of ‘intense’ burning to derive the burn rate. The volume spilled during steady state burning was less than the total volume of LNG spilled. By using the total volume of LNG spilled rather than the volume spilled during steady burning, higher burn rates are calculated. If burn rates are calculated based upon dividing the reported values for spill rate by the pool area, then burn rates vary from 2.6 x10-4 to 9x10-4 m/s. Since the burn rate is a function of the heat transfer from the flame and from the water, the range of burn rates for LNG pool fires on water can be estimated by combining data from pool fire experiments on land and un-ignited spill tests on water. The Montoir tests report an average mass burn rate of 0.14 kg/m2s performed in wind speeds that ranged from about 3 – 10 m/s. The mass burn rate was calculated from dip tube measurements assuming a liquid density of 500 kg/m3. This indicates a burn rate of 2.8 x 10-4 m/s for an LNG pool fire on land. It should be noted that uncertainties in the burn rate measurements were not provided. The range of values for mass flux derived from un-ignited LNG pools on water range from 0.64 x 10-4 to 4.3 x 10-4 m/s with no uncertainty values reported [Boyle, 1973; Burgess, 1970; Feldbaur, 1972; Koopman, 1978]. If these values are added to the Montoir data, the range of burn rate values for pool fires on water would be 3.4 x 10-4 to 7.1 x 10-4 m/s. The higher values in this range, above what was calculated from the LNG pool fire tests on water, could be due to inadequate measurements, differences in LNG composition, and different wind conditions. Pool fire tests conducted at China Lake and Sandia National Laboratories have indicated that wind speeds can significantly affect burn rate as shown in Figure A-1 [Blanchat, 2006 and 2002]. Thus, there is significant uncertainty concerning current burn rate data for LNG, and it is unknown what burn rates would result for pool diameters 100 m or greater.

7

7

6 regression (mm/min)

regression (mm/min)

6 5 4 3 2

5 4 DP Fuel Regression (mm/min)

3

Rake Fuel Regression (mm/min)

2

Linear (Rake Fuel Regression (mm/min))

1

Linear (DP Fuel Regression (mm/min))

0

1 0

2

4 wind speed (m/s)

6

8

0

1

2

3

4

5

6

7

wind speed (m/s)

Figure A-1: Regression rates as a function of wind speed for (a) 18.9 m JP-8 pool fire, 4000 gallons, China Lake [Blanchat, 2006] and (b) 7.9 m JP-8 pool fire, 2200 gallons, Sandi[Blanchat, 2002].

26

It is recommended that a range of burn rates be used when applying a solid flame model by considering the range of 2 x 10-4 to 8 x 10-4 m/s. A3. Flame Height There is great uncertainty in predicting flame height for large pool diameters for coherent fire plumes. Time-averaged flame height is usually defined as the height at which the intermittency reaches a value of 0.5, while maximum height is defined at an intermittency level of 0.05. Intermittency is defined as the fraction of time the flame is at a certain height. It has been demonstrated that the flame height of pool fires decreases for increasing pool diameters. It does become increasingly difficult to determine flame height for increasing pool diameters due to the obscuration of smoke and the periodic appearance of much higher temperature luminous zones. Several flame height correlations based upon pool fires much smaller than the diameters presently considered have been developed. The majority of flame height correlations are based on a combination of flame height measurements and dimensional arguments such as the model by Thomas [Thomas, 1963], or experimental measurement combined with theoretical mathematical models such as the model by Steward [Steward, 1970]. Several correlations such as Moorhouse [Moorhouse, 1982] have based their correlation on Thomas’s dimensional form. That is,

 m&¢¢  L = a  D  ra gD 

b

(1)

where L is flame height, D pool diameter, m&¢¢ burn rate (kg/m2 s), and ρa atmospheric air density. The best fitting coefficients, a and b, for LNG fires from 29 tests ranging from 6.9 m to 15.4 m of equivalent pool diameters were identified by Moorhouse, while Thomas determined the coefficients with experimental data from wood crib fires up to 2 meters. It should be noted that LNG is very different than other hydrocarbons in its propensity to not produce as much smoke. Thus, correlations developed for hydrocarbons other than LNG may have significant error for predicted flame heights. In spite of similarities in fundamental approaches, a fascinating feature of these correlations is that they all predict different flame heights for any given hydrocarbon at a specified pool diameter. Predicted flame height to pool diameter ratios (L/D) for LNG can vary by a factor of 2 to 3 for a given pool diameter as shown in Figure A-2.

27

3.5 Zukoski Thomas 1 Thomas 2 Heskestad Moorhouse (LNG tests) Pritchard & Binding (LNG tests)

height/diameter (L/D)

3 2.5 2 1.5 1 0.5 0 0

100

200 300 diameter (m)

400

500

Figure A-2: Flame height correlations as a function of pool diameter for LNG.

There are at least twenty correlations, but only a few are plotted in Figure A-2 to indicate the range of disagreement. The variation among the correlations may be due to differences in the pool geometry tested and environmental conditions, as well as differences in the measurement technique and definition of flame height. It should also be noted that these correlations assume that the flame is characterized by a single temperature and gas composition regardless of the flame size or soot concentration in the flame. They also do not take into account fuel radiation properties, or turbulent mixing either from the mechanisms due to the fire or induced by the atmosphere. Thus, this justifies their classification of ‘correlation’ as well as their associated uncertainty. Table A-1 shows a comparison among several flame height correlations and the largest LNG pool fire data sets. The burn rates reported from the experiments were used. From this comparison it would indicate that the correlation by Pritchard and Binding [Pritchard, 1992] is most appropriate for LNG, even at very large diameters, but caution should be used in coming to this conclusion since all of the correlations have been developed for a limited range of diameters and a large extrapolation could have significant error. Table A-1: Comparison of several flame height predictions for LNG pool fire tests Diameter (m) 8.5 (test 1 china lake) 9 (test 4 china lake) (2.2 m/s) 20 (Mizner, Eyre - land) (6.2 m/s) 35 (Montoir) (9 m/s)

Experiment (L/D)AVERAGE

L/D predicted Pritchard

Moorhouse

Thomas

Zukoski*

Steward*

Heskestad*

2.8

2.8

2.0

3.0

4.7

4.1

3.6

2.8

2.6

1.9

2.5

3.9

3.7

3.1

2.15

2.2

1.6

1.6

2.9

3.1

2.4

2.2

2.2

1.6

1.5

2.9

3.1

2.4

*no correction term for wind conditions

28

Since most correlations predict a flame height to diameter ratio between 1 and 2 for pool diameters 100 m and greater, it is recommended that when applying a solid flame model a range of L/D values of 1-2 be used. In this analysis, the correlations specified in Figure A-2 were used to obtain an average flame height as a function of pool diameter and burn rate. A4. Flame Tilt and Drag When pool fires are subject to wind they will tilt in the down wind direction and the base dimension of the flame will extend in the downwind dimension, also termed flame drag, while the upwind and crosswind dimensions remain unchanged. The affect of flame tilt and drag is to create an elliptical pool and increase the thermal hazard distances in the downwind direction. Flame tilt and drag have been observed in both the Maplin Sands and Montoir LNG land tests, as well as the tests performed by Moorhouse [Moorhouse, 1982]. LNG pool fire land tests reported flame drag to be between 5 m and 10 m for a 20 m diameter pool in a wind speed of 6.16 m/s, with a flame tilt of 54 degree from vertical [Mizner, 1982]. The Montoir tests reported flame drag to be 10 m for a 35 m diameter pool in wind speed of 9-10 m/s, and 7 m for wind speeds of around 2 m/s [Nedelka, 1989;British Gas, 1988; Tucker, 1988]. The tilt decreased with increasing height, so that the tilt was about 50 degrees from vertical in the bottom portion of the flame (up to L/D ~ 0.5), and 35 degree from vertical for the remaining height. Thus, a single flame tilt value could not be used to describe the flame. From the China Lake tests on water, a flame tilt of 26.5 degrees from vertical was reported for test 4 in a 2.2 m/s wind [Raj, 1979]. For many tests the shape of the pool was observed to become elliptical. The correlations developed by American Gas Association [AGA, 1974], and Moorhouse [Moorhouse, 1982] to predict flame tilt and drag for integral models have been developed from LNG pool fire land tests. It is recommended that an integral model include the flame drag and tilt for facility locations in which non-calm wind conditions exist. A variability of ± 30% of calculated values for flame tilt and drag should be included to account for the variability demonstrated from test data. A5. Atmospheric Attenuation The radiation that is emitted from a flame to the surroundings will be attenuated principally by absorption from CO2 and H2O in the atmosphere. Transmissivity is a measure of this attenuation and is a function of temperature and humidity since water vapor in the atmosphere depends upon temperature and relative humidity. It is very difficult to obtain accurate data on transmissivity because it depends upon knowledge of radiative spectral emission over the surface of the flame, as well as the absorption through the atmosphere to a receiving object. Thus, there is great uncertainty associated with transmissivity. Transmissivity curves are calculated based on LNG pool fires on water from the tests performed at China Lake and range from 0.4 to 0.9 for different humidity levels over a path length of 10,000 m [Raj, 1979]. The uncertainty associated with this calculation was not reported. To account for experimental uncertainty, integral models should account for the variability of transmissivity by considering a range of values. Note that the transmissivity is a function of humidity and distance. Even though the level of uncertainty to incorporate is not known precisely, experiments conducted in outdoor environment will commonly have uncertainties 29

around ±30%. Thus, given a transmissivity function for a particular humidity level, a variability of ±30% should be considered. The upper bound value should not exceed the value of 1.

30

REFERENCES AGA (1974) American Gas Association, LNG safety research program, Report IS 3-1, 1974. Blanchat, T.K., et al. (2006) Well-characterized open pool experiment data and analysis for model validation and development, SAND2006-7508, 2006. Blanchat, T.K., Manning L. (2002) Mock B52 bomb bay fire experiment data and analysis for model validation and development, SAND2002-0145, 2002. Boyle, G. J., Kneebone A. (1973) Laboratory investigations into the characteristics of LNG spills on water. Evaporation, spreading and vapor dispersion, Shell Research Ltd., Thornton Research Centre, Report 6-32, March 1973. British Gas (1988) Results of detailed measurements made on a 35m diameter LNG pool fire experiment conducted at Montoir, France, 3 volumes, British Gas, August 1988. Burgess, D. S., et al. (1970) Hazards associated with the spillage of LNG on water, Report 7448, Bureau of Mines, Pittsburgh, PA. DOE (2005) “Liquefied Natural Gas: Understanding the Basic Facts,” DOE/FE-0489, Department of Energy, Office of Fossil Energy, Washington, D.C., August 2005 EIA (2003) The Global Liquefied Natural Gas Market: Status & Outlook, DOE/EIA-0637, Energy Information Administration, Washington, DC, December 2003. EIA (2008) Annual Energy Outlook, Natural Gas, Energy Information Administration, Washington, D.C., http://www.eia.doe.gov/oiaf/aeo/index.html. Feldbauer, G. F., et al. (1972) Spills of LNG on water – vaporization and downwind drift of combustible mixtures, API Report EE61E-72, 1972. FERC (2004) “A Guide to LNG – What All Citizens Should Know”, Federal Energy Regulatory Commission, Office of Energy Projects, Washington, D.C., www.ferc.gov/industries/lng.asp, 2004. GAO (2007) “Public Safety Consequences of a Terrorist Attack on a Tanker Carrying Liquefied Natural Gas Need Clarification,” GAO-07-316, United States Government Accountability Office, March 2007. Hightower, M. Michael, Gritzo, Louis A., Luketa-Hanlin, Anay, Covan. John M., et. al., “Guidance on Risk Analysis and Safety Implications of a Large Liquefied Natural Gas (LNG) Spill Over Water”, SAND2004-6258, Sandia National Laboratories, Albuquerque, NM, December 2004.

31

Hightower, M. Michael, Kaneshige, Michael J., “Threat and Breach Analysis of an LNG Ship Spill Over Water”, Sandia National Laboratories, Albuquerque, NM, May 2005. Holen, J., Brostrom, M., and Magnussen, B.F. (1990), Finite Difference Calculation of Pool Fires, Twenty-Third Symposium (International) on Combustion, The Combustion Institute, pp. 1677-1683, 1990. Koopman, R.P., et al. (1978) Data and calculations on 5 m3 LNG spill tests, Lawrence Livermore Laboratory, UCRL-52976, Livermore, CA. Liao, S.Y., et al. (2005) Experimental study of flammability limits of natural gas-air mixture, J. of Hazardous Materials, B119, 81-84. Luketa-Hanlin, A. (2006) A review of large-scale LNG spills: experiments and modeling, J. Hazardous Materials, A132, 199-140. Luketa, A., Attaway, S., Hightower, M. (2008) Evaluation of Intentional Breaches of the Larger Class of Liquefied Natural Gas (LNG) Carriers, Draft report. Luketa-Hanlin, A., Koopman, R.P., Ermak, D.L., (2007) On the application of computational fluid dynamics codes for liquefied natural gas dispersion, J. Hazardous Materials, Invited article for special edition, 140, 504-517. Mizner, G. A., Eyre, J. A., (1982) Large-Scale LNG and LPG pool fires, EFCE Publication Series (European Federation of Chemical Engineering) 25, 147-1. Moorhouse, J., (1982) Scaling criteria for pool fires derived from large scale experiments, I. Chem.E. Symposium Series, 71, 165-179. Nedelka, D., et al., (1989) The Montoir 35 m diameter LNG pool fire experiments, Int. Conf. Liq. Nat. Gas, v. 2, 9th, 17-20 Oct 1989, Nice, France. Nicolette, Vernon F., (1996) “Computational Fire Modeling for Aircraft Fire Research”, SAND96-2714, Sandia National Laboratories, Albuquerque, NM, November 1996. Parfomak, P., (2003) Liquefied Natural Gas (LNG) Infrastructure Security: Background and Issues for Congress, Congressional Research Service, The Library of Congress, RL32073, September 2003. Parfomak, P., Vann A., (2007) Liquefied Natural Gas (LNG) Import Terminals: Siting, Safety, and Regulation, Congressional Research Service, The Library of Congress, RL32205, May 2007. Pritchard, M.J., Binding T.M. (1992) FIRE2: A new approach for predicting thermal radiation levels for hydrocarbon pool fires, Sym. major hazards onshore and offshore, 491-505.

32

Poten & Partners, (2006) LNG Shipping Reference Tables, www.poten.com, Copyright 2006. Puttock, J. S., et al., (1982) Field experiments on dense gas dispersion. J. Hazardous Materials 6 13-41. Raj, P.K. et al, (1979) Experiments involving pool and vapor fires from spills of liquefied natural gas on water, Arthur D. Little, ADA 077073, June. Steward, F.R., (1970) Prediction of the height of turbulent diffusion buoyant flames, Combust. Sci. Technol. 2, 203-212. Thomas, P.H., (1963) The size of flames from natural fires, 9th Int. Combustion Symposium, 844-859. Tucker, R.F., (1988) 35 m LNG pool fire tests at Montoir 1987, 3 volumes, Thornton Research Centre, May 1988. USCG (2005) “Liquified Natural Gas – Ensuring its safe and secure marine transportation,” Proceeding of the Marine Safety & Security Council, Vol. 62, Number 3, the Coast Guard Journal of Safety at Sea, United States Coast Guard, Washington, D.C., Fall 2005.

33

DISTRIBUTION: 1 2 5 2 5 2

MS0899 MS0847 MS1108 MS1108 MS1135 MS1135

Technical Library, 9536 (electronic copy) Steve Attaway, 1534 Mike Hightower, 6332 Juan Torres, 6332 Anay Luketa, 1532 Sheldon Tieszen, 1532

50

Robert Corbin United States Department of Energy Office of Natural Gas and Petroleum Technology 1000 Independence Avenue, SW Washington, DC 20585

34

tories: 9/27/12 Sandia Study Shows Large LNG Fires Hotter but Smaller Than Expected

Sandia Study Shows Large LNG Fires Hotter but Smaller Than Expected (http://energy.sandia.gov/?p=7716)

New research shows that large-scale liquefied natural gas fires are hotter but smaller than anticipated, which means regulators can assume that a slightly smaller area would be at risk in the case of an LNG incident, an official from Sandia National Laboratory said Sunday at a meeting of the National Association of Regulatory Utility Commissioners. However, the study showed that extreme cold from spilling LNG combined with extreme heat from an LNG fire would severely damage a tanker, making it difficult to quickly move the vessel, Sandia’s Michael Hightower, told a gas panel at NARUC’s annual meeting in St. Louis, Missouri. Sandia conducted the study for the US

(http://energy.sandia.gov/?attachment_id=7715) Sandia conducts large-scale experiments that provide data for and verification of high-performance computing simulations concerning the vulnerabilities of our vital infrastructure.

Department of Energy, after a 2007 report by the Government Accountability Office recommended more research on the potential for cascading failure of LNG tanks, Hightower said. Read the full article (http://www.platts.com/RSSFeedDetailedNews/RSSFeed/NaturalGas/6676493) at platts.com. Tagged with: fire behavior (http://energy.sandia.gov/?tag=fire-behavior) • Infrastructure Security (http://energy.sandia.gov/?

tag=infrastructure_security) • large scale expriments (http://energy.sandia.gov/?tag=large-scale-expriments) • liquefied natural gas (http://energy.sandia.gov/?tag=liquefied-natural-gas)

Switch to our mobile site (http://energy.sandia.gov/?p=7716&wpmp_switcher=mobile)

energy .sandia.gov /?p=7716

1/1

SANDIA REPORT SAND2011-3342 Unlimited Release Printed December 2011

LNG Cascading Damage Study Volume I: Fracture Testing Report Robert J. Kalan, Jason P. Petti

Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185 and Livermore, California 94550 Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. Approved for public release; further dissemination unlimited.

Issued by Sandia National Laboratories, operated for the United States Department of Energy by Sandia Corporation. NOTICE:

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government, nor any agency thereof, nor any of their employees, nor any of their contractors, subcontractors, or their employees, make any warranty, express or implied, or assume any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represent that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government, any agency thereof, or any of their contractors or subcontractors. The views and opinions expressed herein do not necessarily state or reflect those of the United States Government, any agency thereof, or any of their contractors. Printed in the United States of America. This report has been reproduced directly from the best available copy. Available to DOE and DOE contractors from U.S. Department of Energy Office of Scientific and Technical Information P.O. Box 62 Oak Ridge, TN 37831 Telephone: (865)576-8401 Facsimile: (865)576-5728 E-Mail: [email protected] Online ordering: http://www.osti.gov/bridge

Available to the public from U.S. Department of Commerce National Technical Information Service 5285 Port Royal Rd Springfield, VA 22161 Telephone: Facsimile: E-Mail: Online order:

(800)553-6847 (703)605-6900 [email protected] http://www.ntis.gov/help/ordermethods.asp?loc=7-4-0#online

ii

SAND2011-3342 Unlimited Release Printed December 2011

LNG Cascading Damage Fracture Testing Report Robert J. Kalan Mechanical Environments, Org. 01534 Sandia National Laboratories P.O. Box 5800 Albuquerque, NM 87185 Jason P. Petti Structural & Thermal Analysis, Org. 06233 Sandia National Laboratories P.O. Box 5800 Albuquerque, NM 87185

Abstract As part of the LNG Cascading Damage Study, a series of structural tests were conducted to investigate the thermal induced fracture of steel plate structures. The thermal stresses were achieved by applying liquid nitrogen (LN2) onto sections of each steel plate. In addition to inducing large thermal stresses, the lowering of the steel temperature simultaneously reduced the fracture toughness. Liquid nitrogen was used as a surrogate for LNG due to safety concerns and since the temperature of LN2 is similar (-190oC) to LNG (-161oC). The use of LN2 ensured that the tests could achieve cryogenic temperatures in the range an actual vessel would encounter during a LNG spill. There were four phases to this test series. Phase I was the initial exploratory stage, which was used to develop the testing process. In the Phase II series of tests, larger plates were used and tested until fracture. The plate sizes ranged from 4 ft square pieces to 6 ft square sections with thicknesses from ¼ inches to ¾ inches. This phase investigated the cooling rates on larger plates and the effect of different notch geometries (stress concentrations used to initiate brittle fracture). Phase II was divided into two sections, Phase II-A and Phase II-B. Phase II-A used standard A36 steel, while Phase II-B used marine grade steels. In Phase III, the test structures were significantly larger, in the range of 12 ft by 12 ft by 3 ft high. These structures were designed with more complex geometries to include features similar to those on LNG vessels. The final test phase, Phase IV, investigated differences in the heat transfer (cooling rates) between LNG and LN2. All of the tests conducted in this study are used in subsequent parts of the LNG Cascading Damage Study, specifically the computational analyses.

iii

CONTENTS FIGURES ....................................................................................................................................... vi TABLES ....................................................................................................................................... xii ACKNOWLEDGMENTS ........................................................................................................... xiii EXECUTIVE SUMMARY ......................................................................................................... xiv 1.

Introduction ........................................................................................................................... 1

2.

Testing Purpose ..................................................................................................................... 1

3.

Testing Overview .................................................................................................................. 2

4.

Fracture Testing .................................................................................................................... 3

4.1

Phase I Exploratory Tests ......................................................................................................3

4.2

Phase II-A Moderate-Scale Fracture Testing – A36 Steel .....................................................9

4.3

Phase II-B Moderate-Scale Fracture Testing – Marine Grade Steels ..................................22

4.4

Phase III Large-Scale Fracture Testing................................................................................30

4.4.1

Large-Scale Tests without Water .....................................................................................32

4.4.2

Large-Scale Tests with Water...........................................................................................45

5.

Heat Transfer Testing ......................................................................................................... 49

6.

Summary and Conclusions ................................................................................................. 61

References ..................................................................................................................................... 62 Appendix A – Stress Concentrations ............................................................................................ 63 Appendix B – Test Data ................................................................................................................ 67

v

FIGURES Figure 1. Phase I test plate and trough ............................................................................................ 4 Figure 2. Thermocouple locations for Phase I test plates ............................................................... 4 Figure 3. Filler tube......................................................................................................................... 5 Figure 4. Thermocouple data from initial test ................................................................................ 5 Figure 5. Spray manifold ................................................................................................................ 6 Figure 6. Thermocouple test after painting ..................................................................................... 6 Figure 7. Layout of the thermocouples ........................................................................................... 7 Figure 8. Temperature data from the thermocouple test plate ........................................................ 8 Figure 9. 48 in. x 48 in. x ¼ in. plate and beam test configuration .............................................. 10 Figure 10. Welded 48 in. x 48 in. x ¼ in. Phase II test plate ........................................................ 10 Figure 11. Location of the thermocouples for 48 in plates ........................................................... 11 Figure 12. Test configuration for the 48 in. x 48 in. Phase II plates ............................................. 12 Figure 13. Phase II Test 1 upper surface temperature profile ....................................................... 14 Figure 14. Phase II Test 1 lower surface temperature profile ....................................................... 14 Figure 15. Drilled hole locations Tests 3 and 4 ............................................................................ 16 Figure 16. Notched holes for Tests 5 and 6, center hole (hole 2) shown...................................... 16 Figure 17. Thermocouple data for Test 5...................................................................................... 17 Figure 18. Crack generated in Test 5, 48 in. x 48 in. x ¼ in. plate with notched holes ................ 17 Figure 19. Crack generated in Test 6, 48 in. x 48in. x ¼ in. welded plate with notched holes .... 18 Figure 20. Phase II test plate 72 in. x 72 in. x ¾ in. ..................................................................... 19 Figure 21. Thermocouple data for Test 7...................................................................................... 19 Figure 22. Surface notches cut into 72 in x 72 in plate for (a) Tests 8, (b) Test 9 and (c) Test 10 ................................................................................................................................................ 20

vi

Figure 23. Thermocouple data for Test 10.................................................................................... 21 Figure 24. Cracked formed in 72 in. x 72 in. plate during Test 10 ............................................... 21 Figure 25. The extension of the crack formed in Test 10 beyond the outside of the cooling trough ................................................................................................................................................ 22 Figure 26. Test layout for the Phase II ABS steel plate tests ........................................................ 23 Figure 27. Trough dimensions and thermocouple layout for Phase II ABS steel plate tests ........ 23 Figure 28. Notches used in Tests 11 and 12 ................................................................................. 25 Figure 29. Crack generated in Test 12 ......................................................................................... 26 Figure 30. Thermocouple data for Test 12.................................................................................. 26 Figure 31. Notch geometry used for ABS Grade steel Phase II tests ........................................... 27 Figure 32. Cracking pattern for Test 19 ........................................................................................ 28 Figure 33. Thermocouple data for Test 19.................................................................................... 28 Figure 34. Thermocouple data for Test 20.................................................................................... 29 Figure 35. Cracking pattern for Test 21 ........................................................................................ 29 Figure 36. Thermocouple data for Test 21.................................................................................... 30 Figure 37. Large Phase III structure layout and materials ............................................................ 31 Figure 38. Phase III Test 13 and 14 trough layout, inside the structure (left) and outside the 3 ft vertical Gr. EH plate (right) ................................................................................................... 33 Figure 39. Trough and thermocouple layout for Phase III Test 13 and 14. .................................. 34 Figure 40. Layout of the Phase III tests (Test 13 and 14) ............................................................. 34 Figure 41. Notch added for use in Test 14 .................................................................................... 36 Figure 42. Trough with added notches and paint repair prior to Test 14...................................... 36 Figure 43. Pipe layouts and Bay/Valve locations for Test 13 and 14 (2 ¾ in. star Test 1 designates in Test 13, 4 in. Cross Test 2 added for Test 14) .................................................. 37 Figure 44. Crack formation during Test 14 .................................................................................. 37 Figure 45. Thermocouple data for Test 14.................................................................................... 38

vii

Figure 46. Trough and thermocouple configuration for Test 16 .................................................. 39 Figure 47. Trough and pipe layout for Test 16 ............................................................................. 39 Figure 48. Trough layout outside vertical Gr. EH plate for Test 16 ............................................. 40 Figure 49. Interior crack formation during Test 16 ...................................................................... 41 Figure 50. Thermocouple data for Test 16.................................................................................... 42 Figure 51. Close-up of likely initiation site for Test 16 ................................................................ 42 Figure 52. Fractures caused by LNG spills on actual LNG vessels (Roue, 2011) ....................... 43 Figure 53. Close-up stiffening rib fracture for Test 16 ................................................................. 44 Figure 54. Outside through-cracks for Test 16 ............................................................................. 44 Figure 55. Outside cracking with turning for Test 16 ................................................................... 45 Figure 56. Schematic of the large structure pool tests .................................................................. 46 Figure 57. Polyurethane foam applied to the bottom of the large structure ................................. 46 Figure 58. Test setup for the large structure pool test................................................................... 47 Figure 59. Test 23 notch increased to 8 in. ................................................................................... 48 Figure 60. Crack progression for Test 23 ..................................................................................... 48 Figure 61. Thermocouple data for Test 23.................................................................................... 49 Figure 62. Plate setup used in heat transfer tests .......................................................................... 49 Figure 63. Heat transfer tests dimensions and thermocouple layout ............................................ 50 Figure 64. LN2 heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Bare ................................ 51 Figure 65. LN2 heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Epoxy ............................. 51 Figure 66. LN2 heat transfer thermocouple tata – 6 in. x 6 in. x ¾ in. Epoxy Urethane .............. 52 Figure 67. LN2 heat transfer thermocouple fata – 6 in. x 6 in. x ¼ in. Bare ................................. 52 Figure 68. LN2 heat transfer thermocouple data – 6 in. x 6 in. x ¼ in. Epoxy ............................. 53 Figure 69. LN2 heat transfer thermocouple data – 18 in. x 18 in. x ¾ in. Epoxy ......................... 53 Figure 70. Schematic of the LNG heat transfer tests .................................................................... 54 viii

Figure 71. LNG heat transfer test configuration ........................................................................... 55 Figure 72. Test plate and flow deflector ....................................................................................... 55 Figure 73. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Bare .............................. 56 Figure 74. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Epoxy............................ 56 Figure 75. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Epoxy Urethane ............ 57 Figure 76. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¼ in. Bare .............................. 57 Figure 77. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¼ in. Epoxy............................ 58 Figure 78. LNG heat transfer thermocouple data – 18 in. x 18 in. x ¾ in. Epoxy........................ 58 Figure 79. LNG heat transfer thermocouple data – 18 in. x 18 in. x ¾ in. Epoxy (zoom) ........... 59 Figure 80. LNG vs. LN2 heat transfer thermocouple data for TC1 – 6 in. x 6 in. x ¾ in. Bare .... 60 Figure 81. LNG vs. LN2 heat transfer thermocouple data for TC1 – 6 in. x 6 in. x ¾ in. Epoxy Urethane ................................................................................................................................. 60 Figure 82. Stress Concentration at Circular and Elliptical holes in Plates ................................... 64 Figure 83. Crack Tip Stresses in Plates ........................................................................................ 66 Figure 84. Test 1 – TCs 1-11 ........................................................................................................ 67 Figure 85. Test 1 – TCs 12-22 ...................................................................................................... 67 Figure 86. Test 2 – TCs 1-11 ........................................................................................................ 68 Figure 87. Test 2 – TCs 12-22 ...................................................................................................... 68 Figure 88. Test 3 – TCs 1-11 ........................................................................................................ 69 Figure 89. Test 3 – TCs 12-22 ...................................................................................................... 69 Figure 90. Test 4 – TCs 1-11 ........................................................................................................ 70 Figure 91. Test 4 – TCs 12-22 ...................................................................................................... 70 Figure 92. Test 5 – TCs 1-11 ........................................................................................................ 71 Figure 93. Test 5 – TCs 12-22 ...................................................................................................... 71 Figure 94. Test 6 – TCs 1-11 ........................................................................................................ 72 Figure 95. Test 6 – TCs 12-22 ...................................................................................................... 72 ix

Figure 96. Test 7 – TCs 1-11 ........................................................................................................ 73 Figure 97. Test 7 – TCs 12-22 ...................................................................................................... 73 Figure 98. Test 8 – TCs 1-11 ........................................................................................................ 74 Figure 99. Test 8 – TCs 12-22 ...................................................................................................... 74 Figure 100. Test 9 – TCs 1-11 ...................................................................................................... 75 Figure 101. Test 9 – TCs 12-22 .................................................................................................... 75 Figure 102. Test 10 – TCs 1-11 .................................................................................................... 76 Figure 103. Test 10 – TCs 12-22 .................................................................................................. 76 Figure 104. Test 11 – TCs 1-11 .................................................................................................... 77 Figure 105. Test 11 – TCs 12-22 .................................................................................................. 77 Figure 106. Test 12 – TCs 1-11 .................................................................................................... 78 Figure 107. Test 12 – TCs 12-22 .................................................................................................. 78 Figure 108. Test 17 – TCs 1-10 .................................................................................................... 79 Figure 109. Test 17 – TCs 11-20 .................................................................................................. 79 Figure 110. Test 18 – TCs 1-10 .................................................................................................... 80 Figure 111. Test 18 – TCs 11-20 .................................................................................................. 80 Figure 112. Test 19 – TCs 1-10 .................................................................................................... 81 Figure 113. Test 19 – TCs 11-20 .................................................................................................. 81 Figure 114. Test 20 – TCs 1-10 .................................................................................................... 82 Figure 115. Test 20 – TCs 11-20 .................................................................................................. 82 Figure 116. Test 21 – TCs 1-10 .................................................................................................... 83 Figure 117. Test 21 – TCs 11-20 .................................................................................................. 83 Figure 118. Test 13 – TCs 1-19 .................................................................................................... 84 Figure 119. Test 13 – TCs 36-54 .................................................................................................. 84 Figure 120. Test 13 – TCs 20-35 (TC 25 not used) ...................................................................... 85 x

Figure 121. Test 14 – TCs 1-19 .................................................................................................... 85 Figure 122. Test 14 – TCs 36-54 .................................................................................................. 86 Figure 123. Test 14 – TCs 20-35 (TC 25 not used) ...................................................................... 86 Figure 124. Test 16 – TCs 1-18 .................................................................................................... 87 Figure 125. Test 16 – TCs 27-44 .................................................................................................. 87 Figure 126. Test 16 – TCs 19-26 .................................................................................................. 88 Figure 127. Test 22 – TCs 1-18 .................................................................................................... 88 Figure 128. Test 22 – TCs 27-44 .................................................................................................. 89 Figure 129. Test 22 – TCs 19-26 .................................................................................................. 89 Figure 130. Test 23 – TCs 1-18 .................................................................................................... 90 Figure 131. Test 23 – TCs 27-44 .................................................................................................. 90 Figure 132. Test 23 – TCs 19-26 .................................................................................................. 91

xi

TABLES Table 1. List of thermocouples and mounting methods .................................................................. 7 Table 2 Phase II-A Testing ........................................................................................................... 13 Table 3. Phase II-B Tests with ABS Grade Steel, all with I-Beams ............................................. 24 Table 4. Large Phase III Structure Tests ....................................................................................... 32

xii

ACKNOWLEDGMENTS The authors would like to acknowledge the following Sandia, DOE, and expert panel members. Mike Hightower, Sandia National Laboratories, Overall LNG Project Manager Luis Abeyta, Sandia National Laboratories, Testing Support and Data Acquisition Amarante Martinez, Sandia National Laboratories, Testing Support and Photometrics Carlos Lopez, Sandia National Laboratories, LNG Cascading Damage Thermal Lead Gerald Wellman, Sandia National Laboratories, Test Planning and Design Frank Dempsey, Sandia National Laboratories, Test Planning and Design Doug Ammerman, Sandia National Laboratories, Internal Sandia Review Robert Corbin, DOE Project Manager Christopher Freitas, DOE Commander Nick Caron, U.S. Coast Guard, Expert Review Panel Charles Rawson, U.S. Coast Guard, Expert Review Panel Professor Stan Rolfe, Kansas University, Expert Review Panel Dr. James Rawers, DOE, Expert Review Panel Dr. John Moorhouse, British Gas (formerly of), Expert Review Panel John Dasch, DNV, Expert Review Panel Chris Zerby, FERC, Expert Review Panel Tony Galt, Freeport LNG, Expert Review Panel

xiii

EXECUTIVE SUMMARY The combination of recent and expected future growth of imports of liquefied natural gas (LNG) with the increased safety and security concerns resulting from the incidents of September 11, 2001, have led to an exploration of possible impacts that an attack on potentially hazardous cargos would have on the public and other assets. A number of studies (Hightower, et al., 2004, Hightower et al., 2006) have been performed at Sandia National Laboratories in the aftermath of September 11, 2001, in order to examine the potential hazards from LNG vessels experiencing an unintended release of LNG cargo. An unintended release of LNG has two significant components: 1) the effects of the cryogenic LNG contacting the steel structure of the tanker vessel, and 2) the flammable nature of LNG and the potential fire hazards to the tanker and to the public if in a harbor location. Stored LNG has a temperature of approximately -161oC which is well below the brittle transition for the marine steels that comprise LNG tankers. Rupture of the LNG containment vessel will results in the LNG flowing out and coming into contact with the tankers steel structures. Due to the complex nature of this type of event, the US Government Accountability Office (GAO) has studied the current state of knowledge of the relevant issues/phenomenon (GAO, 2007). This report was compiled in order to identify the areas of additional research required in order to gain an improved understanding of these events. The top two areas recommended for further study were related to improving the understanding of large LNG fire physics and to improving the state of knowledge surrounding the potential for cascading damage to LNG vessels. Under two separate projects, Sandia National Laboratories is examining both areas of concern. This report summarizes one component, the large-scale fracture testing, performed under LNG Cascading Damage Study. An understanding of the potential for fracture of the steel that comprises the LNG vessels is potentially critical in assessing the likelihood of cascading failure of the vessel. Here, cascading failure is defined as damage that causes the spread of sufficient LNG cargo and/or LNG fires that then lead to additional damage or breaches to the vessel beyond the damage that produced the initial spill. This report summarizes the testing performed to better understand the brittle fracture of steel structures subjected to cryogenic liquids. The types and method of testing performed here for the LNG Cascading Damage Study are directly related to how the testing information will be used in the computational analysis component of the project. The goal of the project is to provide an assessment of LNG vessels subjected to unintended LNG spills. This will be accomplished through a series of computational finite element analysis of the vessels. The models created for these analyses are relatively large, but even so, the smallest elements are approximately 4 inches by 4 inches. While this resolution provides sufficient stress/strain resolution to capture the global behavior of the vessel, this resolution is several orders of magnitude too large to capture the stress/strain fields that are generated at the tip of a crack in the steel that comprises the hulls of LNG tankers. The progression of brittle cracks in steel plates depends on a number of factors including the specific geometry, the material properties, the temperature state caused by contact with cryogenic liquids, and the microstructure of the material. Specifically, the dependence on the microstructure of the steel plating from point to point makes the tracking of individual brittle cracks using continuum mechanics in finite element models impossible. However, this level of detail is not necessary to assess the global behavior of the vessel. What is necessary is an estimate of the general crack path and directionality. The “damage” caused by crack initiation and crack propagation will be represented in the finite element models by the removal or “death” xiv

of 4 inch by 4 inch elements along the crack propagation path. The “death” of a finite element is achieved by removing that element from the analysis after a defined criteria has been reached in that element causing material separation. For these analyses, a strain/temperature locus is employed. Due to the transition from ductile to brittle behavior as the temperature drops, the strain required to “kill” a finite element drops drastically as the temperature falls. The temperature at which the strain drops most dramatically varies from steel to steel; however, all of the marine steels used in LNG tanker construction enter the brittle regime at temperature well above LNG temperatures. The initial strain/temperature locus was developed using basic material tests (stress-strain tests at multiple temperatures). These test results are not presented in this report, but are described in the analysis report when discussing the strain/temperature locus calibration. The fracture tests presented here were used in the secondary calibration and validation of the strain/temperature locus. The goal of the large-scale fracture tests was therefore to provide examples of fractured steel plates and structures. Finite element models were created for the relevant fracture structures to test the ability of the finite element code, element death modeling, and the strain/temperature locus to reproduce the general crack path and directionality. These finite element analyses also used 4 inch by 4 inch elements in order to have the same stress/strain resolution as the models of the full vessel. The analyses of the test structures and the full vessel have the temperatures of the finite elements reduced at an appropriate rate to induce thermal strains. The temperature fields were generated with either a heat transfer analysis or by ramping the temperatures down manually. Since the stresses and strains generated in the 4 inch element in the analysis would not match the resolution of the strain obtained with strain gauges, only temperature data was taken for each test. The temperature data was then used to link the tests with the finite element analyses. The details of the test analyses and the locus calibration are provided in a separate report on the computational analyses (Volume III, Petti, et al., 2011). This report focuses on the fracture testing structures, procedures, and results. As part of the LNG Cascading Damage Study, the series of structural tests conducted here were used to investigate thermally induced fracture of steel plate structures. The thermal stresses were achieved by applying liquid nitrogen (LN2) onto sections of each steel plate causing differential thermal contraction. In addition to inducing large thermal stresses, the lowering of the steel temperature simultaneously reduced the fracture toughness. Liquid nitrogen was used as a surrogate for LNG due to safety concerns and since the temperature of LN2 is slightly colder (190oC) than LNG (-161oC). This ensured that the tests could achieve temperatures in the range an actual vessel would see during a spill. There were four phases to this test series. Phase I was the initial exploratory stage, which was used to develop the testing process. The first several tests studied the cooling of steel plates subjected to LN2. In addition, tests were conducted on various thermocouple types. Finally, to better represent the typical condition of steel, the plates were coated with paints typically used in commercial vessels, painting of the steel surface in contact with the LN2 was shown to significantly increase the cooling rate. Since both the interior and exterior surfaces of LNG vessel hulls have surface coatings, all subsequent test phases used surface coatings. In the Phase II series of tests, larger plates were used and tested until fracture. Phase II was divided into two sections, Phase II-A and Phase II-B. Phase II-A used standard A36 steel, while Phase II-B used marine grade steels. The three test structure in Phase II-A included plate sizes ranging from 4 ft square pieces to 6 ft square sections with thicknesses from ¼ inches to ¾ xv

inches. This phase investigated the cooling rates on larger plates and the effect of different notch geometries (stress concentrations used to initiate brittle fracture). Each of the three Phase II-A test structures were tested multiple times. For each subsequent test, stress concentrations were introduced and then made more severe until a fracture was initiated in the plate. Phase II-B used similar test structures and techniques but with marine grade steels. Specifically, ABS Grade A and ABS Grade EH were used in construction of the test articles. Two Gr. A plate and two Gr. EH plate structures were tested. In Phase III, the three test structures were significantly larger than for Phase II. Each of the structures was built with a main plate spanning 12 ft by 12 ft and then a height of 3 ft. These structures were designed with more complex geometries to include features similar to those on LNG vessels. They included multiple material, intersecting plates, and stiffening elements. However, they were not scaled versions of any specific section of an LNG vessel. The Phase III tests showed that stiffening elements and intersecting plates did not arrest brittle cracks. The cracking that occurred in these test followed the extent of the cooled region and propagated, in general, in the direction perpendicular to the maximum stress. The final test phase, Phase IV, investigated differences in the heat transfer (cooling rates) between LNG and LN2.

xvi

1. Introduction The combination of the expected future growth of imports of liquefied natural gas (LNG) with the increased safety and security concerns resulting from the incidents of September 11, 2001, have led to an exploration of possible impacts that an attack on potentially hazardous cargos would have on the public and other assets. A number of studies (Hightower, et al., 2004, Hightower et al., 2006) have been performed at Sandia National Laboratories in the aftermath of September 11, 2001, in order to examine the potential hazards from a LNG tanker vessel experiencing an unintended release of LNG cargo. An unintended release of LNG has two significant components: 1) the effects of the cryogenic LNG contacting the steel structure of the tanker vessel, and 2) the flammable nature of LNG and the potential fire hazards to the tanker and to the public if in a harbor location. Stored LNG has a temperature of approximately -161oC which is well below the brittle transition for the marine steels that comprise LNG tankers. Due to the complex nature of this type of event, the US Government Accountability Office (GAO) commissioned a study on the current state of knowledge of the relevant issues/phenomenon (GAO, 2007). This previous report was compiled in order to identify the areas of additional research required in order to gain an improved understanding of these events. The top two areas recommended for further study were related to improving the understanding of large LNG fire physics and to improving the state of knowledge surrounding the potential for cascading damage to LNG vessels. Under two separate projects, Sandia National Laboratories is examining both areas of concern. This report summarizes one component, the large-scale fracture testing, performed under the LNG Cascading Damage Study. An understanding of the potential for the fracture of the steel that comprises the LNG tankers is critical in assessing the likelihood of cascading failure of the vessel. Here, cascading failure is defined as damage that causes the spread of sufficient LNG cargo and/or LNG fires that then lead to additional damage or breaches to the vessel beyond that produced during the initial spill. This report summarizes the testing performed to better understand the brittle fracture of steel structures subjected to cryogenic liquids. An explanation of why these specific tests were performed and how the results of these tests are used in computational analyses is provided in the next section. The computational analysis portions of the LNG Cascading Damage Study are contained within additional reports (Figueroa and Lopez, 2011, Petti et al., 2011). 2. Testing Purpose The types and methods of testing performed for the LNG Cascading Damage Study are directly related to the information needed and used in the computational analysis component of the project. The goal of the project is to provide an assessment of LNG vessels susceptibility to damage and possibly failure when subjected to unintended LNG spills. This will be accomplished through a series of computational finite element analysis of models of the two most common commercial LNG vessels. The models created for these analyses are relatively large. However, the analysis will be capable of characterizing the materials response on elements, the smallest of which are approximately 4 inches by 4 inches. While this level of resolution provides sufficient stress/strain analysis to capture the global behavior of the vessel, this scale is several orders of magnitude too large to capture the stress/strain fields that may be generated at the tip of a crack in the steel that comprises the hulls of LNG tankers when subjected to thermal stresses resulting from contact with the LNG. In addition,

1

the progression of brittle cracks in steel plates depends on a number of factors including the specific geometry, the material properties, the temperature state caused by contact with cryogenic liquids, and the microstructure of the material. Specifically, the dependence on the microstructure of the steel plating from point-to-point makes it impossible to predict the individual brittle crack behavior using continuum mechanics in finite element models. However, specific knowledge at the microscopic level of detail is not necessary to assess the global behavior of the vessel. What is necessary is an approximate estimate of the general crack path and directionality. The “damage” caused by cracks will be represented in the finite element models by the failure or “death” of a path of 4 inch by 4 inch elements. The “death” of a finite element is achieved by removing that element from the analysis after a defined criteria has been reached for that element, causing material separation. For these analyses, a strain/temperature locus is employed as the failure criterion. Due to the transition from ductile to brittle behavior as the temperature drops, the strain required to “kill” a finite element drops drastically as the temperature falls. The temperature at which the strain drops most dramatically varies from steel to steel; however, all of the marine steels used in LNG tanker construction enter the brittle regime at temperatures well above LNG temperatures. The initial strain/temperature locus was developed using basic material tests (stress-strain tests at multiple temperatures). These test results are not presented in this report, but are described in the analysis report when discussing the strain/temperature locus calibration. The fracture test results presented here were used in the secondary calibration and validation of the strain/temperature locus. The goal of the large-scale fracture tests was therefore to provide examples of fractured steel plates and structures. Finite element models were created for a selected number of the fractured structures to test the ability of the finite element code, element death modeling, and the strain/temperature locus to reproduce the general crack path and directionality observed experimentally. These finite element analyses also used 4 inch by 4 inch elements in order to reflect the same stress/strain resolution as the models of the full vessels. The analyses of the test structures and the full vessel have the temperatures of the finite elements in the cooled regions reduced at an appropriate rate to induce thermal strains. The temperature fields were generated with either a heat transfer analysis or by ramping element temperatures down manually. Since the stresses and strains generated in the 4 inch element in the analysis would not match the resolution of the strain obtained from experimental strain measurements, the temperature data was then used to link the tests with the finite element analyses. The details of the test analyses and the locus calibration are located in Volume III of this study (Petti et al., 2011). This report focuses on the fracture testing structures, procedures, and results. 3. Testing Overview As part of the LNG Cascading Damage Study, a series of structural fracture tests were conducted to investigate the effect of liquid natural gas (LNG) flowing on steel structures. These tests were designed to induce large thermal stresses in the test plates due to differential thermal contraction while simultaneously lowering their ductility. Liquid nitrogen (LN2) was chosen as the cryogenic fluid to use in place of LNG. It is significantly safer to work with and the temperature of LN2 is slightly colder (-190oC) than LNG (-161oC). The use of LN2 ensures that the tests could achieve temperatures in the range an actual vessel would see during a LNG spill. There were four phases in this test series. Phase I was the initial exploratory stage that provided preliminary data to the material's response to LN2 (LNG) and

2

the insight and direction for the series of larger scale tests. In the Phase II series of tests, larger plates were used than for Phase I. The plate sizes ranged from 4 ft square pieces to 6 ft square sections. This phase investigated the cooling rates on these plates and the effect of different notch geometries (stress concentrations used to initiate brittle fracture) from which the crack would initiate. Phase II was broken in to two sections, Phase II-A and Phase II-B. Phase II-A used standard A36 structural steel, while Phase II-B used marine grade steels. In Phase III, significantly larger 12 ft square structures with more complex geometries were tested. The final test phase, Phase IV, investigated the heat transfer (cooling) rate differences between LNG and LN2. Not including the initial exploratory tests conducted as part of Phase I, twenty two tests were conducted in Phases II and III. Cracking was only achieved in a fraction of the tests. The same test article could then be retested using a different, more severe test conditions. Two additional series of the tests were performed for Phase IV. Each of those series included 6 individual tests. This report presents the details of the testing described above. Due to the large number of thermocouples used and the large number of tests conducted, only a small portion of the temperature data is presented within the main report. The complete set of data can be found in the Appendix B. 4. Fracture Testing 4.1

Phase I Exploratory Tests

Initial tests conducted for the LNG study were exploratory in nature. The tests used small plates to develop the basic test methods and procedures using a low pressure dewar to supply the LN2. These tests explored methods to cool steel plates, flow distribution, surface coatings, and thermocouple types. The initial test used ¾ inch thick carbon steel plates (standard A36), which was sectioned into a 24 in. x 24 in. square as shown in Figure 1. The trough used to hold the LNG was made from polyurethane foam. It is sealed to the plate with silicone caulk and held in place with 1 inch angles and studs. The inner trough dimensions were 2 in. wide and 5 in. high. The thermocouples were Type T 30 gauge Teflon and were held in place using Kapton tape (Type T thermocouples are copper-Constantan). They have an operating temperature range of 250°C to 350°C. The location of the thermocouples is shown in Figure 2. There were ten thermocouples (1-10) on the top surface and ten thermocouples (11-20) directly below them on the bottom surface.

3

Figure 1. Phase I test plate and trough

2.00 in

3/4" Plate

3.50 in

Foam Tro ugh

1

2 3

2.00 in 10

8.00 in (typ) 4.00 in (typ)

9

24.00 in 8

4

2.00 in (typ)

5 6.00 in 6 Located Top and Bottom

4.00 in 7

24.00 in

Figure 2. Thermocouple locations for Phase I test plates The trough was filled with LN2 using a low-pressure dewar and a ½ in. filler tube which was placed centered on top of the trough (above TC 4). The filler tube is shown in Figure 3. The temperature data was recording using a National Instrument SCXI-1000 Data Acquisition Unit (DAQ) and a TC-2095 Thermocouple Connector Module. Temperature data from the test is shown in Figure 4. The data shows a large spread in the temperature of the top thermocouples and slow cooling of the plate through the thickness.

4

Figure 3. Filler tube

Figure 4. Thermocouple data from initial test The differences observed in the temperatures along the top thermocouples led to an attempt to distribute the flow of the nitrogen along the length of the trough more evenly. The nozzle was modified as shown in Figure 5 to include a pipe ½ in. diameter pipe with 16 – 7/64 in. diameter holes. The nozzle had a minimal effect in normalizing the temperature distribution along the length of the trough and was not used in future tests.

5

Figure 5. Spray manifold The next exploratory test was conducted to determine whether the Type T Teflon thermocouples fastened with Kapton tape were the most appropriate thermocouples for determining the plate temperature. A test plate shown in Figure 6 was developed to examine alternate fastening methods along with type E thermocouples. Type E thermocouples leads are made of Nickle-Chromium and Copper-nickel. Type E is recommended for use at temperature above 40K. The Seebeck coefficient for Type E is greater than all other standard thermocouples, which along with its low thermal conductivity make it ideally suited for low temperature applications (ASTM, 1993). A complete list of the thermocouples tested is shown in Table 1. The location of the thermocouples is presented in Figure 7. The thermocouples were mounted 1 in. from the center of the plate in a radial pattern every 45° as shown in Figure 7. The test plate was a 12 in. x 12 in. x ¼ in. thick. A dam was made from a 12 in. x 12 in. x 2 in. thick Styrofoam board with a 4 in. diameter hole in the middle. The Styrofoam was attached to the plate with a bead of room temperature vulcanizing (RTV) sealant around the hole.

Figure 6. Thermocouple test after painting

6

Table 1. List of thermocouples and mounting methods Type

Material Type

Mounting Technique

T E

30 ga. Teflon Insulated 30 ga. Teflon Insulated

Kapton Tape Kapton Tape

T

30 ga. Teflon Insulated

T

0.020 in. Sheathed

E

30 ga. Teflon Insulated

E

0.020 in. Sheathed

T

0.020 in. Sheathed

E

0.020 in. Sheathed

Epoxy embedded in 0.020 in. deep by .040 in. wide slot Epoxy embedded in 0.020 in. deep by .040 in. wide slot Epoxy embedded in 0.020 in. deep by .040 in. wide slot Epoxy embedded in 0.020 in. deep by .040 in. wide slot Capicity Discharge (CD )weld using 0.003 in x 0.250 in. NiCr Strap Capacity Discharge (CD) weld using 0.003 in. x 0.250 in. NiCr Strap

T

30 ga. Teflon Insulated

E

30 ga. Teflon Insulated

Epoxy embedded in 0.040 in. hole hole From back side to within 0.020 in. of Surface Epoxy embedded in 0.040 in. hole from back side to within 0.020 in. of surface

E

30 ga. Teflon Insulated

Intrinsic (CD welded) Front surface

Type TTeflon Kapton Tape Type TSheath Groove Kapton Tape

Type ESheath Welded

Ø4.00 in

Ø2.00 in

Ø2.00 in

Ø4.00 in E Type TSheath Groove Epoxy

T

Type ESheath Groove Epoxy Type EIntrinsic

Type ESheath Groove Kapton Tape

Type TTeflon Kapton Tape

Type TSheath Welded Type ETeflon Kapton Tape

Bottom Surface Type ETeflon w/Epoxy and Type TTeflon w/Epoxy Hole to with 0.020 inches of upper surface

Top Surface

Figure 7. Layout of the thermocouples

7

Two tests were run, one with the plate as received (bare steel surface), and one with the exposed surface painted with a spray-on enamel paint (painted version shown in Figure 6). The intrinsic CD welded Type E thermocouple provided the best measurement of the plate temperature. A critical, though not unexpected, observation from these tests is apparent in temperature data presented in Figure 8. The graph shows the temperature of the intrinsic thermocouple on the top surface and the Kapton taped thermocouple on the back surface. The graph clearly shows that there is a higher cooling rate for the painted plate than for the bare plate. The paint layer acts as an insulation layer that affects the boiling behavior of the cryogenic liquid when in contact with the paint relative to bare steel. This change in behavior due to the paint enhances the contact between the surface and the liquid which results in an increased cooling of the steel plate. Therefore, all subsequent tests were conducted using a single flow nozzles, intrinsic Type E thermocouples with a National Instrument DAQ, and paint applied to the wetted plate surfaces. Since both the interior and exterior hull surfaces of LNG vessels are all painted with epoxy primer and/or polyurethane, painting the surfaces in this testing program is the most appropriate choice.

Painted vs Bare Steel for Intrinsic TCs 50

Temperature (C)

0

-50 Bare Steel Backside Bare Steel Topside -100

Painted Steel Backside Painted Steel Topside

-150

-200 0

100

200

300

400

500

600

Time (sec)

Figure 8. Temperature data from the thermocouple test plate

8

4.2

Phase II-A Moderate-Scale Fracture Testing – A36 Steel

The initial tests on the exploratory 24 in. x 24 in. x ¾ in. plates showed a very slow cooling rate through the thickness of the plate. In addition to the lack of paint in those tests, the slow cooling was believed to be due to the large mass of the plate and the relatively small size of the cooling trough. In order to more quickly cool the plate and to provide addition constraint to generate thermal stresses, the second phase of testing includes of a series of 48 in. x 48 in. x ¼ in. thick plates. These A36 carbon steel plates also had W8x40 I-beams welded around their circumference. The induction of thermal strain is caused by a region of material becoming cold while the surrounding material remains warm. The cold material tries to contract while the warm surrounding material holds the cold material from contracting freely. This constraint provided by the warm surrounding material causes the generation of a large amount of tension within the cold material. As the temperature drops, the tension increases in addition to the continued decrease in the fracture toughness of the steel. Once the stress increases sufficiently to cause a flaw in the steel microstructure to reach a critical stress level, a brittle crack initiates and propagates outward with a velocity near that of the speed of sound. The propagation is based on the structure geometry, the loading on the structure, the distribution of the cold region, and the material microstructure. Thus, the I-beams welded to the outer edges of the plate provide additional constraint to the structure helping to induce larger stresses in the cool regions. A diagram of a 48 in. plate with the support welded I-beams and the trough location is show in Figure 9. A foam trough 12 in. wide by 42 in. long and 5 in. deep (inside dimensions) runs along the center of the plate. The trough was filled with LN2 using a 160 liter low-pressure dewar with a single ½ in. fill pipe locate in the center of the top trough cover. The surface of the plate inside the trough was painted with Krylon spray enamel. As describes earlier, the paint increases the heat transfer between the nitrogen and the plate, resulting in more rapid cooling. The trough was fastened to the plate using silicone caulk and two 1 in. steel angles. The bottom surface of the plate beneath the trough was insulated with Styrofoam (14 in. x 48 in. x 3 in. thick) in order to enhance the heat transfer. The bottom surface foam was also held in place by the steel angles. Two plate configurations were initially tested with 48 in. size plates. The first is just a plain ¼ in. thick plate welded to I-beams around the perimeter. The second plate consists of two half sections (24 in. x 48 in. x ¼ in.) welded to each other along the center. The weld runs through the centerline of the long trough dimension axis. Welding changes the local microstructure and thus the overall material's response to the stress loadings. The purpose of the weld is to provide a region of increased stress due to the residual stress in the weld and to provide flaws for possible crack ignition sites. The plates are also welded to I-beams around the perimeter. The welded test plate is shown in Figure 10. In addition to the 48 in. plates with I-beams, one 72 in. x 72 in. plate was also tested as part of Phase II-A tests. The larger plate used the same size trough and therefore employs the extra surrounding warm plate material to provide the constraint in place of the I-beams.

9

For all tests in Phase II-A, eleven type E intrinsic thermocouples were attached to the top and bottom surface of the plate using a CD welder. The location of the thermocouples is shown in Figure 11. The same test configuration for all of the Phase II-A test is shown in Figure 12.

Figure 9. 48 in. x 48 in. x ¼ in. plate and beam test configuration

Figure 10. Welded 48 in. x 48 in. x ¼ in. Phase II test plate

10

23.00 in

44.00 in 16.00 in

8.00 in 0.50 in 1

2

3

12

13

14

4

5

6

7

8

48.00 in

12.00 in 15 16

17

18

19

1.00 in

8.00 in

20

9 1.00 in

21 10 2.00 in

22 11 Thermocouple Loca tion 48" Pla te Te s t

2.00 in

12/17/08 1 - 11 Top Surfa ce 12 -22 B ottom Surfa ce

48.00 in

Figure 11. Location of the thermocouples for 48 in plates

11

Figure 12. Test configuration for the 48 in. x 48 in. Phase II plates

A list of all of the Phase II-A tests using the three test articles described above is summarized in Table 2. As described, each of the three test articles was tested multiple times. Each subsequent test included the addition or slight modification to a stress concentration. These stress concentration include changing the number of holes, the hole geometry, the shape of the holes, and the addition of notches and varying the notch geometry (shape and depth). These holes/notches were increased in severity from test to test until the hole/notch configuration was sufficient to initiate a brittle fracture. The test numbers in Table 2 represent the order in which the test were conducted. Tests 1, 3, and 5 used the 48 in. plain plate. Tests 2, 4, and 6 used the 48 in. welded plate. Finally, Tests 7 through 10 used the 72 in. plate.

12

Table 2 Phase II-A Testing Test

Note

Material

Plate Size

Notch Type

Test Results

Number 1

A36 Steel

48” x 48” x ¼” thick w/ I-beams

No notch

Did not crack

2

A36 Steel

48” x 48” x ¼” thick w/ I-beams

No notch, weld seam along center of plate

Did not crack

3

Same Plate tested in Test 1

A36 Steel

48” x 48” x ¼” thick w/ I-beams

Three 3/8” holes drilled along centerline

Did not crack

4

Same Plate tested in Test 2

A36 steel

48” x 48” x ¼” thick w/ I-beams

Three 3/8” holes drilled along centerline

Did not crack

5

Same Plate tested in Test 1 & 3

A36 Steel

48” x 48” x ¼” thick w/ I-beams

Three 3/8” holes drilled along centerline with notch

Cracked

6

Same Plate tested in Test 2 & 4

A36 Steel

48” x 48” x ¼” thick w/ I-beams

Three 3/8” holes drilled along centerline with notch

Cracked

A36 Steel

72”x 72” x ¾” thick

No notch

Did not crack

7

8

Same Plate tested in Test 7

A36 Steel

72”x 72” x ¾” thick

Small groove 2 1/8” x ¼” deep cut using circular saw

Did not crack

9

Same Plate tested in Test 7 & 8

A36 Steel

72”x 72” x ¾” thick

Small groove 2 7/8” x 5/8” deep x 5/16” wide cut using die grinder

Did not crack

10

Same Plate tested in Test 7, 8 & 9

A36 Steel

72”x 72” x ¾” thick

4.75” Groove in Test 9 notched with jig saw

Cracked

Each of the Phase II-A tests ran for approximately 10 minutes. The temperature profiles for the Test 1 upper and lower thermocouples are presented in Figure 13 and Figure 14, respectively. With the large trough and a thinner plate, the temperature on the upper and lower surface dropped quickly and in a similar manner. The temperature time histories for Test 2 were similar to Test 1. However, as reported in Table 2, the Test 1 and Test 2 plates did not crack.

13

Figure 13. Phase II Test 1 upper surface temperature profile

Figure 14. Phase II Test 1 lower surface temperature profile

14

Since the plane plate and welded plate did not contain a flaw large enough to initiate a brittle fracture given the stress state induced by the cooling applied, stress concentrations were introduced. This was accomplished by drilling three 3/8 in. diameter holes spaced along the long dimension of the trough. The locations of the holes are presented in Figure 15. The holes were sealed with cork material and silicone caulk in order to prevent LN2 from leaking through the plates. This cork and silicone material adds relatively no strength back to the steel plate. The holes act as a stress concentration since the material directly adjacent to the holes will theoretically experience a stress three times that of the test without holes. If the material in the stress concentrated zone contains a microstructural flaw large enough, and the stress level is sufficiently high, and the material is in the brittle failure regime, brittle fracture will occur. The Appendix at the end of this report provides additional discussion on stress concentrations and fracture mechanics. These re-tests (Test 3 and 4) were run for approximately 5 minutes, and again, the plates did not crack. The cooling rates for Tests 3 and 4 were similar to Tests 1 and 2. In order to further increase the stress, notches were introduced to the sides of the 3/8 in. holes using a jigsaw. The notches were between 3/4 in. and 7/8 in. long (total length from notch tip to notch tip). Each notch had a width of approximately 1/16 in. The introduction of the notches increased the stress significantly more than the three times achieved by the holes alone. The center holes for Test 5 and Test 6 are shown in Figure 16. These severe notches did lead to fractures initiating approximately 1.5 minutes in to the cooling of each plate. The temperature data from Test 5 is given in Figure 17. The cooling rates for Test 5 were very close to the previous tests (Tests 1 through 4). The cooling rates for Test 6 were also similar to that of Test 5 and the previous tests. The cracks run in Figure 18 and Figure 19 from the edge of the center hole (hole 2 in each plate, and the location of maximum stress) perpendicular to the long dimension of the trough (this happens to also be perpendicular to the direction of maximum stress). The crack was arrested by the material just under the foam trough where the plate temperature increased sharply since it was not cooled directly during the test. As shown in Figure 19, there was a bifurcation of the crack in the plate with the center weld. This type of bifurcation is typical in brittle fracturing of steel plates. As explained earlier, the goal of this testing program is to explore the propagation of brittle cracks in steel plate structures, and not to determine the exact conditions required for crack initiation. LNG vessels are large complex structures that have sufficient stress concentration to initiate a brittle cracking when subjected to thermal stresses due to contact with LNG. These tests are being used to provide example fractured structures to calibrate/validate a computation failure model as discussed in Section 2. The tests performed on the two 48 in. plates do not appear to differ significantly. The introduction of the weld does not appear to have affected the test outcome. However, this does not in any way demonstrate that welds on LNG vessels will not contain likely sites for flaws and stress concentrations. The welded plate test was one single test on a simple, nonfatigued, geometry. As later tests in this report will demonstrate, combining welds with more complex geometries increases the likelihood of inherent flaws falling within any spill region.

15

44.00 in

0.375 (Three Plac es )

48.00 in

12.00 in

9.00 in

9.00 in 1.00 in

1.00 in

Drille d Hole Size and Loca tion Pla in Plate A s s embly 1/15/09 B ob Ka la n

48.00 in

Figure 15. Drilled hole locations Tests 3 and 4

Figure 16. Notched holes for Tests 5 and 6, center hole (hole 2) shown.

16

50

0

tc 2 tc 3 tc 4

-50 Temperature (C)

tc 5 tc 6

-100

tc 7 tc 13 tc 14

-150

tc 15 tc 16 -200

tc 17 tc 18

-250 0

20

40

60

80

100 Time (sec)

120

140

160

180

200

Figure 17. Thermocouple data for Test 5

Figure 18. Crack generated in Test 5, 48 in. x 48 in. x ¼ in. plate with notched holes

17

Figure 19. Crack generated in Test 6, 48 in. x 48in. x ¼ in. welded plate with notched holes

After testing the 48 in. plates, the A36 carbon steel plate with slightly different dimensions, 72 in. x 72 in. x ¾ in. thick, was tested. The thickness of ¾ in. is much closer to the average hull thickness used in most LNG vessels. The plate and trough are shown in Figure 20. The trough dimensions are the same as those used on the 48 in. x 48 in. plates. In place of the steel beams welded around the perimeter of the plate, the extra material and thickness of the plate provided the warm constraining material. As noted in Table 2, four successive tests were conducted on this plate in order to generate a crack. The plate was first tested with no machining to the plate. Part of this experiment was to investigate the cooling of a thick (¾ in.) plate compared to the 48 in. x 48 in. plates which were only ¼ in. thick. The thermocouple locations for these tests are the same as those for the 48 in. x 48 in. plates. The temperature distribution for the plate (Test 7) is shown in Figure 21. The ¾ in. thick plate used for Test 7 cooled significantly slower than the ¼ in. thick plates in Tests 1 through 6. In addition, the lack of I-beams welded around the perimeter did not prevent the out-of-plane buckling deformation of the plate. The 72 in. plate deformed in a warping, or saddle shape. There was approximately a 1 1/8 in. warping of the plate along the long dimension axis of the trough and approximate 7/8 in. along the short dimension axis of the trough.

18

Figure 20. Phase II test plate 72 in. x 72 in. x ¾ in. 50

0

Temperature (C)

tc 2 tc 3

-50

tc 4

tc 5 -100

tc 6 tc 7 tc 13

-150

tc 14 -200 0

100

200

300

400

500

Time (sec)

Figure 21. Thermocouple data for Test 7

19

600

After the initial test (Test 7) with the 72 in. plate, several additional tests were conducted to study different field techniques for introducing stress concentrations for future tests. All stress concentrations were centered in the trough area with orientations parallel to the short trough dimension. The first attempt introduced a notch cut into the plate using a die grinder. The notch for Test 8 was approximately 2 1/8 in. long by ¼ in. deep at the center as shown in Figure 22(a). This configuration was tested with no crack initiation. Using the die grinder, the notch size was increased to 2 7/8 in. long by 5/8 in. deep by 5/16 in. wide for Test 9 as shown in Figure 22(b). However, the slightly longer and deeper surface notch did not lead to crack initiation for Test 9. From Tests 8 and 9, it was concluded that use of the die grinder to introduce shallow surface notches did not produce sharp enough regions to produce the stress required for fracture. For Test 10, the notch length was increased to 4 inches using the die grinder in addition to cutting completely through the thickness of the plate. In addition, two ¼ in. long thin notches were cut into the ends of the larger die ground notch using a jigsaw. The resulting notch is shown in Figure 22(c). The cooling rates for Test 10 are illustrated in Figure 23. This test resulted in a crack initiated and propagating perpendicular to the long trough dimension as shown in Figure 24. The fracture occurred approximately 4.5 minutes into the test. The crack extended approximately 1.5 inches beyond the outside surface of the trough. This is illustrated in Figure 25. As with Tests 5 and 6, the crack arrested when entering the steep thermal gradient that transitions into warm material. Due to the longer cooling time to fracture, the cooling, and therefore the crack, extended slightly beyond the trough.

Figure 22. Surface notches cut into 72 in x 72 in plate for (a) Tests 8, (b) Test 9 and (c) Test 10

20

50

0

Temperature (C)

tc 2 -50

tc 3 tc 4 tc 5

-100

tc 6

tc 7 tc 13 -150

tc 14

-200 0

100

200

300

400

500

600

Time (sec)

Figure 23. Thermocouple data for Test 10

Figure 24. Cracked formed in 72 in. x 72 in. plate during Test 10

21

Figure 25. The extension of the crack formed in Test 10 beyond the outside of the cooling trough

4.3

Phase II-B Moderate-Scale Fracture Testing – Marine Grade Steels

Following the initial Phase II-A tests, the Phase II-B series of seven tests were conducted on 4 different test articles using same dimensions and I-beam general configuration as employed in Tests 1 through 6 (48 in. x 48 in.). However, the plate thickness was increased to ¾ in. in addition to the use of marine grade steels. ABS Grade A and ABS Grade EH steels were the two steels chosen for the Phase II-B testing. The ABS Gr. A steel has the lowest stress-strain curve and lowest fracture toughness of the marine grades, while ABS Gr. EH has the highest stress-strain curve and the highest fracture toughness. Therefore, testing of these two materials bounds the different marine grade steels. As with Tests 1 through 6, the test plates were constructed with the W8x40 steel I-beams welded around the perimeter. One test used a 40 in. x 12 in. trough, ABS Gr. A steel, and the same thermal couple layout as in previous tests and shown in Figure 11. The remaining three test articles (one Gr. A and two Gr. EH) employed more significantly modified trough dimensions (24 in. wide by 30 in. long) and a new thermocouple layout. These new trough dimensions cause a significant change to the stress field generated during cooling. The trough dimensions and thermocouple layout are shown in Figure 26 and Figure 27. The wetted area of the trough for this test series was painted using Blue Water Marine AC 70 primer and Marine Urethane. The Phase II-B tests are listed in Table 3. The initial two tests (Test 11 and 12) were conducted using the same Gr. A test plate with a 40 in. x 12 in. trough. A second Gr. A plate with the 30 in. x 24 in. trough was used in Tests 18 and 19. Finally, two identical Gr EH test plates (both with 30 in. x 24 in. troughs) were constructed for Tests 17 and 20 and Test 21. Tests 13 through 16 identify tests in Phase III and Phase IV conducted in parallel with Phase II-B.

22

Figure 26. Test layout for the Phase II ABS steel plate tests

Figure 27. Trough dimensions and thermocouple layout for Phase II ABS steel plate tests

23

Table 3. Phase II-B Tests with ABS Grade Steel, all with I-Beams Test

Note

Material

Plate Size

Notch Type

Test Results

48” x 48”x ¾”

3/8” hole with 1” star slot

Did not crack

3/8” hole with 2¾” star slot

Plate cracked

3/8” hole with 2 ½” slot

Did not crack

3/8” hole with 2 ½” slot

Did not crack

3/8” hole with 3 7/8” slot

Plate cracked

3/8” hole with 3 7/8” slot

Did not crack

3/8” hole with 3 7/8” slot

Plate cracked

Number 11

Grade A

40” x 12” trough 12

Same Plate as 11

Grade A

48” x 48”x ¾” 40”x 12” trough

17

Grade EH

48” x 48”x ¾” 30” x 24” trough

18

Grade A

48” x 48”x ¾” 30” x 24” trough

19

Same Plate as 18

Grade A

48” x 48” x ¾” 30” x 24” trough

20

Same Plate as 17

Grade EH

48” x 48”x ¾” 30” x 24” trough

21

Second EH plate

Grade EH

48” x 48”x ¾” 30” x 24” trough

24

The plate used in Test 11 was tested using Gr. A steel and a stress concentration machined into the center of the trough. The stress concentration was started with a 3/8 in. diameter hole and then extended with 8 notches cut in a 1 in. diameter “star” pattern as shown in Figure 28(a). The 40 in. x 12 in. trough dimensions used for Test 11 is very similar to the dimensions used in the Phase II-A tests. This long but narrow trough severely biases the tensile stress fields within the cooled region of the plate. The stress generated in the direction of the long trough dimension is approximated twice the stress generated in the short dimension of the trough. This would theoretically lead to initiations and crack propagation in direction of the short trough dimension. The star pattern was introduced to determine whether or not a crack would initiate in the direction of the long dimension of the trough in addition to the short dimension, or potentially along the diagonal. The first test (Test 11) on this plate did not generate any cracking. Therefore, the notch length was extended to 2 ¾” in only two directions forming a “cross” shape as shown in Figure 28(b). The test with this configuration resulted in a crack (Test 12) across the trough short dimension as shown in Figure 29 approximately 4 minutes and 45 seconds into the test, but no cracking was generated in the long dimension of the trough. The temperature time histories for Test 12 are illustrated in Figure 30.

Figure 28. Notches used in Tests 11 and 12

25

Figure 29. Crack generated in Test 12

50

0

Temperature (C)

TC2 TC3

-50

TC4

TC5 -100

TC6 TC7 TC15

-150

TC16 -200 0

50

100

150

200

250

300

350

Time (sec)

Figure 30. Thermocouple data for Test 12

26

400

Tests 17 through 21 were conducted using three tests articles, one Gr. A plate and two Gr. EH plates. Each was tested with a 30 in. x 24 in. trough. Initially, Tests 17 and 18 used 2 ½ in. long notches parallel to the short tough dimension with Gr. EH and Gr. A plates, respectively. Neither of the plates fractured. The notch lengths were increased to 3 7/8 in. as shown in Figure 31. In addition, a second Gr. EH plate was added with the same notch length and is labeled Gr. EH 2 in Figure 31. The Gr. A plate cracked (Test 19) approximately 8 minutes into the test with the resulting crack shown in Figure 32. Note that there were two cracks generated immediately from the one side of the notch. The use of the 30 in. x 24 in. trough dimensions reduces the level of domination of the maximum stress field and enables a higher likelihood of diagonal crack propagation as seen here. The thermocouple data from Test 19 is illustrated in Figure 33. The Gr. EH plate used in Test 17 was retested for Test 20 using the longer notch but no fracture initiated. The thermocouple data for Test 20 is illustrated in Figure 34. A nearly identical Gr. EH plate was tested (Test 21) with a crack initiating approximately 8 minutes into the test and is shown in Figure 35. A second crack propagation was observed in Test 21 approximately 1 minute after the initial cracking due to a slight extension of the first cracks as the cooling in the plate extended outward. The thermocouple data from Test 21 is illustrated in Figure 36. The Gr. EH plate used in Test 20 was not tested again. Since the two Gr. EH test articles were nearly identical, the conclusion was reached that the EH plate that did not fail was extremely close to failing. Slight differences in the material, the construction, and the notch shape could all have contributed to one failing and the other not failing. This very close threshold of failure was taken into account in the computational strain/temperature failure locus calibration process and is described in the computational analysis report in more detail.

Figure 31. Notch geometry used for ABS Grade steel Phase II tests

27

Figure 32. Cracking pattern for Test 19

Figure 33. Thermocouple data for Test 19

28

Figure 34. Thermocouple data for Test 20

cracks

Figure 35. Cracking pattern for Test 21

29

Figure 36. Thermocouple data for Test 21

4.4

Phase III Large-Scale Fracture Testing

In parallel with the construction of the Phase II-B test articles, the Phase III test structures were also assembled. For Phase III of the LNG fracture testing, a larger structure was designed as shown in Figure 37. The structure consists of a combination of ABS Gr. A and EH steel plates. The structure was designed with some of the general features found within a LNG vessel (e.g., intersecting plates, welds, stiffening ribs, etc). The Phase III test structure was not designed as scaled versions of a section of any existing LNG vessel, but rather to study the behavior and general features found in LNG vessels (e.g., intersecting plates, welds, stiffening elements, etc). Referring to Figure 37, the lower plate (blue) is ¾ in. ABS Gr A steel. The vertical walls are ½ in. ABS Gr. EH steel and extend three feet from the surface of the ¾ in. thick Gr. A plate. Three ¼ in. thick by 4 in. high ABS Gr. A stiffening ribs are placed at 2 ft spacings. The ribs were welded perpendicular to the ¾ in. thick Gr. A base plate. Finally, structural W8X40 I-beams were welded to the edges of the plate to add additional constraint. Structural W8X40 I-beams were added (welded) to the bottom of the structure to act as supports. These supports also allowed access to the back surface of the lower plate for attaching thermocouples and insulation material. Three of these large structures were constructed for Phase III testing. A list of the five Phase III tests using those three large test articles is given in Table 4. The first two test structures (Tests 13 and 14 and Test 16) were cooled using multiple dewars of LN2. This was necessary due to the increased size of the structures and the surface area being cooled. The third large structure (Tests 22

30

and 23) was identical to the first two, but the test was conducted with water in contact with some of the steel structure. This test was of interest due to the scenarios in actual LNG vessels were a portion of the steel hull may have LNG on one side and have water on the adjacent side. The presence of the water may affect the cooling of the steel plates. The details of all three tests are provided below. Note that Tests 17 through 21 were part of Phase II-B. In addition, Test 15 was part of Phase IV.

Figure 37. Large Phase III structure layout and materials

31

Table 4. Large Phase III Structure Tests Test Number 13

Note 5 valves

Steel

Notch Type

Test Result

See Figure 37

2 ¾” star notch in Bay A

Did not crack

See Figure 37

New Notch 4” cross in Frame B

Cracked

See Figure 37

T-slot 2 ¾”

Cracked

See Figure 37

2 ¾”notch

Did not crack

See Figure 37

8” notch

Cracked (after turning off dewars)

Dewars open same time 14

Same structure used in Test 13 Dewars staged

16

6 values Second structure tested

22

6 values First Pool Test

23

Same structure as Test 22 Second Pool Test

4.4.1 Large-Scale Tests without Water The tests conducted (Test 13 and 14) on the first Phase III structure used a 30 inch wide trough constructed perpendicular to the 4 in. stiffening ribs. The layout of the trough is shown in Figure 39. The trough is separated into two sections. The first section is within the 3 ft vertical Gr. EH plates and is approximately 74 in. long and spans three rib compartments within the structure. The other trough section is approximately 24 in. long and is located

32

outside of the vertical Gr. EH plate. Photos of the trough are shown in Figure 38. The trough also extends up the first 18 in. of the inside and outside surface of the 3 ft vertical Gr. EH plate. Fifty-four thermocouples were used during the tests to record temperature data. The layout of the thermocouples is presented in Figure 39. Five low-pressure dewars were used in the two tests (Test 13 and 14) performed on the first large structure. The dewars were connected to five Magnatrol F25M21 solenoid valves to enable remote operation. The layout of the dewar pipes is also shown in Figure 39. One vertical pipe flows into the center of the three inside chambers created by the 4 in. ribs and the foam trough. The other two pipes spray the inside and outside surface of the 3 ft vertical Gr. EH side wall. The flow strikes the 3 ft wall approximately 1 ft above the Gr. A base plate. An eight segment star notch which measure 2 ¾ in. in diameter was machined into the third chamber from the wall (see dark gray spot in Figure 38(a)). This configuration of notches was machined into the structure after the initial painting of the trough area. Therefore, the dark gray spot shows the repainting performed in the notch area. A 4 ft x 8 ft x 2 in. piece of Styrofoam was placed on the underside of the ¾ in. Gr. A plate trough region to act as additional insulation. The layout of the test structure for Tests 13 and 14 is illustrated in Figure 40. During Tests 13 all five solenoid valves were open simultaneously. The test ran for approximately 30 minutes and no fracturing was generated.

Figure 38. Phase III Test 13 and 14 trough layout, inside the structure (left) and outside the 3 ft vertical Gr. EH plate (right)

33

Th ermcou ple and Test Lay out for First Phase III Test

TC on both side of floor (top and Bottomnumberdifferby 35)

N otc h

TC on Both side of vertical gusset orwall D e w ar Pipe Loc a tion

Loc a tion of TC 's on R ibs

12.00 in

17

29 35

32 34 31

33 30 3.00 in

1.00 in

27 26

30.00 in

11

10 A

4.00 in

16

A 28

10.00 in

18

10.00 in

14.00 in

15

19

4.00 in 16.00 in

8.00 in

Dewar Flow on bottom

Dewa r Flow on Wall

Dewar Flow on Wall

9

14 13

12.00 in 6.00 in 26.00 in 20.00 in

12 8

7

6

5 4 3 2 1 1.0 in

V iew A- A Th ermo co u p les o n o n e sid e o f w all

23/24

20/21 22/23

38 16 54 Total

Foa m Trough

The rmoc ouple s on e ac h one s ide of gus s e t

Figure 39. Trough and thermocouple layout for Phase III Test 13 and 14.

Figure 40. Layout of the Phase III tests (Test 13 and 14)

34

As with many of the Phase II tests, a second Phase III test (Test 14) was performed using the same structure that was used in the first test (Test 13). However, the second test on this structure included a redesign of the notch, notch location, and dewar timing. For Test 14, a 4 in. long cross shaped notch was machined into the middle chamber of the trough as shown in Figure 41 and Figure 42 (brown tape covering cross notches). The dark grey areas in Figure 42 show the repainting performed in areas that experienced paint flaking after Test 13. The original star notches in the far bay remained unrepaired (see Figure 43). Longer notches lead to higher stress concentrations and therefore failure was predicted to occur at the 4 in. cross notches and not to original star notches. Also, timing of the 5 dewar values was introduced so that the entire trough region was not cooled simultaneously. During Test 14, the valves were opened at different times. Referring to Figure 43, valve 2 (Bay B) was opened initially. After 10 minutes and 55 seconds, a crack initiated (audible evidence). Within approximated 15 seconds from this audible sound, valve 3 (Bay C) was opened. At a time of 12 minutes and 23 seconds into the test, a second crack initiation/propagation was heard. Again, within approximately 15 seconds from this sound, valves 1, 4, and 5 were opened. At 15 minutes and 45 seconds into the test, a third crack initiation/propagation occurred. As the test continued, significant quantities of leaking LN2 fluid was observed passing through the cracked structure. Finally at 16 minutes and 26 seconds into the test, the final crack initiation/propagation occurred. The valves were close 18 minutes and 35 seconds into the test. The fractures that formed during Test 14 are shown in Figure 44 with the thermocouple data illustrated in Figure 45. Note that the thermocouple data for TC7 increases rapidly around 650 seconds. This was due to the thermocouple losing contact with the plate and/or malfunctioning. The goal of the timed value openings was to learn how it affected the cracks propagated after initiation. The initial region, Bay B, has dimensions 30 in. wide and 24 in. long in the direction of the full trough. As also observed during the Phase II tests, the crack initially propagated in the direction of the short dimension (24 in.), which is the direction of the other sections of the trough (see Figure 44). The second fracture most likely included cracks propagating into Bay C after value 3 was opened. These cracks immediately turned toward the sides of the trough. Note that after the second valve was turned on, the cooling region dimensions were now 48 in. x 30 in. The cooled region now includes Bay C and Bay B, thus the geometry of the cooled region now has the long axis rotated 90 degrees. This rotation of the long dimension therefore rotated the maximum stress direction causing the cracks to turn during propagation. The third observed cracking was most likely the cracks propagating into Bay A and turning toward the sides of the trough after the remaining valves were opened. The final observed crack propagation was likely the extension of multiple existing cracks into material located slightly beyond the borders of the trough. This was observed in the Phase II tests when the cooling continues for some time. Here, the material outside of the direct trough cooling region also begins to cool to the point where crack extension is favorable. Also note that the stiffening ribs were fractured through completely. Since they were nearly as cold as the base plate, they did not provide any type of crack arrestment.

35

Figure 41. Notch added for use in Test 14

Figure 42. Trough with added notches and paint repair prior to Test 14

36

3' wall 3' wall

Trough 3

1

2 B

C

A

D 2 3/4" star Test 1

4" Cross Test 2 5

4 Rib

Rib

Rib

Figure 43. Pipe layouts and Bay/Valve locations for Test 13 and 14 (2 ¾ in. star Test 1 designates in Test 13, 4 in. Cross Test 2 added for Test 14)

Figure 44. Crack formation during Test 14

37

50

tc 2 0

tc 5 tc 6

Temperature (C)

tc 7 -50

tc 8 tc 9 tc 10

-100

tc 37 tc 40

tc 41 -150

tc 42 tc 43 tc 44

-200

tc 45 0

200

400

600

800

1000

1200

Time (sec)

Figure 45. Thermocouple data for Test 14 The next Phase III test (Test 16) used the trough configuration shown in Figure 46, Figure 47, and Figure 48. The trough extended the full width (8 ft.) of the region within the vertical Gr. EH plates and spanned the other direction across two rib sections. The trough ran up the side of the 3 ft vertical Gr. EH wall approximately 18 inches. For the outer region of the structure, the trough extended only approximately 4 in. along the entire 8 ft. span, but also extended up the vertical Gr. EH plate 18 inches along the entire 8 ft. span. There were filler tubes attached to six low-pressure dewars for Test 16. Four of the tubes filled the inner trough regions. The two remaining tubes sprayed the inside and outside surface of the 3 ft. side wall. Similar to Tests 13 and 14, the tubes spray the side wall approximately 12 in. above the ¾ in. thick Gr. A base plate and centered along the 8 ft. span. The five filler tubes that supplied the inner regions of the structure are clearly show in Figure 47. The wetted surfaces were painted with marine primer that was used in the Phase II-B tests. Forty–four thermocouples were attached to the structure. The thermocouple layout is presented in Figure 46.

38

Thermcouple and Test Layout for second Phase III Test TC on both sides of floor 8 34

TC on side of the vertical guss et or wall

13

3 29

(black numbers top surface, green numbers bottom surface)

9

4.00 in

D e w a r P ipe Loca tion

Loca tion of TC 's on Ribs

39

35

30.00 in 2.00 in 12.00 in

30 4

6.00 in

A

10 36 1 27

54.00 in

19 5 31

20

A

14

16

40

42

12.0 in 15.00 in 6.00 in

21 24

2 6

32

11 37

28

15

17

41

43

22

23

25

26

2.00 in

View A-A Thermocouples on one side of wall 36

18 7 33

12 6.00 in

8

38 44

44 Total Foam Trough

Figure 46. Trough and thermocouple configuration for Test 16

Figure 47. Trough and pipe layout for Test 16

39

(Blue numbers )

Figure 48. Trough layout outside vertical Gr. EH plate for Test 16

The stress concentration machined into the structure for Test 16 was started with a 3/8 in. diameter hole with notches then extended away from the hole creating a total notch length of 2 ¾ in. The notch was perpendicular to the ribs and was in the second rib chamber as shown in Figure 49. Due the large number of cracks, red dye penetrate was used to highlight the fracture progression. Again, the long dimension of the cooling region for this test was 96 in. (8 ft.) with the short dimension 48 in. (~52 in. if the section outside of the vertical Gr. EH plates is added). This configuration would lead to a crack initiation and propagation mainly in the short cooling dimension; therefore, the notch was machined only in that direction. All six of the valves were open simultaneously at the beginning of the test. At just over 14 minutes into the test, a crack initiated (audible detection). The resulting cracks are shown in Figure 49 and with the thermocouple data provided in Figure 50. The uniform drop in all of the temperature data at about 850 seconds was caused by the vibration that occurred during the fracture. The data after that point is not considered valid. Examination of the crack pattern suggested there is a high likelihood that the initial crack initiation occurred at the weld “toe” attaching one of the stiffening ribs to the ¾ in. thick Gr. A plate. This rib weld is located adjacent to the “rat hole” near the vertical Gr. EH plate as illustrated in Figure 51. This spot is the intersection of three different plate and weldments resulting in a highly stressed geometry. The resulting stress shock wave propagated through the structure to the machined notches. This was then followed by the subsequent initiation of two additional sites in the weld region between the ¾ in. Gr. A base plate and the vertical Gr. EH plate (upper left of Figure 49). This sequence of initiations is based on a study of the crack propagation and termination pattern. Since only one audible cracking event was observed, all of the initiations occurred nearly instantaneously. It is also clear from Figure 49

40

that the cracks, for the most part, run in the direction of the short dimension of the cooling region. However, with largest cooling region in terms of area of any test performed in this study, Test 16 exhibited many of the features observed in actual cases of large-scale brittle fracture in steel plating as shown in Figure 52. These include crack branching and semirandom propagation. This semi-random nature is based on the stress wave and the microstructure of the material. Though as mentioned earlier, the direction of the gross path of the cracks was dominated by the maximum stress direction which propagated them in the short cooling dimension. A significant outcome of the test was the propagation of the cracks not only through the stiffening ribs (Figure 53), but also through and up the vertical Gr. EH plate (Figure 54). The larger vertical wall with the higher fracture toughness Gr. EH plate had no effect in arresting a propagating brittle crack since the 3 ft. Gr. EH plate was also cold. Finally, the cracks that propagated through the vertical Gr. EH plate and into the horizontal ¾ in. Gr. A plate turned 90 degrees and joined together to form a crack parallel to the vertical Gr. EH plate. Part of this turning and joining is illustrated in Figure 55. The reason for the crack turning lies in the rotation of the maximum stress direction along the narrow cooled region just outside of the vertical Gr. EH plate. It should also be noted that this crack ran through the base Gr. A plate and not along the weldment joining the Gr. A plate and the Gr. EH plate.

Figure 49. Interior crack formation during Test 16

41

50

0

Temperature (C)

tc 4

-50

tc 5 tc 10 tc 14

-100

tc 30 tc 31 tc 36

-150

tc 40

-200 0

100

200

300

400

500

600

700

800

900

Time (sec)

Figure 50. Thermocouple data for Test 16

Figure 51. Close-up of likely initiation site for Test 16

42

1000

Figure 52. Fractures caused by LNG spills on actual LNG vessels (Roue, 2011)

43

Figure 53. Close-up stiffening rib fracture for Test 16

Figure 54. Outside through-cracks for Test 16

44

Figure 55. Outside cracking with turning for Test 16 4.4.2 Large-Scale Tests with Water The last set of tests (Test 22 and 23) conducted using the third and final large structure placed the structure inside a 24 ft diameter pool. The purpose of the tests was to investigate cooling a structure that was in partial contact with water. A schematic of the test configuration is shown in Figure 56. Only a 48 in. wide plate section was in contact with the water as pointed to in Figure 56. Polyurethane foam was used to insulate the remaining portion of the structure from the pool water. The bottom surface was sprayed until forming a 3 in. foam layer as shown in Figure 57. The wetted surfaces were painted with marine paint. For Test 22, the trough layout, notch dimensions, and the dewar layout were the same as for Test 16. A 2 3/4 in slot was cut perpendicular to the ribs in the second chamber. The backside of this area was insulated from the water by the urethane foam. As with Test 16, six dewars were used during the test and all of the dewar valves were open at the start of the test. The pretest setup of the large structure pool test is shown in Figure 58. Several small holes were drilled into the horizontal plate. Plastic tubes were then attached to the top surface of the holes so that water could be seen rising above the surface of the horizontal plate. These holes/tubes were used to attempt to ensure that water was in contact with the bottom surface of the ¾ in. thick Gr. A plate. The only differences from Test 16 were the water and the Polyurethane foam preventing water from contacting a portion of the ¾ in. Gr. A base plate.

45

Figure 56. Schematic of the large structure pool tests

Figure 57. Polyurethane foam applied to the bottom of the large structure

46

Figure 58. Test setup for the large structure pool test Test 22 ran for approximately 30 minutes (exhausted the supply of LN2) without cracking the structure. The notch was subsequently enlarged to 8 in. as shown in Figure 59. Using an identical test setup as for Test 22, the second pool test, Test 23, was run for thirty minute and again no cracking was observed. The valves were closed at 30 minutes, the instrumentation was turned off around 42 minutes, and the disassembly of the test setup was then initiated. However, draining of the pool had not been initiated. Approximately 46 minutes after beginning the test (4 minutes after turning off the data acquisition), a crack initiated. The resulting cracks are shown in Figure 60. The crack initiated from the 8 in. long notches and then formed an “X” pattern. The cracks propagated through and up the vertical Gr. EH plate. In addition, the long cooling time caused significant cooling outside of the direct trough region. This led to a propagation of crack well outside of the trough region (right side of Figure 60). Since the crack initiated at a time when no physical contact with the structure was occurring by the test administrators, the most likely reason for crack initiation is that the structure warmed up through a different path than the structure cooled down through. This slightly different path of warming caused the stresses nearest the tips of the machined notches to increase to a level higher than that experienced earlier in the test. Since the data acquisition system was not collecting data, the exact temperature distribution in the structure at crack initiation is not known. However, Figure 61 illustrated the thermocouple data up to approximately 4 minute before fracture. The temperatures are clearly starting to increase at 42 minutes, but the structure was still extremely cold. The warming trend would have continued for the additional 4 minutes until fracture, but not significantly.

47

Figure 59. Test 23 notch increased to 8 in.

Figure 60. Crack progression for Test 23

48

50

0

Temperature (C)

tc 4 -50

tc 5 tc 10 tc 14

-100

tc 30 tc 31 tc 36

-150

tc 40

-200 0

500

1000

1500

2000

2500

3000

Time (sec)

Figure 61. Thermocouple data for Test 23 5. Heat Transfer Testing The final test series were conducted to collect heat transfer data and to compare cooling rates with LN2 and LNG. The test plates were 6 in. x 6 in. plates with a 5 in. foam trough caulked around the perimeter as shown in Figure 62. There was also one 18 in x 18 in plate with a centered 6 in. x 6 in. trough caulked in the center. The plate thickness varied from ¼ in. to ¾ in. and the plates were tested bare, painted with a marine epoxy primer, or painted with an epoxy primer and urethane top coat. There were twelve thermocouples on the 6 in. plates and 24 thermocouples on the 18 in. plate. The layout of the thermocouples and the description of the six tests conducted with LNG and LN2 are presented in Figure 63.

Figure 62. Plate setup used in heat transfer tests

49

6.00 in

2.00 in

18.00 in

1 7

8.00 in 6.00 in

2

3

8

9

4

5

6

10

11

12

Foam Trough 2.00 in

1.00 in 2.00 in

Small Te s t Pla te

1.00 in

1 13

Plates for Test

18.00 in

8.00 in

2

3

7

8

14

15

19

20

1. 6" x 6" x 3/4" Bare 2. 6" x 6" x 3/4" Epoxy Primer

4

5

16

17

6

2.00 in

18

3. 6" x 6" x 3/4" Epoxy Primer and Urethane

1.00 in

4. 6" x 6" x 1/4" Bare

9

10

21

22

1.00 in

5. 6" x 6" x 1/4" Epoxy Primer 6. 18" x 18" x 3/4" Epoxy Primer

11

12

23

24

2.00 in Large Tes t Plate

Figure 63. Heat transfer tests dimensions and thermocouple layout For the LN2 tests (all six tests collectively designated Test 15), the test plates were filled by pouring in the LN2 from a small 5 liter dewar. The test plates were placed on a Mettler PC8000 precision balance scale and the trough was filled in about 10 seconds. Hand readings were taken and recorded every 15 seconds as the LN2 evaporated. The thermocouple data for the six LN2 heat transfer tests is provided in Figure 64 through Figure 69. A number of thermocouples lost plate contact or malfunctioned and were not plotted. For the ¾ in. thick plates, the surface coatings allow for more efficient cooling as observed in the Phase I tests. However, the differences between the epoxy primer only and epoxy primer with urethane are minimal. The thinner ¼ in. plates cooled much more efficiently than the ¾ in. plates. The addition of the surface coating to the ¼ in. plate did have a noticeable effect, but less than of the ¾ in. plate. For the 18 in. x 18 in. plate tests, only about 200 seconds of data was collected. Due to the increased mass of the plate, the cooling rate to that point was considerably slower than for the 6 in. plates.

50

50

0

tc 1 tc 2

Temperature (C)

tc 3 -50

tc 4 tc 5 tc 6

-100

tc 7

tc 8 tc 9 -150

tc 10 tc 11 tc 12

-200 0

100

200

300

400

500

600

Time (sec)

Figure 64. LN2 heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Bare

50

0 tc 1 tc 2 Temperature (C)

tc 3 -50

tc 4 tc 5 tc 6 tc 7

-100

tc 8

tc 9 tc 10

-150

tc 11 tc 12 -200

0

100

200

300

400

500

600

Time (sec)

Figure 65. LN2 heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Epoxy

51

50

0 tc 1

Temperature (C)

tc 3 -50

tc 4 tc 5 tc 6

-100

tc 7 tc 8 tc 10

tc 11

-150

tc 12

-200 0

100

200

300

400

500

600

Time (sec)

Figure 66. LN2 heat transfer thermocouple tata – 6 in. x 6 in. x ¾ in. Epoxy Urethane

50

0 tc 1

Temperature (C)

tc 2 tc 3

-50

tc 4

tc 6 tc 7

-100

tc 8 tc 9 tc 10

-150

tc 11 tc 12 -200 0

100

200

300

400

500

600

Time (sec)

Figure 67. LN2 heat transfer thermocouple fata – 6 in. x 6 in. x ¼ in. Bare

52

50

0

tc 1 tc 2

Temperature (C)

tc 3 -50

tc 4 tc 5 tc 6

-100

tc 7 tc 8 tc 9

-150

tc 10 tc 11 tc 12

-200 0

100

200

300

400

500

600

Time (sec)

Figure 68. LN2 heat transfer thermocouple data – 6 in. x 6 in. x ¼ in. Epoxy 50 tc 1

tc 2 tc 3

tc 4

0

tc 5 tc 6

tc 7 tc 8

-50

Temperature (C)

tc 9 tc 10 tc 11

tc 12 -100

tc 13 tc 14

tc 15 tc 16

tc 17

-150

tc 18 tc 19

tc 20 tc 21

-200

tc 22 0

100

200

300

400

Time (sec)

500

600

tc 23 tc 24

Figure 69. LN2 heat transfer thermocouple data – 18 in. x 18 in. x ¾ in. Epoxy

53

The LNG heat transfer coefficient tests (all six tests collectively designated Test 24) were conducted remotely as illustrated with the schematic in Figure 70. The low pressure CRYOCYL 180 dewar in Figure 71 was purchased from Chart Industries. The dewar was filled using an LNG tanker brought on site for the LNG fire tests, using a standard truck hose coupling mounted to the dewar support rack. A Magnatrol F25M21 solenoid valve was used to control the flow from the dewar during the test plate filling process. During the tests, the trough filling was monitored by remote camera. A flow deflector, shown in Figure 72 was use to keep the flow from impinging directly on the bottom surface of the plate during filling. The fill pipe and deflector were held approximately 1” above the bottom surface of the test plate. An Interface SM-25 load cell was used to measure the weight of the LNG as a function of time. The thermocouple data and the load cell data were coupled in time by using the timing of the Magnatrol solenoid valve actuation. The thermocouple data for the six LNG heat transfer tests is provided in Figure 73 through Figure 79. For the ¾ in. thick plates, the surface coatings allow for only slightly more efficient cooling. As with the LN2, the differences between the epoxy primer only and epoxy primer with urethane are minimal. The thinner ¼ in. plates cooled much more efficiently than the ¾ in. plates. The addition of the surface coating to the ¼ in. plate again did have a slight effect. For the 18 in. x 18 in. plate tests, approximately 1 hour of data was collected. Due to the increased mass of the plate, the cooling rate to that point was considerably slower than for the 6 in. plates. Figure 79 provides the cooling for the 18 in. plate over the first 600 seconds (10 minutes) only.

Test Plate Fill Hose Assembly (approximate 6 ft)

Liquid Fill & Withdrawal Valve Solenoid Valve

Foam Trough Surrounding Plate (volume 0.6 gallons)

Pressure Relief Valve

Vent valve Cryo-Cyl Burst Disk , PRV and Pressure guage

Test Plate Bonding Cable TC Wires

Foam Insulation Under Test Plate Test Plate Support

Figure 70. Schematic of the LNG heat transfer tests

54

Figure 71. LNG heat transfer test configuration

Figure 72. Test plate and flow deflector

55

Figure 73. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Bare

Figure 74. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Epoxy

56

Figure 75. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¾ in. Epoxy Urethane

Figure 76. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¼ in. Bare

57

Figure 77. LNG heat transfer thermocouple data – 6 in. x 6 in. x ¼ in. Epoxy 50 tc 1

tc 2 tc 3 tc 4

0

tc 5 tc 6

tc 7 tc 8

-50

Temperature (C)

tc 9 tc 10 tc 11

tc 12 tc 13 -100

tc 14 tc 15 tc 16

tc 17 tc 18

-150

tc 19 tc 20

tc 21 tc 22 tc 23

-200 0

500

1000

1500

2000

2500

3000

3500

4000

tc 24

Time (sec)

Figure 78. LNG heat transfer thermocouple data – 18 in. x 18 in. x ¾ in. Epoxy

58

50 tc 1

tc 2 tc 3 tc 4

0

tc 5 tc 6

tc 7 tc 8

-50 Temperature (C)

tc 9 tc 10 tc 11

tc 12 tc 13 -100

tc 14 tc 15 tc 16

tc 17 tc 18

-150

tc 19 tc 20

tc 21 tc 22 tc 23

-200 0

100

200

300

400

500

600

tc 24

Time (sec)

Figure 79. LNG heat transfer thermocouple data – 18 in. x 18 in. x ¾ in. Epoxy (zoom)

Comparisons of the LNG and the LN2 cooling rates in Figure 82 and Figure 83 show more efficient cooling for the LNG for the ¾ in. bare and epoxy urethane plates, respectively. The LNG tests were conducted with an initial temperature approximately 20oC cooler than the LN2 tests. However, even when adjusting for this difference, the LNG cooling rates are still higher than the LN2 rates. Finally, the lowest achieved temperatures in the plates are lower for the LN2 since it is -191oC and LNG is only -161oC. This difference is not significant since the steel will enter the lower shelf fracture toughness regime well above either of these temperatures. The fracture toughness of the materials used in these tests is discussed in Volume III (Petti et al., 2011) of this study.

59

50

Temperature (C)

0

-50

LNG TC1 -100

LN2 TC1

-150

-200 0

100

200

300

400

500

600

Time (sec)

Figure 80. LNG vs. LN2 heat transfer thermocouple data for TC1 – 6 in. x 6 in. x ¾ in. Bare 50

Temperature (C)

0

-50

LNG TC1

-100

LN2 TC1

-150

-200 0

100

200

300

400

500

600

Time (sec)

Figure 81. LNG vs. LN2 heat transfer thermocouple data for TC1 – 6 in. x 6 in. x ¾ in. Epoxy Urethane

60

6. Summary and Conclusions This report summarizes the four phases of large-scale testing conducted as part of the Cascading Damage Study to investigate the effect of cryogenic liquids contacting steel plates. The purpose of these tests were to investigate the cooling of steel plates subjected to cryogenic liquids and to study the development and propagation of cracks in the plates due to the induced thermal stresses and lowering fracture toughness. The plate sizes ranged from simple 4 ft square plates in Phase II to 12 ft x 3ft structures with more complex geometry in Phase III. Twenty-two structural tests were conducted in Phases II and III, nine of which resulted in cracking. In order to generate thermally induced fracturing, stress concentration were introduced with varying degrees of severity and typically in stages. In one instance (Test 16), the crack initiated from the weld toe near the connection of a stiffening plate to the main base plate and vertical intersecting plate. Two tests (Test 15 and 24) were each a collection of 6 individual heat transfer tests. The tests performed in the study are used in the validation of the failure model applied in the computational analysis portion of the project. The details of those analyses are provided in a separate report. Beyond their use in the computational analyses, the test results have provided the following observations and insights: The steel plate cracking observed in these tests is representative of brittle cracking observed in steel hull sections of commercial LNG tankers subjected to accidental LNG spills. Crack initiation requires a sufficient stress concentration given the temperature of the steel; however, LNG vessels are complex and aged structures containing a large population of crack initiators. Fracture propagation generally follows the flow path with the stress fields and microstructure influencing the local fracture patterns. Intersecting plates and stiffening elements do not restrict crack propagation along with the larger base plates if also subjected to the cryogenic temperatures. Crack propagation is arrested when the crack leaves the low-temperature material region and propagates into the warmer region where the metal is ductile. At this point, load redistribution of the structure/tanker could cause the brittle cracks to extend under ductile tearing. However, since these structures only experienced initial gravitational loads, no noticeable ductile tearing extension was observed in these tests. In actual structures, including LNG tankers, other stresses will be imposed that could theoretically lead to further crack extension by ductile tearing. Cooling rates on steel plate sections rely on many factors including the size of the structure, the size of the spill (trough), the plate thickness, the specific cryogenic liquid (LNG vs. LN2), the presence of water at or near the spill, and the presence of a surface coating. The cooling rate observations and data from these tests are used in the computational analyses.

61

References American Society of Testing and Materials (ASTM), (1993) Manual on the Use of Thermocouples in Temperature Measurement: 4th Edition, Manual Series MNL 12. Anderson, T.L., (1995), Fracture Mechanics Fundamentals and Applications, Second Edition. Figueroa, V.G., Lopez, C., O’Rourke, K.K., (2011), LNG Cascading Damage Study Volume II: Flow Analysis for Spills from Moss and Membrane LNG Cargo Tanks, SAND2011-9464, Sandia National Laboratories, New Mexico. Hightower, M., et al. (2004). Guidance on Risk Analysis and Safety Implications of a Large Liquefied Natural (LNG) Spill Over Water, SAND2004-6258. Albuquerque, NM: Sandia National Laboratories. Hightower, M., Luketa-Hanlin, A., Gritzo, L.A., Covan, J.M. (2006). Review of Independent Risk Assessment of the Proposed Cabrillo Liquefied Natural Gas Deepwater Port Project, SAND2005-7339. Albuquerque, NM: Sandia National Laboratories. Hertzburg, R.W., (1996), Deformation and Fracture Mechanics of Engineering Materials, Fourth Edition. Petti, J.P., Wellman, G.W., Villa, D., Lopex, C., Figueroa, V.G., Heinstein, M., (2011), LNG Cascading Damage Study Volume III: Vessel Structural and Thermal Analysis Report, SAND2011-6226, Sandia National Laboratories, New Mexico. Roue, Roger, (2011), Personal communication between Roger Roue, SIGTTO (Society of International Gas Tanker & Terminal Operators Ltd) and M. Hightower (Sandia National Laboratories).

62

Appendix A – Stress Concentrations In order to explain the effects of stress concentrations, a simple example is provided here. Equation 1 and 2 along with Figure 82 illustrate a stress concentration at the edge of a circular or elliptical hole in a plate (Hertzberg, 1996). For a circular hole, the major (a) and minor (b) axes are equal, a = b, and therefore, max = 3 applied. For an elliptical flaw, Eq. 1 can be modified to be a function of the radius of curvature ( ) of the end of the ellipse resulting in Eq. 2. Equation 2 also includes the assumption that the length of the ellipse is much larger than the radius of curvature. Equation 2 illustrates that as the length of the ellipse increases, or notch in our cases, the stress at the edge of the ellipse also increases. In addition, a reduction of also increases the edge stress. The introduction of the machined notches to the drilled holes in our tests increased the stress significantly.

Eq. 1 Eq.2

63

Figure 82. Stress Concentration at Circular and Elliptical holes in Plates

64

As continues to decrease toward 0, the ellipse transitions into a “crack” in which the tip theoretically becomes perfectly sharp. Cracks of prefect sharpness, = 0, do not actually exist, some small blunting is always present. In addition, material yielding prevents the stresses from increasing beyond the material strength. Using Linear Elastic Fracture Mechanics (LEFM) (Anderson, 1995), this mathematically sharp crack tip region is commonly referred to as a “singularity”. The amplitude of the singularity is defined with the stress intensity factor, KI-applied. Equation 3 is used to compute the stress intensity factor for a through crack in an infinite plate as show in Figure 83. The fracture toughness of the structural steels used in LNG construction decreases with temperature. At LNG temperature (-161oC), these steels are extremely brittle and have reached their “lower shelf” fracture toughness. This makes LEFM appropriate in predicting fracture when the applied stress intensity, KI-applied, exceeds the fracture toughness of the material for a given temperature and flaw size (a). Eq. 3

65

singularity

Figure 83. Crack Tip Stresses in Plates

66

Appendix B – Test Data The complete set of thermocouple data are provided in this Appendix for the Phase II and III tests. For Tests 1 through 12, Figure 11 shows the thermocouple locations. For Tests 17 through 21, Figure 27 illustrates the thermocouple locations, For Tests 13 and 14, Figure 39 shows the thermocouple locations. Finally, for Tests 16, 22, and 23, Figure 46 provides the thermocouple locations.

Figure 84. Test 1 – TCs 1-11

Figure 85. Test 1 – TCs 12-22

67

Figure 86. Test 2 – TCs 1-11

Figure 87. Test 2 – TCs 12-22

68

Figure 88. Test 3 – TCs 1-11

Figure 89. Test 3 – TCs 12-22

69

Figure 90. Test 4 – TCs 1-11

Figure 91. Test 4 – TCs 12-22

70

Figure 92. Test 5 – TCs 1-11

Figure 93. Test 5 – TCs 12-22

71

Figure 94. Test 6 – TCs 1-11

Figure 95. Test 6 – TCs 12-22

72

Figure 96. Test 7 – TCs 1-11

Figure 97. Test 7 – TCs 12-22

73

Figure 98. Test 8 – TCs 1-11

Figure 99. Test 8 – TCs 12-22

74

Figure 100. Test 9 – TCs 1-11

Figure 101. Test 9 – TCs 12-22

75

Figure 102. Test 10 – TCs 1-11

Figure 103. Test 10 – TCs 12-22

76

Figure 104. Test 11 – TCs 1-11

Figure 105. Test 11 – TCs 12-22

77

Figure 106. Test 12 – TCs 1-11

Figure 107. Test 12 – TCs 12-22

78

Figure 108. Test 17 – TCs 1-10

Figure 109. Test 17 – TCs 11-20

79

Figure 110. Test 18 – TCs 1-10

Figure 111. Test 18 – TCs 11-20

80

Figure 112. Test 19 – TCs 1-10

Figure 113. Test 19 – TCs 11-20

81

Figure 114. Test 20 – TCs 1-10

Figure 115. Test 20 – TCs 11-20

82

Figure 116. Test 21 – TCs 1-10

Figure 117. Test 21 – TCs 11-20

83

Figure 118. Test 13 – TCs 1-19

Figure 119. Test 13 – TCs 36-54

84

Figure 120. Test 13 – TCs 20-35 (TC 25 not used)

Figure 121. Test 14 – TCs 1-19

85

Figure 122. Test 14 – TCs 36-54

Figure 123. Test 14 – TCs 20-35 (TC 25 not used)

86

Figure 124. Test 16 – TCs 1-18

Figure 125. Test 16 – TCs 27-44

87

Figure 126. Test 16 – TCs 19-26

Figure 127. Test 22 – TCs 1-18

88

Figure 128. Test 22 – TCs 27-44

Figure 129. Test 22 – TCs 19-26

89

Figure 130. Test 23 – TCs 1-18

Figure 131. Test 23 – TCs 27-44

90

Figure 132. Test 23 – TCs 19-26

91

DISTRIBUTION EXTERNAL DISTRIBUTION Bob Corbin Director, Oil and Gas Global Security and Supply United States Department of Energy Office of Oil and Natural Gas FE32 Forrestal Building 1000 Independence Ave SW Washington DC 20585 SANDIA INTERNAL DISTRIBUTION MS1135 MS1108 MS0744 MS0899

R. Kalan, 1534 M. Hightower, 6111 J. Petti, 6233 RIM - Reports Mgmt, 9532

92

+Art

Search

Images

Maps

Play

YouTube

News

Gmail

Drive

Calendar

More

LNG accidents

Search

About 988,000 results (0.21 seconds)

Art MacKay

0

Share

SafeSearch m oderate

Web Images Maps Videos News Shopping More

Any time Past 24 hours Past week Custom range... All results By subject Personal Any size Large Medium Icon Larger than... Exactly... Any color Full color Black and white

Page 2

Any type Face Photo Clip art Line drawing Standard view Show sizes

https://www.google.com/search?q=tankers&hl=en&rlz=1G1ACAW_ENCA318&prmd=imvnsuz&source…

1/9

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

List of pipeline accidents in the United States From Wikipedia, the free encyclopedia

Contents 1 1890s 2 1920s 3 1930s 4 1940s 5 1950s 6 1960s 7 1970s 8 1980s 9 1990s 10 2000s 11 References

1890s 1890 On January 24, a gas explosion destroyed a home in Columbus, Ohio, attracting a crowd of onlookers. While people were still gathered to look at the ruins of the home, a second gas explosion happened in a nearby home. The second explosion caused 4 deaths, and there were 32 injuries from both explosions.[1] 1895 On April 25, a woman in Wilkinsburg, Pennsylvania was investigating the smell of gas in a basement, while using a portable lantern. A series of explosions and followed, injuring that woman, and another woman, and damaging 4 homes. The gas leak was caused by gas being diverted into an older, defective gas main in the area.[2]

1920s 1929 On July 22, two oil company patrolmen were killed by an explosion of a gas pipeline near Castaic, California.[3]

1930s 1930 On April 4, gas leaked into the sewer system in New York City, New York, and later exploded. 6 people were injured, 5,000 were evacuated from nearby buildings, and telephone cables were damaged. [4] 1930 A runaway horse smashed a wagon of lumber against a crude oil pipeline in Ripon, Wisconsin on May 24. The oil ignited and spread to nearby oil tanks, causing a blaze that destroyed a number of buildings. [5] 1930 Excavation in Fairport, New York caused a major gas explosion on July 30. 3 people were killed, 10 en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

1/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

were injured, and a 4 family house was damaged by the blast and following fire.[6] 1931 4 campers near Kilgore, Texas were burned to death when they were surrounded by gas from a pipeline leak that caught fire on April 17. The flames also spread to brush and timber in the area, preventing rescuers from reaching the bodies for 3 hours.[7] 1936 On February 19, a worker inside a sewer in Utica, New York ignited natural gas that had leaked into the sewer system. An explosion was triggered, and the following fire burned for more than 24 hours. 4,000 people were evacuated.[8] 1936 On November 21, a pipeline serving a loading dock in Port Arthur, Texas, ruptured and ignited. The burning oil killed 3 people, and injured 6 others.[9] 1937 An oil pipeline being repaired by gas welding exploded near Pryor, Oklahoma on January 26. 2 of the repair crew, and 4 wives of the repairmen were killed by the explosion and following fire. [10] 1939 On December 12, a pipeline being tested ruptured for 40 miles (64 km), near Wichita Falls, Texas, injuring one person.[11]

1940s 1940 A gas compressor plant exploded in Braintree, Massachusetts on April 4, killing four people and injuring 12 others.[12] 1940 On August 29, a newly hired crew of repairmen were working on fixing a pipeline leak near Buffalo, Oklahoma, when the pipeline exploded and started a fire. Five of the crew were killed, 10 others were burned, and 10 horses burned to death.[13][14] 1943 On January 18, a grass fire near Tyler, Texas spread to a leak in an 8 inch diameter natural gas pipeline. The gas leak was initially small, but grew quickly, until the gas flames were about 200 feet (61 m) high. Gas service was cut to 28,000 people.[15] 1944 The "Big Inch" crude oil pipeline ruptured in Connellsville, Pennsylvania, with the crude spill killing fish along a 12-mile (19 km) stretch of the Laurel Hill creek.[16]\ 1946 A crew working to connect a new gas main in Peru, Illinois on July 4, when the old gas main exploded, killing 5 of the work crew, and injuring 7 others.[17] 1948 On February 28, crude oil spilled from a ruptured pipeline leading to storage tank in Oklahoma City, Oklahoma. Some teen boys in the area saw crude oil bubbling out of manhole covers, and thought that igniting the oil would be a good idea. This caused a string of sewer explosions, causing manhole covers to fly 10 feet (3.0 m) into the air.[18] 1948 On March 18, the 20 inch diameter "Little Big Inch" natural gas pipeline near Petersburg, Indiana, exploded and burned, throwing pieces of the pipe as far as 300 feet (91 m) away from the blast point. 3 homes were destroyed by the fire.[19][20] 1948 October 18: Vapors from a leaking butane pipeline at a refinery in Texas City, Texas spread out along a nearby highway, causing a number of cars to stall. The gas then exploded, killing 4 people, and seriously burning 17 others.[21][22] 1948 On November 19, a "Big Inch" pipeline pumping station exploded and caught fire near Seymour, Indiana, causing $3,000,000 in damage, and injuring 17 workers at the station.[23][24] 1949 A section of the "Little Big Inch" exploded and burned in North Vernon, Indiana on March 4, burning a mother and her infant. It was the fourth explosion on that pipeline in Indiana that year.[25] 1949 A road grader operator was seriously burned when his grader hit a 6-inch gas pipeline west of en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

2/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

Mankato, Kansas on November 17.[26] 1949 On December 8, an explosion and fire occurred at a compressor station for a 24-inch natural gas pipeline in Centralia, Missouri. Flames could be seen for 150 miles (240 km) away.[27] 1949 A leaking gas line caused an explosion at a packing plant in Sioux City, Iowa on December 14. Eighteen workers were killed, and almost 100 inured.[28] 1949 On December 15, a 22-inch natural gas pipeline exploded and burned near Carthage, Tennessee, injuring two people. Flames shot 1,000 feet (300 m) into the air.[29]

1950s 1950 On March 13, an overhead pipeline at a refinery in Martinez, California leaked, causing flammable fumes to spread onto a highway. An automobile ignited the fumes, killed a woman, and injuring 2 other in vehicle. 3 auto were also burned.[30] 1950 The "Big Inch" gas pipeline exploded and burned on July 1, near Beallsville, Ohio. A house and a barn were destroyed by the fire.[31] 1950 Three workers were killed in an underground vault in Los Angeles, California on August 22, when a gas main exploded. There was no fire.[32] 1950 On September 7, a new natural gas pipeline exploded near Big Rapids, Michigan. Two barns were destroyed by the following fire, that was seen for 50 miles.[33] 1950 On November 24, a newly built 30 inch natural gas pipeline ruptured for nearly 3,000 feet (910 m), causing a fire that destroyed 2 homes under construction near King of Prussia, Pennsylvania.[34] 1951 Two men welding on a crude oil pipeline at an oil Terminal in Kansas City, Kansas were severely burned on January 7, when a nearby valve failed, spraying them with crude oil that ignited. Both later died of their burns.[35] 1951 On January 10, two gas explosions, 3 hours apart, hit McKees Rock, Pennsylvania, injuring 8 people, igniting a fire, and causing widespread damage.[36] 1951 A gas main pressure regulator failed in Rochester, New York on September 21, causing a series of explosion that last for 4 hours. 3 people were killed, and 30 homes were destroyed.[37] 1951 A Halloween Parade on October 31 in Pittsburgh, Pennsylvania was interrupted by 4 gas main explosions. 29 people were injured.[38] 1951 A 12 inch diameter temporary gas transmission pipeline exploded and burned near Cranberry, Pennsylvania on November 27, causing a 200-foot (61 m) high flame that could be seen for a number of miles away. The explosion was heard for 10 miles around. A pipeline compressor station under construction at the site was destroyed. A nearby Elementary school was relocated following the failure. [39][40] 1952 Four men working on an 8 inch gas pipeline near Mount Pleasant, Michigan were burned when that en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

3/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

pipeline ruptured as they raised it for reconditioning on September 26.[41] 1952 On December 29, twelve persons were injured in a blast that shook the Lawndale district of Los Angeles, California, when a ditching machine broke a gasoline-carrying pipeline and touched off a fiery explosion.[42] 1953 Five square miles of east Boston, Massachusetts was evacuated on September 9 from a pipeline leaking about 1,000,000 US gallons (3,800,000 L) of gasoline.[43] 1953 On September 10, a gas explosion in Cleveland, Ohio killed one person and injured 50 others.[44] 1953 A US Air Force T-33 trainer jet crashed into a natural gas pipeline bridge over the Mississippi River on November 24 near Greenville, Mississippi, rupturing and igniting the pipeline.[45] 1954 A 40 to 50-year old LP gas distribution line was blamed for causing an explosion in Goldsboro, North Carolina on April 12 that killed 5 people, injured 15 others, and demolished 3 buildings.[46] 1955 The "Big Inch" gas pipeline exploded and burned near Roseville, Ohio on March 7. Flames reached 400 feet (120 m) high, and 8 acres (32,000 m2) of brush & timber burned.[47] 1955 On March 9, a pipeline construction crew of 4 were killed while trying to move a pipeline for the building of a Toll road in Chesterton, Indiana. Two other pipeline workers were injured, and a school a quarter mile away was evacuated.[48] 1955 A burst pipeline at a Refinery in Sunburst, Montana contaminated groundwater and soil in the area. Despite pumping out over 182,000 US gallons (690,000 L) of gasoline, pollution from the accident remained. In 2004, local residents and a school district won a Lawsuit for payments for damages.[49][50] 1955 A bulldozer ruptured and ignited a gas pipeline in Brookshire, Texas. Flames reached 250 feet (76 m), and the bulldozer operator was killed.[51] 1955 On August 10, a gas leak in Ashtabula, Ohio was ignited by electrical equipment or lightning, causing a restaurant to explode. 21 people were killed, 15 more were injured, and 6 buildings destroyed.[52] 1955 A gas pipeline being tested in Detroit, Michigan exploded and burned on September 7, injuring one person, and destroying 50 cars.[53] 1955 On October 10, a crew cleaning the outside of a natural gas pipeline with a heavy rubber ball ruptured a coupler, causing an explosion and fire east of Orleans, Indiana. Two members of the crew ere killed, and 3 others were injured.[54] 1955 A drag line operation in a gravel pit in Irving, Texas ruptured an 8 inch diameter gasoline pipeline on November 30. Gasoline spread out over about 10 acres (40,000 m2), then exploded and burned. 1 home was destroyed, but the family living there was away at the time of the explosion.[55] 1956 On February 11, a corroded gas line from a gas main leaked, causing an explosion that killed 3 people at a meat packing plant in Toledo, Ohio.[56] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

4/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1956 A trench digging machine being used in building a freeway cut into a gasoline pipeline in West Los Angeles, California on February 27. At least 3 people were burned, with 9 homes, a warehouse, and a laundry catching on fire.[57] 1957 On January 16, an explosion and fire occurred at a natural gas compressor station in Liberal, Kansas, killing 3 workers at that station. 11 other workers were injured, and the fire burned for 2 hours. The shut down of this gas pipeline from the explosion affected customers as far away as Ohio in sub-zero weather conditions.[58][59] 1957 Two explosion from a natural gas main killed 3 people in Peoria, Illinois on January 17. 7 others were injured, and a home and a 2 story building were leveled.[60] 1957 A leaking gas main in Reno, Nevada led to three explosions on February 6. 2 people were killed, 42 others injured, and 5 buildings were destroyed.[61] 1957 On June 3, a 26 inch diameter natural gas transmission pipeline exploded and burned near Ellinwood, Kansas, destroying a farm house. One person was injured.[62] 1957 On December 5, a gas line in the basement of a store that was being worked on in Villa Rica, Georgia, exploded. 13 people were killed in the explosion and following fire. At least 6 stores were destroyed.[63] 1958 A natural gas metering station in Kimberly, Idaho exploded on February 17, killing two pipeline company workers, injuring another worker, and destroyed the metering building. There was no fire. [64][65] 1958 On June 1, gas leaking from a pipeline near Big Spring, Texas was ignited and exploded, killing 3 fishermen and seriously burning another fisherman.[66] 1958 A truck missed a curve on a road and crashed into a gas transmission pipeline compressor station near Kings Mountain, North Carolina on September 16. There was an explosion and fire, and the 2 men in the truck were killed.[67] 1958 On October 4, a gasoline pipeline was ruptured by a bulldozer in Hobbs, New Mexico. The gasoline ignited, injuring 3 people, damaging 6 homes, and threatened a number of other homes for a time. [68][69] 1958 On November 9, a jet fuel pipeline ruptured near the Blue Creek in Idaho. Fuel flowed down the creek, and later ignited, damaging one home and destroying 6 bridges. Several were sicked by the fumes from incident.[70][71] 1958 A leaking and burning gasline under a street lead to several explosions at a Hotel in Allentown, Pennsylvania, on December 14. 7 people were killed and 23 others injured.[72] 1959 A worker on gas transmission pipeline was closing a valve, when it exploded near Newton, Pennsylvania on September 25. The worker was killed, and another worker was injured.[73]

1960s en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

5/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1960 An estimated 125,000 persons in southwest Missouri were without gas in subfreezing temperatures for several days due to a ditch-digging machine rupturing a pipeline.[74] 1960 In July, excavation work in Merrill, Wisconsin causes a gas leak and gas explosion that killed 10 people.[75] 1960 A ditching machine used in laying a water main hit an 8 inch diameter natural gas pipeline in Sarasota, Florida on October 5. 9 People were injured in the following explosion and fire.[76] 1960 On October 27, a 16 inch diameter gas transmission pipeline near Checotah, Oklahoma exploded while it was being worked on to repair a leak. 2 of the repair crew died, and 4 others were injured. [77] 1960 A 30 inch gas transmission pipeline exploded and burned at a gas sub station in Huntington, West Virginia on December 19. Windows were broken, 1 homes was damaged, and brush burned, but there were no injuries.[78] 1961 On January 4, a gas pipeline failure near Waynesburg, Pennsylvania ignited, causing a fire that was widely seen in the area. There were no injuries.[79] 1961 On February 22, a pipeline exploded and burned in a refinery in Borger, Texas, killing 9 members of a construction crew, and burning another crewman.[80] 1961 The main City of Miami, Florida Garage was destroyed by a gas explosion on February 23. The blast was caused by a ditch digging machine being used in the garage hitting and rupturing a 2 inch gas pipe. One person was seriously burned by the blast, and 2 fire fighters were injured fight the fire that followed the blast.[81] 1961 A 36 inch gas transmission pipeline exploded near Laurel, Mississippi on June 18. 10 people were injured, and one home was destroyed from flames that went hundreds of feet in the air. A crater 30 feet (9.1 m) long and 20 feet (6.1 m) deep was created by the failure.[82] 1961 A 26 inch diameter gas transmission pipeline exploded and burned near Winchester, Kentucky on September 11. 22 people suffered various burn injuries.[83] 1961 On October 9, vapors from a leaking pipeline on an oil storage tank exploded and burned in Bridgeport, Illinois. 4 oil company workers were killed, and 3 others injured.[84] 1961 On November 19, a gas pipeline exploded and burned near Warrenton, Virginia. The blast created a crater 40 feet (12 m) long, 10 feet (3.0 m) wide, and 6 feet (1.8 m) deep. There were no injuries.[85] 1961 An 18 inch diameter natural gas pipeline exploded and burned near Cadiz, Ohio on November 25. There were no injuries or damage.[86] 1962 Gas leaking from a 10 inch diameter natural gas transmission pipeline exploded on February 20 in Portage, Ohio, injuring 6 people and destroying a home.[87] 1962 On June 14, a backhoe ruptured a gas transmission pipeline near Idaho Falls, Idaho. The escaping gas exploded and ignited later on while a crew was trying to repair the line. One of the crew was killed, and 5 others injured in the fire.[88] 1962 On August 2, a natural gas transmission pipeline exploded and burned in Clearwater, Florida, next to US Highway 19, forcing that road's closure for a time. There were no injuries reported. Investigators found the line had previous mechanical damage as a cause of the failure.[89][90] 1962 A 30 inch diameter gas transmission failed on August 2 in Kansas City, Kansas. The gas flowed for 10 minutes before exploding and igniting. An 8 inch gas distribution pipeline was also ruptured, 11 homes were destroyed, and 23 others were damaged. At least one person was injured.[91] 1962 On September 11, an 8 inch propane/LPG pipeline was ruptured by road building equipment near Eatonton, Georgia. One of the road workers was overcome and asphyxiated by the propane fumes. Propane fumes followed the Oconee River for 10 miles (16 km) into Lake Sinclair.[92] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

6/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1963 On January 2, a gas transmission pipeline ruptured due to a defective weld in San Francisco, California. The gas ignited, one firefighter died from a heart attack, and 9 other firefighters were injured fighting the resulting inferno.[93][94] 1963 An explosion and fire spread through a gas pipeline compressor station in Montezuma, Indiana on March 12, injuring 16 workers.[95] 1963 A crude oil pipeline was hit and ruptured by an earth mover near Fostoria, Ohio, on October 30. The earth mover operator was seriously burned in the resulting fire.[96] 1963 On October 31, a 6 inch diameter butane pipeline was ruptured by an earth mover near West Millgrove, Ohio. The equipment operator was critically burned by the following explosion and fire. [97] 1963 On November 17, flammable liquids leaking from a pipeline disposal pit were accidentally ignited, killing a teen planning to cook alongside a creek in South Carolina.[98] 1964 A Santa Fe Railroad Freight Train apparently ignited fumes from a leaking propane pipeline near Bosworth, Missouri on February 4. The explosion and fire ignited 4 diesel locomotives and some box cars, and derailed other box cars. One member of the Rail Crew was injured.[99] 1964 On February 7, 2 workers installing insulation on a valve in a manhole in Richardson, Texas were overcome by gas when an 8 inch pipeline in the vault ruptured, and were killed.[100] 1964 A front loader ruptured a gas pipeline in Fort Worth, Texas on February 28, seriously burning the loader operator.[101] 1964 On May 12, a bulldozer hit and broke a valve on an LPG pipeline near Demopolis, Alabama while grading land. The resulting fire caused fears of flames spreading to an underground storage facility, but the fire was later controlled. There were no injuries.[102] 1964 A crude oil pipeline ruptured in Gilbertown, Alabama on October 29. More than 72,000 US gallons (270,000 L) of oil were spilled.[103] 1964 A gas line being moved in Miami, Florida exploded and burned on November 18. 4 people were injured.[104] 1964 On November 25, a recently replace natural gas transmission pipeline exploded and burned in Saint Francisville, Louisiana, killing 5 workers of the pipeline, and injuring at least 23 others. [105] 1965 On January 6, a house in Garnett, Kansas was destroyed by an explosion, and later on gas was found leaking from a 2 inch gasline in the street front of it, and was suspected as the cause. A young boy was killed. The leak may have also caused another nearby house explosion the previous November. [106] 1965 On January 21, an 8 inch diameter propane transmission pipeline 15 miles (24 km) east of Jefferson City, Missouri leaked. The propane spread along the ground, and exploded several hours later, scorching an area over a mile wide. A girl being dropped off at a school bus stop was severely burned and later died, and 2 other people were burned.[107] 1965 A 32 inch diameter gas transmission pipeline, north of Natchitoches, Louisiana, belonging to the Tennessee Gas Pipeline exploded and burned from Stress corrosion cracking(SCC) on March 4, killing 17 people. At least 9 others were injured, and 7 homes 450 feet from the rupture were destroyed. This accident, and others of the era, led then-President Lyndon B. Johnson to call for the formation of a national pipeline safety agency in 1967. The same pipeline had also had an explosion on May 9, 1955, just 930 feet (280 m) from the 1965 failure.[108][109][110][111][112] 1965 A crude oil pipeline ruptured east of Blanding, Utah on April 3, spilling about 5,000 barrels (790 m3) of crude oil into the San Juan River. The ruptured pipeline was reported to flow "wide open" for over an hour.[113] 1965 On July 24, a natural gas pipeline exploded and burned when workers were welding on a tie-in en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

7/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

pipeline onto it near Tescott, Kansas. One of the workers died, and 15 others were injured.[114] 1965 On August 21, a 9-year-old girl was killed and eight people were injured in a pipeline explosion in western Van Wert County, Ohio. The explosion threw up flames that could be seen from 40 miles (64 km) away and scorched a 100-acre (0.40 km2) area of farmland. Nancy Anna May Rigdon was killed in her bed in a house 300 yards from the blast site. The rest of her family was injured but survived. Investigators said the explosion was caused by gas leaking from an eight-inch pipeline apparently ignited by a spark from a passing train.[115][116] 1965 A 8 inch diameter gasoline pipeline ruptured in Sylvania, Ohio on August 23. The danger of fire or explosion forced evacuations of residents in a 2-square-mile (5.2 km2) area. There was no fire.[117] 1965 On October 25, a ruptured pipeline spilled naphtha in Mount Cory, Ohio, forcing evacuations until the naphtha evaporated.[118] 1966 A 6 inch diameter natural gas pipeline ruptured in Norfolk, Nebraska on January 28, shutting off gas to 20,000 people in 10 communities on January 28.[119] 1966 On December 14, a leaking propane pipeline near Swedenborg, Missouri made a car stall. Other came to aid the stalled car, and someone lit a cigarette, igniting the fumes. 8 people were burned and hospitalized.[120] 1967 A leaking gas main in the Jamaica section of New York City, New York caught fire on January 13. 2 pieces of FDNY equipment responding to the gas leak report were burned, as well as numerous buildings. The fire spread to 13 alarm size, with 63 fire companies being used to control the situation. The cause of the leak was the failure of a moisture scrubbing "drip pot" on the pipeline.[121][122] 1967 A 6 inch diameter propane pipeline exploded and burned while it was being worked on in Meeker, Oklahoma on January 10. One of the workers was killed, and another injured.[123] 1967 Manufacturers Light and Heat Company announced they were requesting to the Federal Power Commission permission to allow a new pipeline to replace 73.5 miles of older pipeline, which was having 200 to 450 leaks a year in Eastern Pennsylvania.[124] 1967 On May 16, a pile driver ruptured a propane pipeline in Dearborn, Michigan. The escaping gas caught fire, with 2 construction workers being killed, and 4 others seriously burned.[125] 1967 A leaking pipeline released 30,000 barrels (4,800 m3) of JP-4 grade jet fuel in Wilmington, California on June 30. There was no fire.[126] 1968 A petroleum products pipeline was discovered to be leaking on January 27, near Kokomo, Mississippi. Damage to cotton crops and water wells was discovered soon afterward.[127] 1968 On April 6, natural gas leaking from a pipeline in Richmond, Indiana built up in a sporting goods store and exploded. Gunpowder in the that store exploded later on. 41 people were killed, 150 were injured, and 15 buildings destroyed.[128][129][130] 1968 On April 15, gasoline odor was detected at a drinking fountain in Glendale, California. The source of the water well that fed the fountain was determined to be a 8 inch pipeline that was leaking. Between 100,000 and 250,000 US gallons (950,000 L) of gasoline were leaked into the local groundwater.[131] 1968 On May 29, a bulldozer ruptured a 1-inch gas service line at a children's nursery in Hapeville, Georgia, causing an explosion and fire. Seven children and two adults were killed, and three children were seriously injured in the accident.[132][133] 1968 An 8 inch diameter propane pipeline rupture in a landslide ruptured near Plainfield, Ohio on June 2. 2 people were killed, 3 others injured by burns, and 7 buildings and 7 vehicles were destroyed.[134] 1968 A contractor laying a new pipeline broke an old pipeline in Norwalk, Ohio on August 7, spilling en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

8/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

gasoline for 4 hours into the Huron River.[135] 1968 On August 22, a 16 inch diameter gasoline pipeline ruptured at General Mitchell Field, spilling almost 200,000 US gallons (760,000 L) of gasoline, and forcing closure of one runway. Previous damage to the pipeline by heavy equipment working in the area was identified as the cause of the rupture. [136][137] 1968 A coal company digging machine hit an 8 inch LPG pipeline in Fulton County, Illinois on September 3, killing one person and injuring 4 others.[138] 1968 2 teen boys shooting a rifle ignited gasoline leaking from a petroleum pipeline pumping station near Midland, Pennsylvania on November 3. A large brush fire ensued. Both boys had moderate burns. A stuck relief valve on the pipeline was the cause of the leakge.[139][140] 1968 An LPG pipeline, near Yutan, Nebraska ruptured on December 5. Repair crews responded to the pipeline rupture, and thought LPG vapors were dispersed, but ignited the vapor cloud by driving into it. Five repairmen were killed. After the accident, the Nebraska State Fire Marshal ordered MAPCO to reduce its operating pressure, and to hydrostatic retest 52 miles (84 km) of that pipeline. During the tests, 195 longitudinal seams failed.[141][142] 1968 On December 18, a 30 inch diameter gas pipeline exploded and burned at a gas processing plant in Gibson, Louisiana. One plant worker was injured.[143] 1969 On January 13, a 22 inch diameter crude oil pipeline ruptured in Lima, Ohio, spilling over 2,000 US gallons (7,600 L) of oil into the sewer system. Cracks from welding were blamed for the failure.[144] 1969 A leaking crude oil pipeline caused a slick 35 miles (56 km) long in the Dry Creek near Greybull, Wyoming on February 24.[145] 1969 A 10 inch pipeline carrying aviation gasoline was ruptured by explosives on March 17 in Canyon, California The fuel caught fire short after that.[146] 1969 On May 6, a gas pipeline in Pittsburgh, Pennsylvania, that had been moved, was undergoing pressure testing when a cap on it blew off, hitting and rupturing another nearby gas pipeline. That pipeline exploded and burned, killing 1 worker, injuring 9 other workers, and damaging 3 homes.[147] 1969 Overpressure of a low pressure natural gas distribution system in Gary, Indiana caused numerous small fires and explosions. A gas company worker's errors allowed much higher than normal gas pressure in a gas distribution system. 56 square blocks were evacuated, 7 people were injured, 6 homes destroyed, and 19 other homes damaged.(June 3, 1969)[148][149] 1969 On September 9, a converted natural gas pipeline running at 789 psi near Houston, Texas ruptured, causing a massive fire. Construction work downstream of the accident led to a pressure build up that caused the rupture. 7 people were injured, 13 homes were destroyed, and many others damaged.[150][151] 1969 On December 25, a land leveler ruptured a 22 inch natural gas transmission pipeline in Hermiston, Oregon. Gas at 600 psi sprayed from the pipeline. A warning sign about the existence of the gas pipeline was 10 feet (3.0 m) away from the rupture site.[152]

1970s 1970 A leak natural gas pipeline exploded in Houma, Louisiana on January 24, killing 3 people, and demolishing half a block of downtown buildings.[153] 1970 Early on September 2, residents of Jacksonville, Maryland, detected gasoline odors and noticed gasoline in a small creek flowing beneath a nearby road. Because fumes were still present in the late afternoon of September 2, a resident notified Colonial Pipeline at 6:19 p.m. about the situation. About 12 hours later, on the morning of September 3, an explosion and fire occurred in a ditch in which contractor en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

9/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

personnel for Colonial were manually digging to further expose the pipeline & catch gasoline trickling from the ground. Five persons were injured, none fatally. The NTSB later pointed out that workers had failed to use a flammable vapor percent detector during the work. The leak point was found 4 days later. The failure resulted in a release of 30,186 gallons (718 barrels) of gasoline and kerosene.[154][155][156] 1970 On September 7, a pipeline leaked thousands of gallons of gasoline onto farmland near Ontario, Oregon. Roads were closed in the area until the gasoline was removed.[157] 1970 The 1970 Propane vapor cloud explosion in Port Hudson, Phillips Pipeline Company on December 9, 1970. propane gas explosion, Franklin County, Missouri. A leak led to propane cloud explosion with a force estimated up to 50 tons of TNT. The NTSB cited past external and internal corrosion issues, and poor welds on the uncoated pipeline as concerns.[158][159][160] 1970 Explosion of a 30-inch diameter 1100 psi inlet natural gas pipeline, bringing offshore natural gas into a gas drying plant in southern Louisiana. Two plant personnel were killed. Rupture was at a junction of a 12inch gas line to the 30-inch main line. (December 1970) 1970 A restaurant owner opened a gas line valve in New York, New York, not knowing that part of the gas line was open and unconnected. The gas in the building exploded, killing 15 people, & injuring more than 60 others. (December 11, 1970)[161] 1970 On December 28, a 12 inch diameter pipeline ruptured in Jackson, Wisconsin, spilling 200 barrels (32 m3) of fuel oil into a wildlife sanctuary.[162] 1971 A faulty valve on a 3 inch diameter natural gas pipeline was suspected of causing a gas leak that resulted in 3 separate explosions, including a house explosion in Lambertville, New Jersey that killed 7 people.[163] 1971 On June 5, an ammonia pipeline failed near Floral, Arkansas, releasing 80 tons of ammonia.[164] 1971 2 gas explosions in North Richland Hills, Texas on October 4. Gas migrated into 2 homes from leaking gas pipes.[165] 1971 A gas company repair crew was overcome in a service vault on November 17, in Pittsburgh, Pennsylvania. 2 workmen were overcome initially, and 4 others attempting to rescue them were also overcome by gas asphyxiation. All 6 died.[166] 1972 The second pipeline leak in a month into the Tippecanoe River in Indiana on a Buckeye Pipeline company (now Buckeye Partners ) line hit on January 12. The Buckeye Pipeline was owned by the bankrupt Penn Central Railroad, preventing money from being spent on repairs. One EPA official stated "they know they have a leaky system".[167] 1972 On January 11, a 10 inch diameter pipeline ruptured in Clinton, Montana, spilling 3,000 barrels (480 m3) of diesel fuel, with some of it reaching the Clark Fork River.[168][169] 1972 During the blowdown of a pipeline dehydrator, LPG fumes caught fire at Conway, Kansas on January 29. 1972 On February 12, a pipeline rupture spilled 16,000 US gallons (61,000 L) of diesel fuel into the Spokane River.[170][171] 1972 A Natural gas explosion at Annandale, Virginia, on March 24.[172] 1972 On June 15, a crew was welding on a gas main in Bryan, Ohio that had been shut off, when someone inadvertently openned a valve that fed gas into that main. The gas ignited, and exploded, serious injuring 2 workers.[173] 1972 A 12 inch diameter high pressure propane pipeline, near Butler, Alabama, was ruptured by a road grader. A short time after the line was ruptured, a car drove into the vapor cloud. The car stalled, and trying to restart it was suspected to have ignited the vapor cloud, killing four people. (June 20, 1972)[174][175] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

10/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1972 A gasoline pipeline ruptured and ignited at a Plantation Pipeline Terminal in Bremen, Georgia on September 6. For a time, there were fears the fire might spread to nearby fuel storage tanks, but the fire was limited to the pipeline.[176] 1972 In October, a crude oil pipeline ruptures near Shiprock, New Mexico, spilling 285,000 US gallons (1,080,000 L) of crude oil into the San Juan River, polluting it for 200 miles.[177][178] 1972 On October 30, a bulldozer working on a power company construction project ruptured a gas main in Lake City, Minnesota. Leaking gas accumulated in, then exploded in a nearby variety store, killing 6 and injuring 9.[179] 1972 A leak in a weld on a 36 inch diameter gas transmission pipeline on November 18 in Bend, Oregon forced the shutdown of gas service to 3,000 customers.[180] 1973 A cracked gas main leaked in Adamsville, Alabama, on February 7. The escaping gas exploded, killing 3 people and injuring 2 others. A string of other gas main cracking incidents occurred in this city, killing one other person, and injuring t2 others.[181] 1973 Installation of a sewer was suspected of damaging a gas line in Coopersburg, Pennsylvania on February 21. Leaking gas later exploded in an apartment building, killing 5 people, injuring 22 others, and destroying the building.[182] 1973 In Austin, Texas, a natural gas liquids (NGL) pipeline ruptured due to an improper weld. A passing car or truck set off a vapor cloud explosion and fire. Six people were killed, and 2 others injured. (February 22, 1973)[183][184] 1973 On May 2, a 10 inch diameter pipeline ruptured in Murray, Idaho, causing a mist of diesel fuel to cover homes and trailers. Between 7,000 and 10,000 US gallons (38,000 L) of fuel were lost. Some of the fuel reached a nearby creek. There was no fire.[185][186] 1973 Improper sampling procedures on an LPG pipeline killed one worker and injured another from freezing at Dayton, Ohio on May 3. 1973 In the summer, a pipeline ruptured in Diamond, Louisiana. The escaping gas fumes were ignited by a lawnmower, killing 2 people.[187] 1973 A pipeline failed near Findlay, Ohio on June 27, spill about 150,000 US gallons (570,000 L) of jet aviation fuel into the Ottawa Creek. A failed gasket caused the spill.[188][189][190] 1973 A crude oil pipeline ruptured in Los Angeles, California on October 18. Crude flowed along several streets for a time.[191] 1973 On December 4, a pipeline break releases 31,000 barrels (4,900 m3) of oil near Argyle, Minnesota.[192] 1973 On December 6, a pump station on an ammonia pipeline near Conway, Kansas, was started against a closed valve, and the pipeline failed in a previously damaged section. Two persons who drove through the ammonia vapors were hospitalized; several rural residents were evacuated from the area; and 89,796 US gallons (339,910 L) of anhydrous ammonia were lost.[193] 1974 A 22 inch diameter natural gas transmission pipeline failed in Prairie du Rocher, Illinois on January 2. The resulting fire caused no serious damage, but 7,000 people in the area were left without gas heating for several subfreezing days.[194] 1974 On March 2, a 30 inch diameter gas pipeline failed at 797 pounds pressure inside a 34-inch diameter casing pipe under a road near Monroe, Louisiana. 10 acres of forest were burned, but there were no injuries or deaths. A substandard girth weld was the cause. The failure of automatic valves on the pipeline to close upon a pressure drop were also cited in contributing to the size of the accident.[195][196] 1974 A gas transmission pipeline ruptured near Farmington, New Mexico on March 15, killing a family of 3 en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

11/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

in a truck driving nearby when the gas ignited. Corrosion along the longitudinal seem weld of the pipe section caused the failure.[197][198] 1974 A gas line in a commercial building in New York, New York, was ruptured by falling equipment in a basement on April 22. The escaping gas later exploded, injuring more than 70 people.[199] 1974 A previously damaged gas main ruptured in Philadelphia, Pennsylvania on May 3, causing an explosion that killed 2, and caused extensive damage to 4 row homes. Earlier plumbing work was suspect to have caused the gas line damage.[200] 1974 On May 21, a 6 inch gas-gathering pipeline, ruptured at the edge of a rural road south of Meridian, Mississippi. Three vehicles entered the area which contained the escaping gas, and stalled near the rupture. The gas ignited at 10:05 p.m., and five persons died as a result. The 3 vehicles were destroyed and 40 acres (160,000 m2) of woodland were burned. Although less than 4 years old, the 6-inch pipe had corroded internally and had been embrittled by hydrogen.[201] 1974 A 30 inch gas transmission pipeline failed and gas ignited near Bealeton, Virginia, on June 9, from hydrogen stress cracking. Failure alarms at the nearest upstream gas compressor station did not activate, and the pipeline failure was first notice by a compressor station employee happening to see the large fire from the pipeline rupture.[202][203][204] 1974 On August 13, an ammonia pipeline failed near Hutchinson, Kansas after a pump station was started against a closed valve. 3 police officers were treated for ammonia inhalation; approximately 200 persons were evacuated from the area of the vapors; trees, lawns, shrubbery, and crops were burned; and an estimated 11,000 fish were killed.[205] 1974 On September 14, a propane pipeline to an underground storage cavern failed in Griffith, Indiana. The propane later caught fire. 1,000 residents were evacuated during the incident.[206] 1974 A 12 inch diameter gas gathering pipeline exploded and burned near Meta, Kentucky on November 24. There were no injuries reported. Acts of previous vandalism against the pipeline company had happened before.[207][208] 1974 A crew repairing a leaking crude oil pipeline near Abilene, Texas, were overcome by sour crude oil fumes on December 1. Six of the repair crew died. The leak was cause by improper welding.[209] 1975 A crude oil pipeline at Lima, Ohio ruptured after a valve was accidentally closed against a pumping pipeline on January 17. The spraying crude oil ignited, killing a Terminal Operator. [210] 1975 On January 23, a propane chiller exploded violently during maintenance work on it near Iowa City, Iowa. 2 workers were killed and 3 others injured by the failure.[211] 1975 In March, a leak was discovered in a 14 inch diameter petroleum products pipeline in Mecklenburg County, North Carolina. Plantation Pipeline repaired the pipeline immediately, and began efforts to recover the spilled petroleum. From that time through June 1983, approximately 2,022 barrels of spilled petroleum products were recovered from standpipes at the leak site. Remediation efforts stopped in October 1984. Later tests raised questions on the possibility of not all of the spill products were recovered.[212] 1975 A 12 inch diameter crude oil pipeline ruptured near Harwood, Missouri, on March 26. Heavy rain slowed the cleanup.[213] 1975 A natural gas liquids (NGL) pipeline ruptured due to previous mechanical damage at Devers, Texas. 4 people were killed in a following vapor cloud fire. The pipeline had been damaged when a valve was installed on the pipeline. (May 12, 1975)[214] 1975 An explosion in June 1975 at a home in East Stroudsburg, Pennsylvania, was caused by natural gas leaking into the home from an open main in the middle of the street. One person was killed. In 1973, workers hired by the gas company had falsified records showing the main had been closed.[215] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

12/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1975 An LPG pipeline ruptured near Romulus, Michigan, due to previous mechanical damage to the pipeline, and over pressurization from operator error, caused by closing a valve against a pumping pipeline, at a storage facility. Nine people were injured in the following vapor cloud fire. Flames 500 feet (150 m) high engulfed a 600-foot (180 m)-diameter area, destroyed four houses and damaged three others, burned 12 vehicles, and consumed 2,389 barrels (379.8 m3) of propane. (August 2, 1975)[216][217] 1975 On October 13, employees at a gas processing plant at Goldsmith, Texas heard leak gas, and investigated. Before the leak could be found, a 12 inch diameter pipeline there exploded, killing 3 of the crew, injuring 2 others, and causing extensive plant damage.[218] 1976 A repair crew working on natural gas gathering compressor station at Cedardale, Oklahoma on January 7, opened the wrong valve in an attempt to increase gas flow. Natural gas & Natural Gas Liquids flow out of an open 12 inch pipeline, and were ignited by an open flame heater. 5 of the crew were killed, and 2 seriously burned.[219] 1976 A gas leak at the Pathfinder Hotel in Fremont, Nebraska, exploded, killing 23 people on January 10. A compression coupling had pulled apart, causing gas to leak into the Hotel's basement.[220] 1976 An LPG/NGL pipeline ruptured near Whitharral, Texas, leading to vapor cloud fire that killed one, severely burning 4 others who later died, destroyed two homes, and burned an area about 400 yards wide. Electrical resistance weld (ERW) seam failure is suspected for the failure. From January 1968 to the date of the Whitharral accident, 14 longitudinal pipe seam failures had occurred on that pipeline system, which resulted in 6 other fatalities, and the loss of over 60,000 barrels (9,500 m3) of LPG.(February 25, 1976) [221][222][223]

1976 An improperly assembled compression coupling failed on a gas distribution line in Phoenix, Arizona on February 8, causing a house explosion that killed 2 people.[224] 1976 On March 27, a two-story building in Phenix City, Alabama, exploded and burned from a gas leak. The explosion and fire killed the six persons in the building. The NTSB found that gas at 20-psig pressure had leaked from a cracked, 3-inch cast iron gas main.[225] 1976 A front loader hit an 8 inch petroleum products pipeline in Los Angeles, California, during a road widening project along Venice Boulevard. 9 people were killed, a plastic factory was destroyed, and other serious property damage occurred. (June 16, 1976)[226][227] 1976 A road grader hit a 20 inch gas transmission pipeline near Calhoun, Louisiana. Six people were killed in the ensuing fire, 6 families were left homeless, and a mobile home and 2 houses were destroyed. (August 9, 1976)[228][229] 1976 On August 13, a flash fire in the basement of a house in Bangor, Maine, occurred while a gas company crew was checking for the cause of low gas pressure at the home. The fire killed one gas company employee, burned two other employees, and caused minor damage to the house. One of the crew was using a match to light the basement of the home, and another crew member was smoking when the fire started. [230] 1976 On August 29, an explosion and fire destroyed a house at Kenosha, Wisconsin. Two persons were killed, four persons were injured, and two adjacent houses were damaged. The destroyed house was not served by natural gas. However, natural gas, which was escaping at 58 psig pressure from a punctured 2inch plastic main located 39 feet (12 m) away, had entered the house through a 6 inch sewer lateral that had been bored through to install the gas line.[231][232] 1976 An explosion and fire at a gas pipeline compressor station in Orange Grove, Texas killed one plant worker, and injured another on December 7.[233] 1977 On January 2, a gas pipeline ruptured and burned near Nursery, Texas. Some power poles were destroyed, but there were no injuries.[234] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

13/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1977 A gas pipeline exploded and burned in Stockton, California on February 4. Another gas pipeline fire had occurred nearby 4 days earlier. There were no injuries.[235] 1977 An explosion on July 8 at Alyeska Pipeline Service Co. Pump Station No. 8 kills one worker, injures 5 others, and destroys the pump station. A US House of Representatives Committee later announced the cause was workers not following the proper procedures, causing crude oil to flow into a pump under repair at the time.[236][237][238][239] 1977 On July 20, the Trans-Alaskan Pipeline was shut down for the 4th time in a month, when it was hit in a valve by a front loader. More than 40,000 US gallons (150,000 L) of crude oil was spilled.[237] 1977 A 12 inch diameter propane pipeline ruptured near Ruff Creek, in Greene County, Pennsylvania, from stress corrosion cracking. The resulting propane vapor cloud ignited when a truck driven into the cloud stalled, then created a spark when it was restarted. Subsidence of underground coal mines in the area may have hastened the failure. (July 20, 1977)[240] 1977 A cast iron gas main broke in Cherokee, Alabama on July 30. Gas migrated into a home through a recently back filled sewer line trench, and exploded 5 days later.[241] 1977 In August, a car drove through the leaking liquid from a petroleum pipeline in Lakewood, California. The pooled liquid appeared to be mud, but it exploded and burned, injuring a woman in the car.[242] 1977 On August 15, crude oil spilled at Alyeska Pipeline Pump Station No. 9. There was no fire, but a fire or explosion at that station could have shut down that pipeline, since Pump Station No. 8 was out of service from the previous month's accident there. This was the seventh accident on this pipeline since the start up of the Alaska pipeline on June 20, 1977. The NTSB released three recommendations on September 9, 1977, to correct certain design and operating deficiencies in the pump rooms of each station of the Alyeska system.[243][244] 1977 On September 5, 2 brothers in a moving truck drove into a vapor cloud from a leak at a gas compressor plant in New Cuyama, California, igniting the cloud. One was killed immediately, and the other died 11 days later.[245] 1977 On September 10, a pipeline rupture spilled 69,000 US gallons (260,000 L) of gasoline into a creek in Toledo, Ohio. Corrosion of the pipeline caused the failure.[246][247] 1977 A gasline inside a building in San Francisco, California leaked and exploded, injuring 7 and heavily damaging that building. Gas repair crews were working on the line at the time.[248] 1977 On October 12, a bulldozer ruptured a propane pipeline near Albany, Georgia, causing nearby train traffic to be halted. The bulldozer engine was left running, nearly igniting the vapors. [249] 1977 A backhoe being used to install a pipeline hit an adjacent 6 inch diameter propane pipeline on November 21 in Hutchison, Kansas. Fire broke out, but there were no injuries.[250] 1977 Construction workers punctured a 12 inch gas pipeline in Atlanta, Georgia, with an I-beam on December 1. No fire or explosion followed, but thousands of people were evacuated from nearby buildings.[251] 1977 A compression coupling joint between a plastic and a steel gas line pulled apart in Lawrence, Kansas on December 15. The gas migrated into 2 buildings, and exploded, killing 2 people, and injuring 3 others.[252][253] 1978 Earth movement was suspected in causing a gas transmission pipeline to rupture and burn near Stevenson, Washington on January 23. There were no injuries.[254] 1978 On February 15, a gas pipeline being tested with compressed air exploded at a seam on the pipe in Cincinnati, Ohio on February 15, injuring 8.[255] 1978 A portion of the Alyeska Pipeline east of Fairbanks, Alaska was ruptured by an explosive device on en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

14/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

February 15. Crude oil spilled in a 600-foot (180 m) diameter spot.[256] 1978 An improperly plugged gas line leaked into service vault in Oklahoma City, Oklahoma at a shopping center, overcoming 5 gas company workers on April 28. Four of the repairmen died of asphyxiation. None of the repair crew had respirators at the job site.[257] 1978 A gas company crew in Mansfield, Ohio accidentally tied a high pressure gas main into a low pressure gas main on May 17. Much higher gas flames in gas appliances caused damage in 16 homes, and about 2,000 gas meters were shut off during the incident.[258] 1978 An LPG pipeline at Donnellson, Iowa, ruptured from past mechanical damage and improper lowering for road improvements. The vapor cloud ignited several minutes after the rupture. Three people were killed and 2 others severely burned. (August 4, 1978)[259][260][261] 1978 On August 7, in Lafayette, Louisiana, natural gas at 15 psig pressure escaped from a corrosion leak in an inactive 1-inch steel service line and migrated beneath a concrete slab and into a building where it ignited. The resulting explosion and fire injured six persons and destroyed the building and its contents.[262] 1978 On August 28, natural gas, which had escaped from a circumferential fracture in a socket heat-fusion coupling on a 2-in. polyethylene (PE) main, operating at 40-psig pressure, migrated beneath a one-story house in Grand Island, Nebraska, exploded, and then burned. One person was injured; the house was destroyed; and three adjacent houses were damaged.[263] 1978 About 7,600 US gallons (29,000 L) of gasoline were spilled in Hampton, Pennsylvania on August 30. Workers boring for a sewer line had hit the fuel pipeline. Later, the 2 construction firms responsible were fined only $500 each.[264][265] 1978 A gas pipeline in Brookside Village, Texas ruptured and exploded, killing five people, and injuring 43 others. Seven mobile homes were also destroyed, (October 24, 1978)[266] 1978 A crude oil pipeline leaks into the Farmington Bay Waterfowl Management Area west of Farmington, Utah on November 8. About 42,000 US gallons (160,000 L) of crude were spilled. The rupture was caused by pumping against a valve that had been closed for earlier pipeline maintenance. [267] 1978 A ruptured 2 inch diameter gasline leaking caused a home to explode in Spokane, Washington on January 6, killing the homeowner.[268] 1979 On January 16, an explosion and fire destroyed five commercial buildings and damaged several other buildings in London, Kentucky. Two persons were injured. External corrosion was suspected as the cause. A prearranged pressure increase in the pipeline was also a factor.[269][270] 1979 An 18 inch diameter natural gas transmission pipeline failed underneath the Florida Turnpike in West Palm Beach, Florida, resulting in a 2 hour road closure.[271] 1979 On April 18, a 24-inch natural gas transmission pipeline pulled out of a compression coupling during a line-lowering project under Iowa State Highway 181 in a rural area near Dallas, Iowa. Within seconds, the natural gas ignited and burned a 900-foot (270 m) by 400-foot (120 m) area. Two cars, a pickup truck, and a trailer housing construction equipment were destroyed. A backhoe was damaged and windows were broken in a nearby farmhouse. Five of the eight injured workers were hospitalized. The gas company's accident records indicated that this 24-inch pipeline had experienced 12 previous failures since it was constructed.[272] 1979 On May 11, 2 explosions and a following fire killed 7 people, injured 19 others, and destroyed 3 buildings in Philadelphia, Pennsylvania. Soil erosion under an 8 inch cast iron gas main caused the main to break and release gas.[273] 1979 On May 13, a 36 inch diameter Colonial Pipeline rupture released 336,000 US gallons (1,270,000 L) of fuel oil that damaged vegetation and killed fish near Spartanburg, South Carolina. Cracks made in the en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

15/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

railroad shipping of the pipe before installation were the cause.[274] 1979 A "spud" dropped by a pile driving barge in the Gulf of Mexico near Pilottown, Louisiana ruptured a 4 inch diameter natural gas pipeline on June 5. The escaping gas ignited, and seriously burned the barge. 4 crew members went missing and were presumed dead.[275] 1979 On June 10, the pilot of a helicopter reported sighting oil on the surface of the Atigun River near the route of the Alyeska Pipeline Service Company's 48-inch crude oil pipeline. Repair crews found a 7 inch crack which passed through a longitudinal weld. Five days after the first leak, at 3:15 p.m. on June 15, the pilot of an Alyeska helicopter on a routine surveillance flight reported a leak north of pump station No. 12 near the Little Tonsina River. A crack near a wrinkle in the pipe was found there. The June 10 spill resulted in a release of approximately 1,500 barrels (240 m3) of crude oil; the June 15 leak resulted in a release of approximately 300 barrels (48 m3) of crude oil; these losses were estimated by Alyeska personnel at the leak site. The spills were too small to be verified by the Alyeska metering system.[276] 1979 On June 16, operator error at Colonial Pipeline causes a rail shipping induced crack section of 36 inch diameter pipeline to rupture in Greenville County, South Carolina. 395,000 gallons of fuel oil were spilled, causing vegetation, fish, & wildlife kills.[277][278] 1979 A leaking pipeline releasing gasoline in Granger, Indiana caused the evacuation of 400 people on July 3.[279] 1979 An anchor handling boat, PETE TIDE II, damaged an unmarked gas pipeline with a grappling hook offshore from New Orleans, Louisiana. Two of the crew were missing and presumed dead in the fire that followed. (July 15, 1979)[280] 1979 On July 25, an explosion and fire destroyed a duplex apartment house in Albuquerque, New Mexico. Two persons were killed, and two persons were hospitalized for burns; adjacent houses were damaged. Earlier in the day, a crew from Mountain Bell Telephone Company (Mountain Bell) had been using a backhoe at the intersection of Bridge Boulevard and Atrisco Road to locate a telephone cable. The backhoe snagged a gas service line but the fact that it was pulled from a 1-inch coupling under the house was not discovered at that time.[281] 1979 A 34 inch diameter Lakehead (now Enbridge) pipeline ruptured near Bemidji, Minnesota, leaking 10,700 barrels (1,700 m3) of crude oil on August 20. The pipeline company initially recovers 60 percent of the spilled oil. Later in 1988, the Minnesota Pollution Control Agency required Lakehead to extract more oil using new technology; removal continued on, with studies still underway in the area.[192][282][283][284] 1979 On August 20, a bulldozer operating near Orange, Texas, began to clean a farm drainage ditch. The corner of the blade cut into a propane line, which crossed beneath the ditch. Propane at 350 psig escaped and was ignited within seconds. The resulting fire killed one person and injured another, and caused considerable property damage.[285] 1979 A crude oil pipeline ruptured and spilled oil into a creek new Walnut Grove, Missouri on August 25. 2 miles (3.2 km) of the creek were contaminated, and 32,000 fish killed.[286] 1979 On September 4, the M/V WHITEFACE struck a high-pressure gas pipeline in Lake Verret, Louisiana. A resulting explosion killed a crewman aboard the vessel.[287] 1979 On October 6, an explosion caused by liquefied natural gas (LNG) vapors destroyed a transformer building at the reception facility of the Columbia LNG Corporation, Cove Point, Maryland. Odorless liquefied natural gas leaked through an inadequately tightened LNG pump seal, vaporized, passed through approximately 210 feet (64 m) of underground electrical conduit and entered the substation building. One person was killed, and one person was seriously injured. Damage to the facility was estimated at about $3 million. The fire hydrants and deluge water spray system were inoperable after the explosion because the en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

16/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

water main that supplied the system was broken at a flange above ground inside the substation.[288] 1979 On October 24, an explosion and fire destroyed the county clerk's office building and the adjoining courthouse building, gutted a connecting building which was under construction, and damaged the adjacent houses in Stanardsville, Virginia. Thirteen persons were injured and property was damaged extensively. The following NTSB investigation revealed that natural gas had leaked from a break in a 1 1/4-inch coated steel service line, which had been snagged by a backhoe which was being used to dig a footing for an addition to the county clerk's office building.[289] 1979 On October 30, a natural gas explosion and fire demolished a townhouse in Washington, D.C., and damaged nearby buildings and cars. No one was inside the townhouse at the time, but three persons in a stopped car were injured when debris from the explosion shattered a car window. After the accident, an inspection of the gas service line that served the townhouse revealed that it had been struck by excavating equipment.[290] 1979 A natural gas transmission pipeline exploded in West Monroe, Louisiana on November 11, causing 3 subdivisions to be evacuated. A gas pipeline explosion had taken place nearby 8 years before.[291]

1980s 1980 On January 2, crude oil leaked from a fractured 22-inch pipeline at a levee crossing at Berwick, Louisiana. At 9:54 a.m., the crude oil ignited. One person was killed, one person was injured, and six homes were either destroyed or damaged. The pipeline's monitoring system failed to detect a loss of over 1,800 barrels (290 m3) of oil. A defective sleeve weld cause the pipeline to fail.[292] 1980 On January 30, an 8-inch-diameter, refined petroleum products pipeline owned by The Pipelines of Puerto Rico, Inc., was struck and ruptured by a bulldozer during maintenance work on a nearby waterline in the Sector Cana of Bayamon, Puerto Rico, about 10 miles (16 km) southwest of San Juan. Gasoline from the rupture sprayed downhill and ran off into a small creek. About 1 1/2 hours later, the gasoline vapors were ignited by an undetermined source and exploded; the subsequent fire killed one person and extensively damaged 25 houses and other property.[293] 1980 On February 21, an explosion and fire destroyed four stores in a shopping complex and severely damaged an adjoining restaurant in Cordele, Georgia. Of the eight persons who were injured, three died later as a result of their injuries. Property damage was extensive. The NTSB investigation of the accident has revealed that natural gas leaked from a 1-inch steel service line, which had been pulled from a 1-inch compression coupling from a backhoe working in the area, and migrated under a concrete slab floor and into a jewelry store where it was ignited by an unknown source.[294] 1980 A Colonial Pipeline Dispatcher ignored established procedures for dealing with a pressure surge, causing a double rupture of a 32 inch steel petroleum products pipeline on March 6. One break, where the pipe had been thinned by corrosion in a casing under a road, caused the release of 8,000 barrels (1,300 m3) of aviation-grade kerosene adjacent to route 234 near Manassas, Virginia. Before being fully contained, the kerosene had flowed into Bull Run River, and had entered the Occoquan Reservoir, a source of drinking water for several northern Virginia communities. The other break, where a crack in the pipe wall initiated during rail shipment of the pipe from the steel mill finally propagated to failure, caused the release of 2,190 barrels (348 m3) of No. 2 fuel oil near Locust Grove, a rural area in Orange County, near Fredericksburg, Virginia. Before being fully contained, the fuel oil had flowed into the Rapidan River and then into the Rappahannock River, a source of drinking water for the City of Fredericksburg.[295][296] 1980 Sabotage during a labor strike was suspected in a gasoline pipeline explosion in Marcus Hook, en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

17/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

Pennsylvania on March 7. The following fire burned for about 17 hours.[297][298] 1980 On April 16, gasoline at a pipeline terminal in Roseville, Minnesota, sprayed from the fractured castiron base of a station booster pump at 72 psig pressure, vaporized, and exploded after it was ignited by the spark of an electric switch in the mainline pump control room 50 feet (15 m) downwind of the booster pump. One man was killed, 3 others injured, and extensive damage was done to the terminal. About 3,500 barrels (560 m3) of petroleum products burned and property damage was estimated at $3 million.[299] 1980 On May 27, near Cartwright, Louisiana, an anhydrous ammonia pipeline was struck by a bulldozer, which was being used to prepare a well site, and the pipeline ruptured. Over 100 people were evacuated from the area.[300] 1980 A road grader ruptured an NGL pipeline in Aurora, Colorado on August 11. Firefighters had barely evacuated residents in the area when the vapors exploded, burning one firefighter.[301] 1980 An oil pipeline ruptured and burned while it was being repaired at an oil storage Terminal in Piney Point, Maryland on September 12, 1980. One worker was killed, and 5 others injured in the fire.[302] 1980 On October 9, a 2-inch-diameter compression coupling located on the upstream side of a gas meter set assembly in the boiler room of the Simon Kenton High School in Independence, Kentucky, pulled out of its connection with a 2-inch-diameter gas service line. Natural gas at 165-psig pressure escaped through the 2-inch-diameter opening and, seconds later, exploded and burned. A basement wall was blown down, an adjacent classroom was damaged, and one student was killed. About 30 minutes later, a second explosion occurred, which injured 37 persons and extensively damaged the school. The gas main was being uprated at the time.[303] 1980 A bulldozer digging a ditch for a new pipeline hit a 16 inch crude oil pipeline near San Ysidro, New Mexico on October 22. The operator was fatally burned.[304] 1980 A pipeline carrying naphtha ruptured under a street in Long Beach, California, causing a fire that destroyed one home and damaged several others. Two people were injured. Lack of communication of pipeline valve setups, and pressure relief valves set to open at too high a pressure were identified by the NTSB as causes of the accident. (December 1, 1980)[305][306] 1980 A dirt pan machine being used for road construction hit a propane pipeline in Sumner, Georgia on December 10, causing slight injuries to the dirt pan operator. US Highway 82 and a rail line were closed, and several families evacuated until the vapors dispersed. There was no fire.[307] 1980 On December 22, a pipeline carrying jet fuel ruptured in Las Vegas, Nevada, spilling fuel for 2 hours. Later, the fuel ignited, forcing road closures. One firefighters was overcome by fumes. Between 50,000 and 100,000 US gallons (380,000 L) of jet fuel were spilled. Prior construction in the area was suspected of damaging the pipeline.[308] 1980 A natural gas pipeline exploded and burned at a gas plant in Ulysses, Kansas on December 28. There were no injuries[309] 1981 An ammonia pipeline leaked near Hutchinson, Kansas on July 31, injuring 5 people, including 3 children at a Bible Camp. A 2-mile (3.2 km) radius from the leak was evacuated, including 90 from the Bible Camp.[310] 1981 On August 25, in downtown San Francisco, California, a 16-inch natural gas main was punctured by a drill that an excavation contractor was using. Escaping natural gas blew upward and carried into the Embarcadero Complex and other nearby buildings. There was no ignition; however, the gas stream entrained an oil containing polychlorinated biphenyl (PCB). Fall-out affected an eight-square-block area of the city's financial district covering buildings, cars, trees, pedestrians, police, and firemen. Approximately 30,000 persons were safely evacuated from the area in 45 minutes. No one was killed or seriously injured, although en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

18/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

many persons were sprayed with the PCB oil mist. There were delays in shutting down the gas, due to inaccurate diagrams.[311] 1981 On September 15, during routine maintenance, a pipeline exploded and burned between a gas plant and a petroleum plant in Goldsmith, Texas. While workers were fighting the fire, another part of the pipeline burst and burned. 6 workers were burned, and another had other injuries. There were a total of 7 fires from 7 pipeline ruptures.[312][313] 1981 A 12 inch diameter pipeline near Ackerly, Texas, was hit by a rathole drill on September 27, releasing an ethane-propane mix. There was then an explosion & fire that killed 4 people.[314] 1981 On November 30, at Flatwoods, West Virginia, gas, leaking into a test section of a 26-inch-diameter gas transmission pipeline, ignited as a welder engaged in installing an end cap placed a tack weld on the east end of a 180-foot (55 m)-long section of pipe. The resultant explosion blew off-the east end cap, which struck and killed the welder's helper.[315] 1981 On December 5, hunters near Yutan, Nebraska tried out a new high power rifle by shooting what they thought was a log in a creek bed. The log was actually an LPG pipeline, and 12 to 16 families needed to be evacuated for their safety from the resulting vapor cloud. There was no fire.[316] 1981 On December 9, a pipeline carrying gasoline ruptured near Joliet, Illinois, spilling 30,000 US gallons (110,000 L) of gasoline into the Des Plaines River.[317] 1981 A gas pipeline in Ottawa, Kansas caused 2 explosions and a raging fire that destroyed 2 mobile homes on December 31. There were no injuries reported.[318] 1982 On January 28 at Centralia, Missouri, natural gas at 47 psig entered a low pressure distribution system which normally operated at 0.40 psig after a backhoe bucket snagged, ruptured, and separated a 3/4-inchdiameter steel pressure regulator control line at a regulator station. The backhoe, which was owned and operated by the city of Centralia, was being used to clean a ditch located adjacent to the pressure regulator station. The high-pressure gas entering customer piping systems in some cases resulted in high pilot light flames which initiated fires in buildings; while in other cases, the pilot light flames were blown out, allowing gas to escape within the buildings. Of the 167 buildings affected by the overpressure, 12 were destroyed and 32 sustained moderate to heavy damages. Five persons received minor injuries.[319] 1982 – An LPG pipeline was ruptured by road construction in North Richland Hills, Texas on April 16. 800 to 1,000 nearby residents were evacuated. There was no fire. The construction crew workers said the pipeline was 5 feet (1.5 m) away from where it was shown on a map they were using.[320] 1982 A backhoe ruptured a 2 inch-diameter gas pipeline in three places in Tacoma, Washington, causing evacuations. There was no fire or explosion.[321] 1982 On June 28, a natural gas explosion demolished a house, killed five persons, and critically injured one person in Portales, New Mexico; the critically injured person died later at a burn treatment center. The gas service line to the house had been damaged 37 days earlier when a contractor's backhoe pulled up the line during conduit excavation work for the local telephone company.[322] 1982 On September 7, natural gas at 15 psig escaping from the open ends of a 2 1/4-inch cast-iron gas main located in a deep, narrow excavation in Dublin, Georgia, was ignited by an unknown source. Four City of Dublin gas department employees who were working in the excavation were critically burned. [323] 1982 – On October 1, a steel plate, which had been welded by a work crew to cap temporarily the open end of a section of a 22-inch diameter gas transmission pipeline, blew off at an initial pressure of possibly 260 psig. Escaping natural gas from the pipeline, which had accumulated due to a leak in a nearby gate valve, ignited almost immediately and the entire work area and a portion of U.S. Route 65 were momentarily engulfed in flames. Seven persons who were working to replace a section of the pipeline under the road en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

19/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

about 2 miles (3.2 km) south of Pine Bluff, Arkansas, were burned.[324] 1982 On October 29, a crew mechanic working on new gas service lines at Burke, Virginia, was overcome by leaking gas and died.[325] 1982 On November 4, a tile plow installing field drainage tile on a farm located 4 miles (6.4 km) west of Hudson, Iowa, struck and punctured a well-marked, 20-inch natural gas transmission pipeline. Natural gas escaping at about 820 psig ignited immediately, and the ensuing fire killed five persons.[326][327] 1982 On December 8, a five-member crew was working on a gas compressor at Bonicord, Tennessee, when a gas explosion occurred. All five crew members were injured seriously, but were able to evacuate the building. One crew member died later that day, and two others died a few days later.[328] 1983 On February 1, a corroded gas service line caused a natural gas explosion and flash fire that destroyed a house, killed two persons, and injured three persons in Pryor, Oklahoma, and damaged an adjacent house.[329] 1983 A gas pipeline failed and caused a fire with flames 250 to 300 feet (91 m) tall near Marlow, Oklahoma on February 15. There were no injuries.[330] 1983 An 8-inch-diameter LPG pipeline was hit by a rotating auger used for planting trees near West Odessa, Texas. After several minutes, the escaping LPG at 1,060 psi ignited, killing 5 people and injuring 5 others. Flames went as high as 600 feet into the air.(March 15, 1983)[331][332][333] 1983 On March 27, a pump for a petroleum products pipeline broke, causing up to 420,000 gallons of diesel fuel to spill into the Bowie River in Collins, Mississippi.[334][335] 1983 A 36-inch-diameter gas transmission pipeline exploded and burned in Caldwell, Ohio on May 21, destroying two homes, burning 100 acres of vegetation, and closing nearby Interstate 77. There were three minor injuries.[336] 1983 On June 4, a front loader accidentally dug into a 10-inch-diameter petroleum pipeline near Coeur d'Alene, Idaho, spilling over 20,000 US gallons (76,000 L) of unleaded gasoline into a creek, killing everything downstream for 3 miles.[337] 1983 A 16 inch diameter gas pipeline ruptures and burned near Athens, Texas, on July 19. A nearby section of the same pipeline had ruptured the year before.[338] 1983 On September 23, gas service pressure surged up in a section of Boston, Massachusetts. 3 major structure fires, numerous smaller fires, and an explosion at a restaurant followed. There was no serious injuries. A flooded gas regulator vault was the cause.[339] 1983 A crude oil pipeline exploded and burned at an Oil Terminal in Lima, Ohio on December 26. [340] 1984 An 8 inch NGL pipeline near Hurst, Texas, was hit by a front loader, and the escaping gases ignited, causing burns to the equipment operator. (February 28, 1984)[341] 1984 On June 19, six employees of a contractor working for Washington Gas Light Company (WGL) in Rockville, Maryland, were using mechanical saws to cut a section of 22 inch diameter steel pipeline when residual gas at atmospheric pressure in the isolated section of the pipeline was ignited. A flash fire ensued, and four contractor employees who were operating the saws and a WGL superintendent were burned. [342] 1984 Two natural gas pipelines exploded and burned near Falls City, Texas.[343] 1984 On September 24, a failed gas main of ABS plastic caused an explosion and fire in Phoenix, Arizona. 5 people died and 7 others injured in the accident. Liquid in the pipe had caused it to break down.[344][345] 1984 A tugboat hit and ruptured a gas pipeline on the Houston Ship Channel on October 16. There were no injuries, but the Channel was closed for a time.[346] 1984 Fast moving water in the Cado Creek near Durant, Oklahoma led to 2 pipelines being ruptured on en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

20/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

October 27 & 28. About 1,500 barrels (240 m3) of petroleum were spilled.[347] 1984 On November 25, a 30-inch gas transmission pipeline, constructed in 1955 and operating at 1,000 psig pressure, ruptured at a location about three miles (5 km) west of Jackson, Louisiana. Gas blowing from the rupture fractured the pipe into many pieces and created a hole in the earth about 90 feet (27 m) long, 25 feet (7.6 m) wide, and 10 feet (3.0 m) deep. The escaping gas was quickly ignited by one of several potential sources of ignition. The resulting fire incinerated an area extending from the rupture about 950 feet (290 m) north, 500 feet (150 m) south, and 180 feet (55 m) to the east and to the west. Within this sparsely populated area, five persons involved with the pipeline construction work were killed, and 23 persons were injured. Additionally, several pieces of construction equipment were damaged extensively. Lack of proper ground support under the pipeline when a nearby section of that pipeline was upgraded and replaced was identified as a factor in the failure.[348][349][350] 1985 Natural gas from a leaking line traveled through soil and caused a massive gas explosion in El Paso, Texas on January 8. Eleven people were injured, 2 homes were destroyed, and 88 other homes were damaged by the blast.[351] 1985 On February 22, 1985, a police patrolman on routine patrol smelled strong natural gas odors in Sharpsville, Pennsylvania. A gas serviceman was ordered to the scene. Before the serviceman arrived at the site of the reported leak, the Sharpsville Inn and a connecting building exploded and burned, killing two persons. Firefighters arriving on scene moments later encountered a second, smaller explosion, which injured one firefighter. The delay in the gas serviceman getting to the incident was a contributing factor. [352] 1985 On April 5, a lightning caused computer malfunction resulted in a pipeline rupture that sent thousands of gallons of gasoline into the Yellow Leaf Creek in Alabama.[353] 1985 A 30 inch diameter gas pipeline operating at about 960 psi, weakened by atmospheric corrosion, ruptured, and tore out about 29 feet (8.8 m) of the carrier pipe, blew apart about 16 feet (4.9 m) of a 36inch-diameter casing pipe, blasted an opening across Kentucky State Highway 90, and cut out a pearshaped crater approximately 90 feet (27 m) long, 38 feet (12 m) wide, and 12 feet (3.7 m) deep near Beaumont, Kentucky. 5 people were killed in one home, and 3 injured. The fireball from the incident could be seen 20 miles away.(April 27, 1985)[354][355] 1985 Workers on the extension of the North Dallas Tollway ruptured a 12 inch gasoline pipeline on June 19, causing a massive gasoline spill along a creek bed north of Dallas, Texas. The gasoline later ignited. One person had moderate injuries, several office buildings were damaged by fire, and some automobiles were damaged.[356] 1985 On July 23, in a rural area about 8 miles (13 km) south of Kaycee, Wyoming, a girth weld cracked during a pipeline re-coating project on a 23-year-old, 8-inch-diameter pipeline. The cracked girth weld allowed the release, atomization, and ignition of aircraft turbine fuel under 430 pounds pressure, killing one person, burning six persons, destroying construction equipment.[357][358] 1985 A gasoline leak of up to 42,000 US gallons (160,000 L) from a ruptured 10 inch pipeline ignited on August 2 in Indianapolis, Indiana, causing a 200-foot (61 m) high fireball that killed three people, and injured 3 others working to clean up the spill along a creek.[359][360] 1985 On September 23, a 12 inch diameter gasoline pipeline fitting was hit by a backhoe, and sprayed about 35,000 US gallons (130,000 L) of gasoline 45 feet (14 m) into the air in Staten Island, New York. There were evacuations, but no fire.[361] 1985 On December 6, a natural gas explosion and fire destroyed the River Restaurant in Derby, Connecticut. Gas escaping from a broken gas main at a pressure of about 1 pound per square inch had escaped, migrated into the restaurant basement, ignited, exploded, and burned. Of the 18 persons inside the en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

21/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

restaurant at the time, 6 were killed and 12 were injured; 1 passerby and 1 firefighter were also injured. After the accident the street adjacent to the restaurant was excavated where a 24 inch diameter sewer system had just been installed; An 87-year-old, 3-inch diameter, cast-iron natural gas main was found broken.[362] 1986 A 30 inch gas pipeline ruptured due to corrosion near Lancaster, Kentucky. 3 people had serious burns, and 5 others had lesser injuries. External corrosion made worse by difficulties of Cathodic protection in rocky soil was the cause. (February 21, 1986)[363][364] 1986 An 8 inch high-pressure petroleum pipeline ruptures in Muskegon County, Michigan on February 22, spilling gasoline into creeks.[365] 1986 A backhoe snagged a natural gas distribution line in Fort Worth, TX, causing a break that leaked gas into an unoccupied building. Later, that building exploded, injuring 22 people, destroying the unoccupied building, and damaging 40 other buildings. 57 automobiles in the unoccupied building were damaged or destroyed. (March 12, 1986)[366] 1986 A new water main was being installed in Chicago Heights, Illinois on March 13. While excavating, an active gas service line was snagged. Gas company crews responded to the wrong site, causing delays in getting the leaking gas line shut down. Just as crews finished closing the valve on the leaking line, the nearby house exploded and began to burn; one of the two persons inside this house was killed, and the other was injured. Two neighboring houses were damaged, and one gas company employee, two construction crew members, and four persons in the general area were injured by the explosion and subsequent fire. Although gas company personnel arrived on the scene approximately 10 minutes before the explosion and shut off the gas at the meter, neither they nor the contractor's crew had made an effort to warn or evacuate the residents of the house.[367] 1986 On June 28, a pipeline ruptured and spilled diesel fuel into Trail Creek in Michigan City, Indiana. The fuel later ignited. Thousands of fish were killed.[368] 1986 Early on July 8, a Williams petroleum products pipeline ruptured in Mounds View, Minnesota. Gasoline at 1,434 psi sprayed a residential area around 4:20 am local time, then ignited. A woman and her 7 year old daughter suffered fatal burns, at least two others were injured, and many homes damaged or destroyed. Confusion by the pipeline company led to a delay in shutting down the pipeline. Electrical resistance welded (ERW) seam failure caused the rupture. During a hydrostatic test of this pipeline following the accident, 7 ERW seams failed. Studies of available data by OPS staff in early 1988 showed that ERW seams have been involved in 145 service failures in both hazardous liquid and natural gas pipelines since 1970 to early 1988, and that of these failures, all but 2 occurred on pipe manufactured prior to 1970.[369][370][371][372][373][374] 1986 A gas transmission pipeline fails and burns in a compressor station near Prattville, Alabama, on July 12. The fire spread by melting flange gaskets on 2 other gas transmission pipelines in the station. 4 homes and several cars were destroyed in the following fire, with flames reaching 300 feet (91 m) high. There were no injuries.[375] 1986 between 800 and 1200 residents were evacuated in East Chicago, Indiana after a gasoline tank at a pipeline Terminal ruptured on September 4. 28 people were overcome by gasoline fumes. There was no fire.[376] 1986 A petroleum products pipeline failed near Billings, Montana, causing the evacuation of nearby businesses. There was no fire.[377] 1986 On September 8, a pipeline failed under the Red River near Gainesville, Texas. Fumes from the pipeline sent 14 to hospitals for treatment.[378] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

22/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1986 A 30 inch diameter natural gas pipe line under Pelahatchie Bay in Mississippi exploded near the water's edge on the north shore of the bay on December 6.[379] 1986 On December 25, a pipeline ruptures and spills furnace oil into the Des Plaines River near Chicago, Illinois. About 1,000 US gallons (3,800 L) of furnace oil was spilled. Corrosion seemed to cause the pipeline failure.[380] 1987 A petroleum pipeline ruptured and burned near Corsicana, Texas on March 12, forcing the closure of an Interstate highway, and cause some evacuations.[381] 1987 A work crew burning the remains of a house near Ladysmith, Virginia ruptured a nearby petroleum products pipeline with a bulldozer on March 26, igniting diesel fuel from the line. 2 of the worker were injured.[382] 1987 On April 4, an LPG pipeline exploded at a Terminal in Iowa City, Iowa. Due to the fire spreading to a pipeline for nearby underground gas storage, residents within a 2 1/2 mile radius of the Terminal were evacuated for a time. The fire burned until April 20. The cause was an ERW seam failure in a pipeline. During a hydrostatic test of that pipeline, 20 more pipeline segment seams failed.[383][384][385] 1987 On June 11, a "rock ripper" at a construction site punctured a 32 inch petroleum products pipeline in Centreville, Virginia. Gasoline sprayed from the rupture, but there was no fire. More than 15,500 US gallons (59,000 L) of gasoline were released. Thirteen emergency response personnel suffered from exposure to the gasoline fumes.[386] 1987 In July, a fishing vessel, working in shallow waters off Louisiana, the menhaden purse seiner Sea Chief, struck and ruptured an 8" natural gas liquids pipeline operating at 480 psi. The resulting explosion killed two crew members. Divers investigating found that the pipe, installed in 1968, was covered with only 6" of soft mud, having lost its original 3-foot (0.91 m) cover of sediments.[387] 1987 On July 23, a construction crew working on an Interstate 90 project east of Coeur d'Alene, Idaho struck the 10 inch diameter Yellowstone Pipeline, causing a leaking that sprayed out over 200 barrels (32 m3) of gasoline. The pipeline was supposed to have 30 inches of soil cover, but had only 2 inches of cover. There was no fire.[388] 1987 A gas leak on a busy road in Wilmington, North Carolina suddenly ignited while gas company workers were trying to plug that leak, burning them and firefighters on standby nearby in August. 19 people were burned, with a fire department Assistant Chief later dying from the burns he received.[389] 1988 On January 5, a Colonial Pipeline mainline ruptured, spilling about 100,000 gallons of home heating oil in Deptford, New Jersey. The cause of the pipeline failure was corrosion.[390][391] 1988 The rupture of a large interstate gas line at Pocono Ridge development in Lehman Township, Pennsylvania, left a crater about 8 feet (2.4 m) deep and ejected a 6-foot (1.8 m) section of pipe over the treetops before it landed 50 yards away. One hundred thirty people were evacuated. No one was injured.[215] 1988 On January 18, a natural gas explosion destroyed the building housing the K&W Cafeteria and the lobby of the Sheraton Motor Inn at Winston-Salem, North Carolina. Two adjoining motel wings suffered structural damage. Of the four persons in the lobby/cafeteria building at the time of the explosion, three sustained minor injuries. The fourth person sustained a fractured ankle. One motel guest also sustained minor cuts.[392] 1988 On February 8, an offshore pipeline near Galveston, Texas, that may have been damaged by an anchor, ruptures, spilling about 15,576 barrels (2,476.4 m3) of crude oil into the Gulf.[393] 1988 On April 9, a 20 inch diameter crude oil pipeline failed in a Peoria County, Illinois subdivision. About 200,000 US gallons (760,000 L) of crude were spilled, contaminating 2 private lakes.[394] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

23/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1988 A pair of MAPCO LPG/NGL pipelines failed in an explosion south of Topeka, Kansas on July 22. 200 nearby residents had to be evacuated, and there was serious damage to US Route 75 nearby from the explosion & following fire. An ERW seam selective corrosion failure in one of the pipelines caused the failure.[395][396] 1988 On August 31, a gas company crew struck and ruptured a fitting on a 4-inch plastic gas main in Green Oaks, Illinois. While the crew was attempting to excavate a nearby valve to shut off the flow of gas, the backhoe struck an unmarked power cable. The gas ignited and four gas company employees were injured.[397] 1988 A crude oil pipeline ruptured, spilling about 132,000 gallons of crude oil in Encino, California on September 10. The crude flowed into storm drains, and then into the Los Angeles River. Electrical interference to Cathodic protection from other pipelines was suspected to have cause the corrosion that caused the failure. The crude oil pipeline was on top of a steel water pipeline, which would directly interfere with Cathodic protection efforts.[398][399][400][401][402] 1988 On September 16, a natural gas explosion in Overland Park, Kansas, involved gas leaking from corrosion holes in the customer-owned line. Gas migrated underground to the house and was ignited. The house was destroyed and the four residents were injured.[403] 1988 In November, corrosion of a 14-inch underground pipeline owned and operated by the Shell Oil Company, a predecessor of Shell Pipeline Corporation (Shell), resulted in the release of an estimated 120,000 US gallons (450,000 L) of gasoline. A pool of gasoline about 450 feet (140 m) by 50 feet (15 m) appeared among fields of corn and soybeans. The site of the release was in Limestone Township in Kankakee County, about 4 miles (6.4 km) west of Kankakee, Illinois. Approximately 2,100 people live within a 1-mile (1.6 km) radius of the November 1988 release point.[404] 1988 On November 25, natural gas explosion and fire in Kansas City, Missouri, involving a break in a customer owned service line at a threaded joint that was affected by corrosion. One person was killed and five persons injured in the explosion that severely damaged the residence.[403] 1988 A Koch Industries and Ashland Oil subsidiary 16 inch diameter crude oil pipeline failed near Dellwood, Minnesota, spilling about 200,000 US gallons (760,000 L) of crude on a farm. Snow complicated the cleanup. The leak occurred late December 1, but was not discovered until early December 2. An ERW seam fatigue crack caused the failure.[405][406][407] 1988 A 22 inch diameter crude oil pipeline ruptures near Vienna, Missouri on December 24, spilling more than 860,000 US gallons (3,300,000 L) of crude oil into the Gasconade River. A pipeline worker in Oklahoma failed to notice the pipeline's plummeting pressure gauges for at least two hours. An ERW seam defect in the pipe was determined to be the cause of the failure.[408][409] 1989 A crude oil pipeline rupture on January 24 in Winkler County, Texas spills over 23,000 barrels (3,700 m3) of oil. 6 acres (24,000 m2) of land were covered in oil, and groundwater was contaminated.[410] 1989 February 10, a natural gas explosion and fire in Oak Grove, Missouri, involved the failure of a customer owned service line at a threaded joint. Two persons were killed and their house was destroyed in the explosion.[403] A leaking gas distribution line caused an explosion in Topeka, Kansas on March 25, killing one person. This was the latest in a string of gas distribution line failures that lead to an NTSB investigation into the regional gas company. 600,000 gas services lines were replaced as a result of the investigation.[411] 1989 On May 25, a petroleum products pipeline failed, after the San Bernardino train disaster, California. Damage from a train derailment cleanup caused a CalNev petroleum products pipeline to rupture, spraying nearby homes with gasoline. Three were killed, 31 were injured, and 15 homes were damaged or destroyed en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

24/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

in following fire.[412][413][414] 1989 A crude oil gathering pipeline ruptures near Craig, Colorado on June 2, spilling 10,000 US gallons (38,000 L) of crude into the Yampa River. Federal maintenance oversight of gathering pipelines ended in 1985.[415] 1989 On October 3, the United States menhaden' fishing vessel NORTHUMBERLAND, owned and operated by the Zapata Haynie Corporation (vessel owner), was backing and maneuvering in 9 to 11 feet (3.4 m) of water when the stern of the vessel struck and ruptured an offshore 16 inch natural gas transmission pipeline. Natural gas under 835 pounds per square inch pressure was released. An undetermined source on board the vessel ignited the gas, and within seconds, the entire vessel was engulfed in flames. The fire on the vessel burned until 4:30 a.m. on October 4, when it burned itself out. Leaking gas from the pipeline also continued to burn until the flow of gas subsided and the fire self-extinguished about 6 a.m. on October 4. Eleven of fourteen crew members died as a result of the accident.[416] 1989 An explosion at a valve in a natural gas processing station on October 25 near Evanston, Wyoming kills one worker, and injures 4 others.[417] 1989 A farmer hit a propane pipeline near Butler, Illinois on December 8, forcing evacuation of that town. There was no fire.[418] 1989 On December 18, a Colonial Pipeline petroleum pipeline failed near Locust Grove, Virginia. 212,000 US gallons (800,000 L) of kerosene spilled into the Rapidan and Rappahannock Rivers. On New Year's Eve, following a rapid thaw and heavy rains, containment dams broke and kerosene flowed downstream toward Fredericksburg, Virginia. Fish and game were killed, the City's water supply was cut off, and drinking water had to be hauled from Stafford County for seven days. This was the seventh major leak from Colonial Pipeline in Virginia since 1973.[419][420] 1989 New York City Con Edison Steam Pipe explosion, rupture killing three people in the 3rd Ave./Gramercy Park area.

1990s 1990 On January 2, an Exxon underwater pipeline located at the mouth of Morse Creek discharged approximately 13,500 barrels (2,150 m3) of No. 2 heating oil into the Arthur Kill waterway between New Jersey and Staten Island, New York.[421][422] 1990 A propane pipeline ruptured and burned, near North Blenheim, New York, on March 13. Stress from previous work done on the pipeline caused a pipeline rupture and vapor cloud that moved downhill into a town. Two people were killed, seven persons injured, and more than $4 million in property damage and other costs resulted when the cloud ignited.[423][424][425] 1990 on March 30, a 10-inch-diameter pipeline, ruptured from overstress due to a landslide in Freeport, Pennsylvania, resulting in the release of approximately 1,300 barrels (210 m3) of mixed petroleum products. Spilled petroleum products entered Knapp's Run, a small creek emptying into the Allegheny River and, eventually, the Ohio River. The product release resulted in extensive ground and water pollution and interrupted the use of the Allegheny River as a water supply for several communities. Damage to the pipeline and environmental cleanup and restoration costs exceeded $12 million.[386] 1990 On May 6, a spool on a pipeline ruptured off of the Louisiana coast. 13,600 barrels (2,160 m3) of crude oil were estimated to have spilled.[426] 1990 On August 29, a private contractor laying conduit for underground power lines ruptured a pipeline that fouled a Western Branch creek with diesel fuel in Chesapeake, Virginia. Over 67,000 US gallons (250,000 en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

25/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

L) of fuel were spilled.[427] 1990 On August 29, a natural gas explosion and fire destroyed two row houses and damaged two adjacent houses and three parked cars in Allentown, Pennsylvania. One person was killed, and nine people, including two firefighters, were injured. A cracked gas main, that was stressed by soil erosion from a nearby broken water line, was the cause of the gas leak.[428][429] 1990 On November 5, a crude oil pipeline ruptures near Ethel, Missouri, fouling over 35 miles (56 km) of the Chariton River. 44,000 to 66,000 US gallons (250,000 L) of crude were spilled.[430] 1990 At least 3 leaks that spill over a thousand gallons of oil were found in a pipeline in Cerritos, California, it was announced on November 23. One of the failed section of pipeline was 6 to 7 years old.[431] 1990 On December 9, a gas system valve between one of Fort Benjamin Harrison Indianapolis, Indiana gas distribution systems and a discontinued steel gas system segment was inadvertently opened, allowing natural gas to enter residential buildings that had previously received their gas from the discontinued segment. Gas accumulating in Building 1025 of Harrison Village was ignited by one of many available sources, and the resulting explosion killed 2 occupants and injured 24 other persons One building was destroyed, and two were damaged,[432] 1991 On January 31, a Mobil Company crude oil pipeline ruptured near Valencia, California, spilling up to 75,000 gallons of crude oil. The same day, a report was released showing that particular pipeline had a 99.8% chance of a leak in the next 5 years.[433][434][435] 1991 A Lakehead (now Enbridge) crude oil pipeline near Grand Rapids, Minnesota ruptured on March 2. More than 40,000 barrels of crude went into the Prairie River. About 4 million US gallons (15,000 m3) of oil had spilled from that pipeline from the early 1970s to 1991, per Minnesota records. A resident in the area noticed the smell of oil and alerted the local fire department. Approximately 300 people living in homes near the site were evacuated for safety, but were allowed to return to their homes later in the night.[192][436][437][438][439][440] 1991 A ruptured propane pipeline on March 2 forced the evacuation of 2,500 from several subdivisions in Richland County, South Carolina for a time. There was no fire.[441] 1991 On June 29, over 60,000 US gallons (230,000 L) of fuel oil and gasoline leaked from a 10 inch diameter Koch Industries pipeline in Carson, Wisconsin from a 3 inch crack. A previous significant leak had occurred on this pipeline in that area the year before. Local officials urged Koch to upgrade it's leak monitoring equipment. Koch later replaced 12 miles (19 km) of that pipeline in the area[442][443][444][445][446] 1991 On July 17, workers were removing a corroded segment of the Consumers Power Company’s (CP) 10-inch-diameter transmission line pipeline in Mapleton, Michigan. As a segment of the pipeline was being removed, natural gas at 360-psig pressure exerted about 12 tons of force on an adjacent closed valve (H143), causing it and a short segment of connected pipe to move and separate from an unanchored compression coupling. The force of the escaping gas killed one worker (a welder), injured five other workers, and collapsed a steel pit that housed valve H-143.[447] 1991 About 42,000 US gallons (160,000 L) of crude oil spilled from a broken pipeline at a barge facility at High Island, Texas on September 5.[448] 1991 On December 19, a 36-inch-diameter Colonial Pipeline ruptured from prior excavation damage about 2.8 miles (4.5 km) downstream of the pipeline's Simpsonville, South Carolina, pump station. The rupture allowed more than 500,000 US gallons (1,900,000 L) of diesel fuel to flow into Durbin Creek, causing environmental damage that affected 26 miles (42 km) of waterways, including the Enoree River, which flows through Sumter National Forest. The spill also forced Clinton and Whitmire, South Carolina, to use alternative water supplies.[386][449] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

26/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1991 On December 28, two explosions in rapid succession occurred in apartment No. 3 of a two-story, eight-apartment, wood-frame structure in Santa Rosa, California. Two people were killed and three others were injured. Fire after the explosions destroyed that apartment and three other apartments in the front of the building.[450] 1992 A pipeline offshore of Grand Island, Louisiana in the Gulf of Mexico leaked thousands of gallons of crude oil.[451] 1992 On January 17, while a gas company crew was doing routine annual maintenance work at a regulator stations in Chicago, Illinois, high-pressure gas entered a low-pressure system. The gas—under as much as 10 psig of pressure—escaped through gas appliances into homes and other buildings, where it was ignited by several unidentified sources. The resulting explosion and fires killed 4 people, injured 4, and damaged 14 houses and 3 commercial buildings.[452] 1992 On April 7, a salt dome cavern used to store LPG & similar products was overfilled, leading to an uncontrolled release of highly volatile liquids (HVLs) from a salt dome storage cavern near Brenham, Texas, formed a large, heavier-than-air gas cloud that later exploded. Three people died from injuries sustained either from the blast or in the following fire. An additional 21 people were treated for injuries at area hospitals. Damage from the accident exceeded $9 million.[453][454][455] 1992 A natural gas explosion destroyed a house in Catskill (town), New York, on November 6. The twostory wood-frame house had not had active gas service since 1969. The explosion killed a woman in the house, seriously injured her daughter, and slightly injured two children in a neighboring house. Gas had escaped from a nearby cracked gas main.[456] 1992 On December 3, a ruptured natural gas liquid pipeline caused a vapor cloud to drift across I-70 near Aurora, Colorado. The Cloud later ignited, burning 6 motorists.[457][458][459] 1993 On March 28, a pressurized 36-inch-diameter (910 mm) petroleum product pipeline owned and operated by Colonial Pipeline Company ruptured near Hemdon, Virginia. The rupture created a geyser which sprayed diesel fuel over 75 feet (23 m) into the air, coating overhead powerlines and adjacent trees, and misting adjacent Virginia Electric Power Company buildings. The diesel fuel spewed from the rupture into an adjacent storm water management pond and flowed overland and through a network of storm sewer pipes before reaching Sugarland Run Creek, a tributary of the Potomac River. The cause was latent third party damage.[460][461] 1993 On April 6, a crude oil pipeline ruptured & spill up to 125,000 gallons of crude oil into a stream bed in Kern County, California, forcing a temporary closure of the nearby Golden State Freeway. [462] 1993 On June 9, a cinder block duplex at in Cliffwood Beach, New Jersey, exploded as a New Jersey Natural Gas Company (NJNG) contractor was trenching in front of the building. The explosion killed 3 residents of the duplex, and seriously injured 3 others.[463] 1993 On July 22, a city of St. Paul Department of Public Works backhoe hooked and pulled apart a Northern States Power Company (NSP) high-pressure gas service line in St. Paul, Minnesota. An explosion and natural gas-fueled fire resulted about 20 minutes after the backhoe hooked the service line. The explosion force caused part of the building to land on and flatten an automobile traveling southwest on East Third Street, and the driver died instantly. The explosion and ensuing fire also killed an apartment occupant and a person outside the building and injured 12 people.[464] 1993 On July 26, a 6-inch pipeline in Nebraska was exposed by scour in a creek bed and its banks, and was struck by flood debris, which caused it to rupture. The rupture resulted in the release of 2,203 barrels (350.2 m3) of anhydrous ammonia [465] 1993 An ammonia pipeline failed in Sperry, Oklahoma on August 20. 80 homes in the area were evacuated. en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

27/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

Several people were treated for ammonia inhalation injuries.[466] 1993 On November 3, Amoco Pipeline was fined $12,500 for a 1971 pipeline leak that contaminated a drinking well and caused other pollution problems for people living near Garfield, Minnesota. [467] 1993 On December 2, a 10 inch diameter Conoco pipeline ruptured, spilling 8,400 US gallons (32,000 L) of gasoline into a creek in Washington, Missouri.[468] 1993 An explosion and fire on a gas transmission pipeline on December 20, near Mellen, Wisconsin, cut off the gas supply to 3,500 customers in the area.[469] 1994 In January, a pipeline ruptures, dumping almost 162,500 US gallons (615,000 L) of oil in a river, the Marais des Cygnes River in Osawatomie, Kansas. In addition to a $804,700 fine, BP Amoco agreed to spend at least $145,300 on a supplemental environmental project involving reconstruction improvements to Osawatomie's water intake.[470] 1994 On February 1, the third explosion in 7 years hit a LPG/NGL pipeline Terminal in Iowa City, Iowa. 11 workers at the Terminal escaped injury, and 6 families within 1 1/2 miles of the Terminal were evacuated. The 2 previous explosions were in 1987 and 1989.[471] 1994 The Texas Eastern Transmission Corporation Natural Gas Pipeline Explosion and Fire : Previous damage caused a 36 inch diameter natural gas transmission pipeline to rupture at Edison, New Jersey on March 23, 1994. Several apartment buildings were destroyed in the massive fire. One woman died of a heart attack, and at least 93 others had minor injuries. Delays in shutting off one of the pipeline's valves was cited as contributing to the damage.[472] 1994 A 2-inch-diameter steel gas service line that had been exposed during excavation separated at a compression coupling about 5 feet (1.5 m) from the wall of a retirement home in Allentown, Pennsylvania on June 9. The escaping gas flowed underground, passed through openings in the building foundation, migrated to other floors, and exploded. The accident resulted in 1 fatality, 66 injuries, and more than $5 million in property damage.[473] 1994 A residents near O'Fallon, Missouri detected a petroleum smell, early on September 22. The local Fire Department was called several hours later, and noticed an oily mist in the area, and found a leaking pipeline. The owner of the 10 inch diameter petroleum products later claimed the spill volume was less than 1,000 US gallons (3,800 L), but later calculations and batch volume measurements indicate a spill of 29,000 to 37,000 US gallons (140,000 L). EPA officials later admit someone lied about the spill volume. Over the next 10 years, 8 attempts at remediation were made, before the O'Day Creek was cleaned of all petroleum products.[474][475] 1994 On October 8, a lightning strike shut a valve on a crude oil pipeline, while the oil was flowing, triggering a pressure buildup that ripped a 50-square-inch hole in a section of the pipe that was already weakened by corrosion. Pipeline employees — unaware of the rupture in the pipe — turned on the pumps after the pipeline shut down automatically, sending oil pouring into the creek for about an hour. The spill created a 12mile (19 km)-long slick on Nueces and Corpus Christi bays along the Texas Gulf Coast. Nearly seven years later, delicate coastal marshes that serve as a nursery for shrimp, flounder, crabs and other marine life have not fully recovered. The estimated spill size was 2,151 barrels (342.0 m3), but that was debated as being too small a size. The pipeline eventually agreed to pay more than $45 million in damages. [476] 1994 In October, record high flooding along the San Jacinto River in Texas lead to the failure of 8 pipelines crossing that river. Due to the flooding many other pipelines were also undermined. More than 35,000 barrels (5,600 m3) of petroleum and petroleum products were released into the river. Ignition of the released products resulted in 547 people receiving (mostly minor) burn and inhalation injuries. Spill response costs exceeded $7 million, and estimated property damage losses were about $16 million.[477][478] en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

28/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

1994 A natural gas explosion and fire destroyed a one-story, wood frame building in Waterloo, Iowa on October 17. The force of the explosion scattered debris over a 200-foot (61 m) radius. 6 persons inside the building died, and one person sustained serious injuries. 3 persons working in an adjacent building sustained minor injuries when a wall of the building collapsed inward from the force of the explosion. The explosion also damaged nine parked cars. A person in a vehicle who had just exited the adjacent building suffered minor injuries. Additionally, two firefighters sustained minor injuries during the emergency response. Two other nearby buildings also sustained structural damage and broken windows.[479][480] 1994 A leak of at least 20,000 US gallons (76,000 L) of diesel fuel was discovered on a Koch Industries pipeline near Plover, Wisconsin on November 29. The leak brought this pipeline's total spill volume to 100,000 US gallons (380,000 L) on a 91 miles (146 km) pipeline section through several years.[481] 1995 Since starting operations in 1954 until 1995, Yellowstone Pipeline had 71 leaks along the Flathead Indian Reservation in Montana, spilling 3,500,000 US gallons (13,000,000 L) of petroleum products. Eventually, the Flathead refused to sign a new lease with Yellowstone.[482] 1995 A 26 inch diameter gas transmission pipeline ruptured and burned near Castle Rock, Washington on March 6. There were no injuries.[483] 1995 On March 20, a natural gas transmission pipeline leaked and burned near Chipola, Louisiana. There were no injuries reported.[484] 1995 On March 27, a bulldozer operator ruptured a 40 inch diameter gas transmission pipeline in Huntersville, North Carolina, causing an explosion. The operator was knocked off the bulldozer, then was run over by the driverless bulldozer.[485] 1995 On December 2, 3 contractors were killed, and another injured, when a vacuum used to control flammable fumes accidentally reversed during welding at a pipeline facility near McCamey, Texas. [486] 1995 A bulldozer hit a 16 inch diameter gas pipeline in North Attleboro, Massachusetts on December 9, forcing evacuations of a nearby shopping mall. An estimated 40,000 people were evacuated.[487][488] 1995 On December 19, a gas explosion at a twin dwelling in Norristown, Pennsylvania, killed 2 people and injured another person. Gas had migrated from a crack in a 6 inch cast iron gas main in the street. [486] 1996 A gas pipeline failure excised a 30-foot (9.1 m) section of pipe, and the gas later ignited, causing a vegetation fire in East Stroudsburg, Pennsylvania on January 6. Later inspections found numerous flaws on this pipeline.[489] 1996 On February 5, a pipeline ruptured and spilled diesel fuel into a creek in Fairview Heights, Missouri.[490] 1996 A 20 inch diameter pipeline ruptured at a location near Gramercy, Louisiana, on May 23, 1996. The ruptured pipeline ultimately released about 475,000 US gallons (1,800,000 L) of gasoline into a common pipeline right-of-way and marsh land. Gasoline also entered the Blind River, causing environmental damage and killing fish, wildlife, and vegetation in the area. The pipeline controller did not at first recognize the pipeline had failed, and continued to ignore alarms from the pipeline SCADA system.[491] 1996 A 36 inch diameter Colonial Pipeline ruptured at the Reedy River, near Fork Shoals, South Carolina, June 26. The ruptured pipeline released about 957,600 US gallons (3,625,000 L) of fuel oil into the Reedy River and surrounding areas. The spill polluted a 34-mile (55 km) stretch of the Reedy River, causing significant environmental damage. Floating oil extended about 23 miles (37 km) down the river. Approximately 35,000 fish were killed, along with other aquatic organisms and wildlife. The estimated cost to Colonial Pipeline for cleanup and settlement with the State of South Carolina was $20.5 million. No one was injured in the accident. The pipeline was operating at reduced pressure due to know corrosion issues, but pipeline operator confusion led to an accidental return to normal pressure in that pipeline section, causing en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

29/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

the rupture.[492][493] 1996 On August 4, 420,000 US gallons (1,600,000 L) of unspecified petroleum product spilled from a Lakehead pipe near Donaldson, Minnesota.[436] 1996 A Koch butane pipeline ruptured, causing an explosion and fire, near Kemp, Texas, on August 24. Two teenagers were killed after driving into the unseen butane cloud while going to report the pipeline leak. A mobile home was also destroyed by the fire. The leak was caused by external corrosion. The pipeline was only 15 years old at the time.[494][495][496][497] 1996 On October 23, in Tiger Pass, Louisiana, the crew of a Bean Horizon Corporation dredge dropped a stern spud into the bottom of the channel in preparation for dredging operations. The spud struck and ruptured a 12-inch-diameter submerged natural gas steel pipeline. The pressurized natural gas released from the pipeline enveloped the stern of the dredge and an accompanying tug, then ignited, destroying the dredge and the tug. No fatalities resulted from the accident.[498][499] 1996 A Colonial Pipeline stubline in Murfreesboro, Tennessee was undergoing maintenance on November 5. The pipeline was returned to service, but a valve on that pipeline was accidentally left closed from the maintenance, causing pressure to rupture the pipeline.[500] 1996 On November 21, an explosion occurred in a shoe store and office building in Rio Piedras, Puerto Rico. Thirty-three people were killed, and at least 69 were injured. Crews from the local gas provider, Enron, had not found any gas leaks previously, despite complaints of propane odor in the buildings.[501] 1997 A leak was detected on a 12 inch diameter pipeline near Mount Morris, Illinois on May 9. Between 125,000 and 130,000 US gallons (490,000 L) of gasoline were spilled. A month later, gasoline was till being extracted from the area.[502] 1997 On May 30, Colonial Pipeline spilled approximately 18,900 US gallons (72,000 L) of gasoline, some of which entered an unnamed creek and its adjoining shoreline in the Bear Creek watershed near Athens, Georgia. During the spill, a vapor cloud of gasoline formed, causing several Colonial employees to flee for safety. This spill resulted from a calculation error related to a regular procedure. No one checked the calculations, nor did Colonial have a procedure in place to check such calculations. [493] 1997 A gas pipeline rupture and fire, in Indianapolis, Indiana, on July 21. A 20-inch-diameter steel natural gas transmission pipeline ruptured and released natural gas near an intersection adjoining the Charter Pointe subdivision. The gas ignited and burned, killing one resident and injuring another. About 75 residents required temporary shelter. Six homes were destroyed, and about 65 others sustained damage significant enough to be documented by the local investigation team. A nearby directional drilling operation had hit & weakened the pipeline before the failure.[503] 1997 In August, residents in Vacaville, California noticed petroleum fumes, but a leaking pipeline was not found until September 10. A hairline crack from the pipe's manufacturing was the cause, and 20,000 to 60,000 US gallons (230,000 L) of petroleum products had leaked by the time the source was found.[504] 1997 Over a period of years, more than 420,000 US gallons (1,600,000 L) of gasoline spilled from small leaks in Colonial Pipeline near Darling Creek in St. Helena Parish, Louisiana, before Colonial finally discovered the leak in December 1997. As of September 1999, a plume of gasoline, including leaded gasoline, extended over approximately 14 acres (57,000 m2) on the groundwater surface, more than 60 acres (240,000 m2) of groundwater had been contaminated, and some of the gasoline had entered Darling Creek.[493] 1998 On January 23, at least 800 barrels (130 m3) of light crude oil was spilled into the Gulf of Mexico 50 miles (80 km) south of Galveston, Texas by a leaking pipeline.[505] 1998 A rupture in Colonial Pipeline in a landfill at Sandy Springs, Georgia, discovered on March 30, en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

30/52

9/24/12

List of pipeline accidents in the United States - Wikipedia, the free encyclopedia

resulted in the release of more than 30,000 US gallons (110,000 L) of gasoline. When the pipe was excavated, it was found to be buckled and cracked. The NTSB found that the pipeline ruptured because of settlement of soil and trash underneath the pipeline.[506] 1998 On April 4, a tow of the M/V Anne Holly, comprising 12 loaded and 2 empty barges, which was traveling northbound on the Mississippi River through the St. Louis Harbor, struck the Missouri-side pier of the center span of the Eads Bridge. Three of these barges drifted toward the President Casino on the Admiral, a permanently moored vessel below the bridge on the Missouri side of the river. A natural gas leak resulted when the natural gas supply line to the Admiral was severed in the course of the accident. When the line broke, natural gas began escaping. Although the escaping gas did not ignite, the gas leak had continued for about 3 hours before being stopped.[507] 1998 In South Riding, Virginia, on July 7, a natural gas explosion and fire destroyed a newly constructed residence in the South Riding community in Loudoun County, Virginia. A family consisting of a husband and wife and their two children were spending their first night in their new home at the time of the explosion. As a result of the accident, the wife was killed, the husband was seriously injured, and the two children received minor injuries. Five other homes and two vehicles were damaged.[508] 1998 Lightning struck a Florida Gas Transmission Co. natural gas compressor plant near Perry, Florida on August 13, causing an explosion and massive fire. A second explosion later followed, injuring 5 firefighters & pipeline company employees. 6 nearby homes were also destroyed.[509] 1998 On September 22, a 8,810 barrels (1,401 m3) crude oil spill from a Lakehead (now Enbridge) pipeline near Plummer, Minnesota was caused by an excavator hitting that pipeline. [510] 1998 On December 3, a natural gas liquids pipeline near Moab, Utah failed and ignited near Highway U191, injuring 4 pipeline workers. Asphalt in the road was melted, and traffic was stopped.[511] 1998 A natural gas pipeline rupture and subsequent explosion, in St. Cloud, Minnesota, on December 11. While attempting to install a utility pole support anchor in a city sidewalk in St. Cloud, Minnesota, a communications network installation crew struck and ruptured an underground, 1-inch-diameter, highpressure plastic gas service pipeline, thereby precipitating a natural gas leak. About 39 minutes later, while utility workers and emergency response personnel were taking preliminary precautions and assessing the situation, an explosion occurred. As a result of the explosion, 4 persons were fatally injured; 1 person was seriously injured; and 10 persons, including 2 firefighters and 1 police officer, received minor injuries. Six buildings were destroyed. Damage assessments estimated property losses at $399,000.[512] 1999 Natural Gas Explosion and Fire at a gas pressure station, Wytheville, Virginia, destroying a home and motorcycle store.[513] (January 3, 1999) 1999 In Bridgeport, Alabama, on January 22, while digging a trench behind a building, a backhoe operator damaged a 3/4-inch steel natural gas service line and a 1-inch water service line. This resulted in two leaks in the natural gas service line, which was operated at 35 psig. One leak occurred where the backhoe bucket had contacted and pulled the natural gas service line. The other was a physical separation of the gas service line at an underground joint near the meter, which was close to the building. Gas migrated into a building nearby, where it ignited. An explosion followed, destroying three buildings. Other buildings within a twoblock area of the explosion sustained significant damage. Three fatalities, five serious injuries, and one minor injury resulted from this accident.[514] 1999 On January 23, a construction crew ruptured a 10 inch diameter petroleum products pipe near Germantown, Wisconsin, spilling about 41,000 gallons of gasoline.[515][516] 1999 A pipeline rupture in Knoxville, Tennessee, and released over 53,000 US gallons (200,000 L) of diesel fuel into the Tennessee River on February 9. A brittle-like crack was found on the pipe in an area of en.wikipedia.org/wiki/List_of_pipeline_accidents_in_the_United_States

31/52

Engineering Sciences Thermal Science

The Phoenix Series of Liquefied Natural Gas Pool Fires

Figure 1: LNG cargo carrier entering Boston harbor for offload.

Figure 2: Large scale LNG pool fire experimental site.

Large-scale experiments enable researchers to examine big spills that would occur from LNG tankers For more information: Technical Contacts: Thomas K. Blanchat 505-844-9061 [email protected] Michael M. Hightower 505-844-5499 [email protected] Science Matters Contact: Alan Burns 505-844-9642 [email protected]

Due to the growing demand for natural gas nationwide, the number of liquefied natural gas (LNG) tanker deliveries to U.S. ports (Figure 1) is expected to increase, thus raising concerns about accidental spills or other events. The risks and hazards of a LNG spill will vary depending on its size, the environmental conditions, and the harbor. Risks include not only harm to nearby people, but also significant property damage and economic impact due to long-term interruptions in the LNG supply. Therefore, methods to ensure the security of LNG terminals and shipments in the event of an incident are critical from both public safety and property perspectives. Much progress has been made in LNG threat consequence and vulnerability assessment. A general approach to risk evaluation has been developed, and is used as a basis in site-specific risk assessments. However, there are significant knowledge gaps in the science of very large-scale LNG pool fires. These gaps create serious uncertainties that may either under- or over-estimate latent hazards. Generally, the surface emissive power of a pool fire is a function of pool size and will increase

to reach a maximum value then decrease to reach a limiting value with increasing diameter. For LNG, the limiting power is uncertain. To fill the knowledge gaps and reduce uncertainty, it became necessary to leverage Sandia’s considerable expertise in thermal science to stage very large (> 25-m diameter) LNG pool fires that would surpass by a factor of ten anything that had been attempted previously. To accomplish this task, the Sandia team came up with a simple, low-cost concept of excavating a shallow 120-m diameter pool for the water, and then using the soil to create a deep, insulated 310,000 gallon reservoir to hold the LNG needed for the test (Figure 2). Concrete pipes from the center of the reservoir transported the LNG to the center of the water pool (Figure 3). A simple removable plug allowed gravity to control the flow rate (Figure 4). The considerable safety issues were reservoir integrity, thermal hazards (from cryogenic to extreme heat), asphyxiation, explosion, drowning, and aviation traffic. An advanced threedimensional transport simulation was used to evaluate both the thermal performance of the reservoir and components, the

transport of gaseous boil-off during the cool-down process, and the design of the diffuser in the middle of the pool used to translate the linear momentum of the LNG in the pipes to a radially-spreading pool. Data was captured via cameras (gyroscopically stabilized and suspended from helicopters), spectroscopic diagnostics, and heat flux sensors. Experiments were completed on two LNG spills, with diameters of 25 m and 85 m. Datasets will now allow model development and validation for extrapolation to a scale expected for a spill (300-500 m). The data had unexpected

results in that the fire diameter was smaller than the spreading LNG pool diameter (Figure 5). Previous studies with stagnant pools in pans had resulted in fires the same size as the pool. However, in all such studies, the pans have edges that can result in flame stabilization that would not be available on the open water. The data further showed that, in both very light and significant cross-winds, the flame will stabilize on objects projecting out of the water, suggesting that the ship itself will act as a flame holder.

Figure 3: Schematic of test facility showing LNG reservoir (right), and concrete piping going to the center of the pool.

Figure 5: Image of 85-m pool fire that shows flames not extending across spreading (white area) LNG pool.

Figure 4: Schematic of gravity fed flow and removable plug in center of pool.

Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. SAND2010-5893P 09/2010

+Art

Search

Images

Maps

Play

YouTube

News

Gmail

Drive

Calendar

More

gas tanker explosion

Search

About 1,070,000 results (0.25 seconds)

Art MacKay

0

Share

SafeSearch m oderate

Web Images Maps Videos News More

Any time Past 24 hours Past week Custom range... All results By subject Any size Large Medium Icon Larger than... Exactly... Any color Full color Black and white

Any type Face Photo Clip art Line drawing

Page 2

Standard view Show sizes

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

1/10

Page 3

Page 4

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

2/10

Page 5

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

3/10

Page 6

Page 7

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

4/10

Page 10

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

6/10

Page 11

Page 12

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

7/10

Page 13

Page 14

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

8/10

Page 15

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

9/10

+Art

Search

Images

Maps

Play

YouTube

News

Gmail

Drive

Calendar

More

gas pipeline explosion

Search

About 1,450,000 results (0.17 seconds)

Art MacKay

0

Share

SafeSearch m oderate

Web Images Maps Videos News More

Any time Past 24 hours Past week Custom range... All results By subject Any size Large Medium Icon Larger than... Exactly... Any color Full color Black and white

Any type Face Photo Clip art Line drawing

Page 2

Standard view Show sizes

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

1/9

Page 3

Page 4

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

2/9

Page 5

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

3/9

Page 6

Page 7

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

4/9

Page 8

Page 9

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

5/9

Page 10

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

6/9

Page 11

Page 12

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

7/9

Page 13

Page 14

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

8/9

Show m ore results

Google Images Home

Switch to basic version

Help

Give us feedback

Google Home ​Advertising Programs ​Business Solutions ​Privacy & Terms ​About Google ​

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

9/9

+Art

Search

Images

Maps

Play

YouTube

News

Gmail

Drive

Calendar

More

gas explosion

Search

About 44,300,000 results (0.16 seconds)

Art MacKay

0

Share

SafeSearch m oderate

Web Images Maps Videos News More

Any time Past 24 hours Past week Custom range... All results By subject Any size Large Medium Icon Larger than... Exactly... Any color Full color Black and white

Any type Face Photo Clip art Line drawing

Page 2

Standard view Show sizes

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

1/9

Page 3

Page 4

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

2/9

Page 5

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

3/9

Page 6

Page 7

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

4/9

Page 8

Page 9

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

5/9

Page 10

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

6/9

Page 11

Page 12

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

7/9

Page 13

Page 14

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

8/9

Show m ore results

Google Images Home

Switch to basic version

Help

Give us feedback

Google Home ​Advertising Programs ​Business Solutions ​Privacy & Terms ​About Google ​

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

9/9

Tanker Fires

ASSAM STATE FIRE SERVICE ORGANISATION L.P.G. Tankers' Fire On 1-11-1998, one L.P.G. Bullet Tanker caught fire at Khanapara, Guwahati, where there was a BLEVE (Boiling Liquid Expanding Vapour Explosion) in which 8 persons died and 5 Fire Servie Officers and men were injured seriously. Out of the the seriously injured, the Dy.Director(Tech.) Shri D.K. Chakraborty died at the Hospital and one Sr.Station Officer, Shri Thaneswar Nath, survived miraculously after a long treatment. In this incident, 10 Water Tender Pumps alongwith 40 fire fighter fought the fire for more than 8 hours and saved the densely populated area from further casualities.

Prepare for Worst if LNG Tankers Are Targeted By Craig Hooper For South Boston and Dorchester, liquefied natural gas (LNG) is a permanent neighbor. Every week tankers carrying combustible liquefied natural gas trundle past Castle Island, heading toward the Tractebel/Distrigas operated LNG receiving terminal in Everett. These enormous ships, surrounded by a small security armada, are hard to miss. So too is the colorful Keyspan LNG Gas Tank, a long-standing landmark for travelers on the Southeast Expressway. Most local officials consider Boston's liquefied natural gas infrastructure a potential terror target. Last month, former White House official Richard Clarke inflamed local fears by revealing, in his tell-all book Against All Enemies, that al Qaeda associates used Algerian LNG tankers as a conduit into Boston. Reporters pounced on Clarke's claims and used a lot of ink to detail indignant FBI or White House counterclaims. Unfortunately, the stories offered scant analysis of the key question behind Clarke's concern: Does liquefied natural gas infrastructure endanger Boston? Natural gas infrastructure makes a tempting economic target for adversaries seeking to wage economic mischief. During the Cold War, Western operatives duped Russian agents into purchasing booby-trapped components for use in a natural gas pipeline. In June 1982, the newly constructed pipeline blew up, producing, according to former Air Force Secretary Thomas C. Reed, "the most monumental non-nuclear explosion and fire ever seen from space." The explosion, detailed in a February New York Times

file:///C|/Users/Art/Desktop/LNG/1.%20SORT/WholeBay/Tanker%20Fires.html (1 of 4) [26/09/2012 4:39:05 PM]

Tanker Fires

column by William Safire, was estimated to have yielded an explosion of 3 kilotons or better. Natural gas is not something to be treated lightly. The natural gas industry insists that any threat posed by LNG infrastructure is overestimated and that the risk of an LNG-fueled explosion is limited. But U.S. military leaders think otherwise, and treat LNG facilities with extreme caution. When American military planners discussed retaking Kuwait in 1991, the US Marine Corps insisted a Kuwaiti LNG plant be destroyed before committing troops near the facility. A proposal to preemptively destroy the LNG plant was nixed by General Schwarzkopf, who reportedly claimed, "I do not want to destroy Kuwait in order to save it." In January of this year, an explosion and fire at an Algerian natural gas facility killed more than 20 people, offering further evidence that liquefied natural gas can turn lethal. Keyspan's Gas Tank is the only obvious piece of permanent liquefied natural gas infrastructure in South Boston and Dorchester. The thirty-year old facility, sitting quietly by I-93 and tucked under the approaches to Logan Airport, has a relatively spotless operational history. To Keyspan's credit, the LNG tank has suffered only a handful of leaks due to minor mechanical and human errors. Keyspan contends that the tank poses little danger to Dorchester. In the days after the attacks of September 11, 2001, the Boston Globe reported that Keyspan representative Mike Connors insisted the tank could not explode and that heat from a fire at the tank would not harm neighbors. Given that both the Boston Globe and Boston Herald have since reported on studies that say that an LNG fire can release enough heat in thirty seconds to scorch unprotected skin a half-mile away, Keyspan may well be relying upon outdated safety assumptions. Is the double-walled Keyspan gas tank tough enough to take on present-day terrorists? One popular terrorist weapon, the modern, hand-held anti-tank grenade launcher, is becoming increasingly powerful. The 2004-2005 issue of Jane's Infantry Weapons, a compendium of modern weaponry, suggests that a light-weight, easily hidden rocket-propelled grenade called an RPG-7, can, when firing a modern warhead, penetrate over three feet of military-quality armor and break through more than a meter

file:///C|/Users/Art/Desktop/LNG/1.%20SORT/WholeBay/Tanker%20Fires.html (2 of 4) [26/09/2012 4:39:05 PM]

Tanker Fires

of reinforced concrete. On land, the rocket threat can be beaten. At the Keyspan LNG tank, a variety of tasteful and relatively inexpensive steps can quickly make Dorchester's liquefied natural gas tank an unappealing target for rocket-wielding terrorists. Simple barriers significantly reduce the chance that hand-held missiles might puncture the tank and ignite the contents. Insuring LNG tanker security is another matter. The terror threat to these high-profile targets is real. According to a recent report published in a Lebanon-based newspaper, the Daily Star, an American counterterrorist official was quoted as saying that al Qaeda has developed a manual to teach how rocket-propelled grenades can "turn liquefied natural gas (LNG) tankers into floating bombs." This threat isn't the product of idle terrorist brainstorming. Terrorists have already successfully targeted tankers. During the Iran/Iraq war, Iranian radicals, sailing small craft, used anti-tank missiles to damage merchant shipping in the Persian Gulf. While no LNG tankers were hit, tankers of similar design were assaulted. The Lloyd's Register/Fairplay database on shipping accidents records that, on July 3, 1988, Iranian terrorists attacked a liquid petroleum gas (LPG) tanker, the Berge Strand. Anti-tank grenades punctured the dual-hulled ship and compromised several of the tanker's LPG storage tanks. While the Berge Strand wasn't carrying fuel at the time, successful penetration of the fuel storage tanks may validate the larger terror threat. The LNG industry is quick to cite a different Persian Gulf attack as evidence that LNG carriers are able to withstand attack. The fully loaded LPG tanker Gaz Fountain survived after being set ablaze by aircraft-launched rockets. What goes unreported by the LNG industry is that the attack occurred in the open sea, allowing firefighters to operate unimpeded and without regard for secondary blazes ignited by LNG's concentrated combustive power. Even though every gallon of LPG holds several hundred times less potential energy than a gallon of LNG, the Gaz Fountain was kept miles out of port, and experts were flown in to prevent an explosion. An attack in a crowded, closed harbor offers little in the way of conveniences for firefighters and may make controlling an LNG tanker fire much more difficult.

file:///C|/Users/Art/Desktop/LNG/1.%20SORT/WholeBay/Tanker%20Fires.html (3 of 4) [26/09/2012 4:39:05 PM]

Tanker Fires

Any attacked or damaged tanker must leave Boston the way it entered - passing by South Boston and Dorchester. In the event of an LNG emergency at the Keyspan LNG tank or in the harbor, many local neighborhoods may require rapid evacuation. Are the neighborhoods prepared? No. Are the companies that operate Boston's LNG infrastructure encouraging greater neighborhood readiness? No. This is exactly the sort of complacency that led America to a disaster on September 11, 2001. Help isn't on the way. Governor Mitt Romney, the LNG industry and the federal government are working hard to stymie forthright debate over the risks and benefits of Boston's LNG infrastructure. Rather than spend a bit more to better secure LNG, Boston's LNG-associated companies seem brazenly confident that America's growing demand for LNG will outweigh Boston's overwhelming desire for a prudent and secure LNG infrastructure. Boston may need this clean-burning, environmentally friendly fuel, but, given the gathering terror threat, Boston's present safety arrangements are unacceptable. Boston can live with LNG only if our elected officials, insurance companies and entire neighborhoods demand that Boston prepare for the worst. Boston holds a strong position; if Boston's LNG operators come to the table and go the extra mile to be a good neighbor here, the entire LNG sector will benefit. If not, then the LNG industry risks losing Boston and will have a tough time finding a foothold anywhere else in the Northeast. Craig Hooper is a graduate student at the Harvard Graduate School of Public Health. He lives on Sagamore Street in Dorchester.

file:///C|/Users/Art/Desktop/LNG/1.%20SORT/WholeBay/Tanker%20Fires.html (4 of 4) [26/09/2012 4:39:05 PM]

+You

Search

Images

Maps

Play

YouTube

News

Gmail

Documents

Calendar

More

earthquake fire japan

Search

Sign in

SafeSearch m oderate

About 9,490,000 results (0.16 seconds)

Web Images Maps Videos News More

Japan‑Earthquake‑Fire‑Fire.jpg 24sevenpost.com 567 × 369 - Massive quakes in Japan 8.9 and many many more Near Tokyo caught fire and at ... Similar More sizes

Any time Past 24 hours Past week Custom range... All results By subject Any size Large Medium Icon Larger than... Exactly... Any color Full color Black and white Page 2

Any type Face Photo Clip art Line drawing Standard view Show sizes

www.google.ca/search?q=LNG+earthquake+fire+japan&hl=en&tbm=isch&prmd=imvns&source=lnms&…

1/9

Page 3

Page 4

www.google.ca/search?q=LNG+earthquake+fire+japan&hl=en&tbm=isch&prmd=imvns&source=lnms&…

2/9

+Art

Search

Images

Maps

Play

YouTube

News

Gmail

Drive

Calendar

More

Cleveland gas explosion

Search

Art MacKay

0

Share

SafeSearch m oderate

About 2,620,000 results (0.29 seconds)

Web Images Maps Videos News More

10155380‑large.jpg cleveland.com 380 × 325 - East Ohio Gas explosion gallery (5 photos). It happened on a Friday inMore 1944,sizes ... Similar

Any time Past 24 hours Past week Custom range... All results By subject Any size Large Medium Icon Larger than... Exactly... Any color Full color Black and white

Any type Face Photo Clip art Line drawing

Page 2

Standard view Show sizes

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

1/8

Page 3

Page 4

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

2/8

Page 5

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

3/8

Page 6

Page 7

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

4/8

Page 8

Page 9

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

5/8

Page 10

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

6/8

Page 11

Page 12

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

7/8

Page 13

Show m ore results

Google Images Home

Switch to basic version

Help

Give us feedback

Google Home ​Advertising Programs ​Business Solutions ​Privacy & Terms ​About Google ​

https://www.google.ca/search?q=cleveland+gas+explosion&hl=en&prmd=imvns&source=lnms&tbm=is…

8/8

Appendix C3 C3-1 Chronological List of LNG Accidents C3-2 Marine Safety and Security Requirements C3-3 Design and Safety Standards Applicable to Natural Gas Projects

C3-1 CHRONOLOGICAL LIST OF LNG ACCIDENTS

Cabrillo Port LNG Deepwater Port

March 2006

CHRONOLOGICAL LIST OF LNG ACCIDENTS Major LNG Incidents Incident Date

Ship/Facility Name

Location

1944

East Ohio Gas LNG Tank

1965 1965

Jules Vernet

1965

Methane Princess

1971

LNG ship Esso Brega, La Spezia LNG Import Terminal

1973

Texas Eastern Transmission, LNG Tank

Injuries/ Fatalities

Cleveland, Ohio, US

NA

128 deaths

Canvey Island, UK

A transfer operation

Italy

Staten Island, NY, US

Canvey Island, UK

1973

1974

Ship Status

Massachusetts

March 2006

Loading Disconnec ting after discharge Unloading LNG into the storage tank

NA

Ship/ Property Damage NA

LNG Spill/ Release

Comment

NA

LNG peakshaving facility. Tank failure and no earthen berm. Vapor cloud formed and filled the surrounding streets and storm sewer system. Natural gas in the vaporizing LNG pool ignited.

1 seriously burned No

Yes

Yes

Overfilling. Tank covered and deck fractures.

No

Yes

Yes

Valve leakage. Deck fractures.

Yes

First documented LNG rollover incident. Tank developed a sudden increase in pressure. LNG vapor discharged from the tank safety valves and vents. Tank roof slightly damaged. No ignition.

No

Industrial incident unrelated to the presence of LNG (construction incident). During the repairs, vapors associated with the cleaning process apparently ignited the mylar liner. Fire caused temperature in the tank to rise, generating enough pressure to dislodge a 6-inch thick concrete roof, which then fell on the workers in the tank.

NA

40 killed

Yes

NA

No

NA

No

Yes

Yes

Glass breakage. Small amount of LNG spilled upon a puddle of rainwater, and the resulting flameless vapor explosion, called a rapid phase transition (RPT), caused the loud "booms". No injuries resulted.

Loading

No

Yes

Yes

Valve leakage. Deck fractures.

C3.1-1

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Major LNG Incidents Ship Status

Injuries/ Fatalities

Ship/ Property Damage

LNG Spill/ Release

Comment

Methane Princess

In port

No

Yes

No

Touched bottom at Arzew.

Philadelphia Gas Works

NA

No

Yes

NA

Not caused by LNG. An iso-pentane intermediate heat transfer fluid leak caught fire and burned the entire vaporizer area.

NA

1 worker frozen to death

NA

Yes

Aluminum valve failure on contact with cryogenic temperatures. Wrong aluminum alloy on replacement valve. LNG released, but no vapor ignition (LNG liquefaction facility).

Loading

No

No

Yes

Tank overfilled.

Incident Date

Ship/Facility Name

1974 1975

1977

Arzew

1977

LNG Aquarius

1979

1979 1979

Columbia Gas LNG Terminal

Mostefa BenBoulaid Ship Pollenger Ship

Location

Algeria

NA

1 killed, 1 seriously injured

Yes

Yes

An explosion occurred within an electrical substation. LNG leaked through LNG pump electrical penetration seal, vaporized, passed through 200 feet of underground electrical conduit, and entered the substation. Since natural gas was never expected in this building, there were no gas detectors installed in the building. The normal arcing contacts of a circuit breaker ignited the natural gas-air mixture, resulting in an explosion. (LNG regasification terminal)

?

Unloading

No

Yes

Yes

Valve leakage. Deck fractures.

?

Unloading

No

Yes

Yes

Valve leakage. Tank cover plate fractures. Stranded. Severe damage to bottom, ballast tanks, motors water damaged, bottom of containment system set up. Shaft moved against rudder. Tail shaft fractured. Stranded. Ballast tanks all flooded and listing. Extensive bottom damage.

Cove Point, Maryland, US

1979

El Paso Paul Kayser Ship

At sea

No

Yes

No

1980

LNG Libra

At sea

No

Yes

No

1980

LNG Taurus

In port

No

Yes

No

March 2006

C3.1-2

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Major LNG Incidents Ship Status

Injuries/ Fatalities

Ship/ Property Damage

LNG Spill/ Release

Melrose

At sea

No

Yes

No

1985

Gradinia

In port

No

Not reported

No

1985

Isabella

Unloading

No

Yes

Yes

1989

Tellier

Loading

No

Yes

Yes

1990

Bachir Chihani

At sea

No

Yes

No

Incident Date

Ship/Facility Name

1984

Location

1993

Indonesian liquefaction facility

Indonesia

2002

LNG ship Norman Lady

East of the Strait of Gibraltar

2004

Skikda I

Algeria

NA

No

At sea

No

NA

27 killed 56 injured (The casualties are mainly due to the blast, few casualties due to fire)

NA

Yes

NA

Comment Fire in engine room. No structural damage sustained - limited to engine room. Steering gear failure. No details of damage reported. Cargo valve failure. Cargo overflow. Deck fractures. Broke moorings. Hull and deck failures. Sustained structural cracks allegedly caused by stressing and fatigue in inner hull.

NA

LNG leak from open run-down line during a pipe modification project. LNG entered an underground concrete storm sewer system and underwent a rapid vapor expansion that overpressured and ruptured the sewer pipes. Storm sewer system substantially damaged.

No

Collision with a U.S. Navy nuclear-powered attack submarine, the U.S.S Oklahoma City. In ballast condition. Ship suffered a leakage of seawater into the double bottom dry tank area.

NA

On January 2004: No wind, semi-confined area (cold boxes, boiler, control room on 3 sides). The fire completely destroyed the train 40, 30, and 20, although it did not damage the loading facilities or three large LNG storage tanks also located at the terminal. Complete details are pending until completion of ongoing accident investigation.

Sources: University of Houston, "LNG Safety and Security," October 2003. http://www.beg.utexas.edu/energyecon/lng/. Cited with permission; Sonatrach, "The Incident at the Skikda Plant: Description and Preliminary Conclusions", March 2004.

March 2006

C3.1-3

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

C3-2 MARINE SAFETY AND SECURITY REQUIREMENTS

Cabrillo Port LNG Deepwater Port

March 2006

MARINE SAFETY AND SECURITY REQUIREMENTS 1 Structural Safety Features of the FSRU Some of the major safety features of the FSRU required that would reduce the likelihood of an accidental cargo release and would mitigate any release, regardless of cause are listed in the following table. Safety Feature

Description

Double Hull Construction

The FSRU and LNG carriers would be constructed with an outer and inner hull to provide protection against collisions and resultant cargo loss. These hulls are separated from each other by structural members and separated from the Moss spherical tanks by the tank mounts. Thus a collision would need to penetrate three layers to result in cargo spillage.

Separation of cargo holds and piping systems

IGC code requires the structural separation of cargo holds from other spaces, as well as separation of cargo piping from other piping systems. Amongst other things, this helps keep cargo leaks away from potential ignition sources and keeps cargo from inadvertently being pumped through the wrong pipes.

Accessibility for Inspection Access

IGC code requires that a tank be constructed so that at least one side is visible and accessible to inspectors. This allows proper periodic inspection of the tank for integrity and signs of corrosion or stress.

Leak Detectors in Hold spaces

IGC code requires that gas detectors and low temperature sensors be placed in a cargo hold in order to cargo leakage. An alarm sounds if either is detected and appropriate repairs and precautions can be undertaken.

Tank Requirements for Cargo Containment

ICG code requires that a tank be constructed with materials that can withstand the temperatures involved so as to properly contain the cargo, and have adequate relief valve systems to avoid over pressurization.

Structural Analysis

IGC code requires structural analysis of the cargo containment system and specifies individual tank stress limitations.

Secondary containment and thermal management

IGC code requires partial secondary containment to contain leaks and prevent contact of cryogenic liquid with the inner hull. This prevents thermal stress. In addition, insulation in conjunction with a primary and backup heating system must be installed that would keep the cargo from exceeding the thermal limitations of the material selected for the inner hull should the leak prevention system fail.

Tank Construction and Testing Requirements

IGC codes address standards for workmanship, quality, and testing of tanks under construction. Each tank on the FSRU will have had its welds non-destructively tested, and have had a pressure test to insure integrity before cargo is pumped aboard.

Isolation, Construction and Testing Requirements for Piping and Pressure Vessels

IGC code specifies piping thickness, leak testing, pressure testing, isolation requirements, welding requirements and many other aspects of pressure vessel and piping design and construction. This insures the integrity of these systems before any cargo is brought aboard.

March 2006

C3.2-1

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Safety Feature

Description

Emergency Shutdown Valves and Shutdown Systems

IGC code requires remote control shutdown systems for ceasing of cargo and vapor transfer in an emergency. This system must have the ability to be activated from at least two locations on board the FSRU and will also be automatically activated in the event of a cargo fire.

Pressure Venting Systems

IGC code specifies that appropriate venting of the cargo be installed to keep the cargo under the design pressure of the tank and keep relief valves from needing to operate. The FSRU will use some of this gas for fueling the Submerged Combustion Vaporizers, and will add the rest to the gasified product being pumped to shore.

Vacuum Protection Systems

IGC code requires the installation of relief valves that would prevent under pressurization of cargo tanks in the event that cargo was pumped out without adequately providing for vapor return. The FSRU will have sufficient vapor return capacity to keep the pressures at appropriate levels, however this system will prevent under pressurization should this system fail to be actuated or fail to work properly.

Fire Protection Systems

IGC code requires that LNG carriers have a saltwater fire main system for fighting fires throughout the ship, and fixed dry chemical and CO2 systems for cargo areas and compressor rooms, respectively.

Cargo Tank Instrumentation

IGC code requires that each cargo tank be outfitted with an integrated instrumentation/alarm system that notifies the crew of possible leaks via gas detection and temperature sensors; and tank liquid levels, temperatures and pressures. These systems, as well as the pressure relief systems mentioned above, provide many-layered protection against cargo release either through equipment malfunction or human error.

Additional Gas Detection Systems

IGC code also requires gas detection systems and alarms in spaces where cargo is located, including compressor spaces, spaces where fuel gas is located, and other spaces likely to contain gasified cargo. Venting systems for certain spaces and portable gas detectors are also required.

Automatic Safety Shutdown Systems

IGC code requires that cargo loading areas and the docks be equipped with LNG vapor and fire detection systems that automatically shut down the transfer systems in the event of a leak or fire. These shutdowns can also be manually operated by personnel on the dock (in this case, the FSRU) or LNG carrier.

Loading Arm Emergency Release Couplings

The FSRU loading arms are designed to isolate the flow of cargo and break away from their connection to the carrier if relative motion exceeds safety parameters. This prevents damage to the arms, and averts the spill of cargo which would result from a broken arm. Quantities spilled during this process would be only a few gallons, most of which would be caught in drip trays to prevent deck thermal damage.

2 Operational Measures for Accidental Release Prevention In addition to the design regulations described above, the international and national entities with authority to impose such regulations have also provided operational guidelines to reduce the likelihood and impact of an LNG release aboard carriers. The FSRU, as a Deepwater Port of the United States, is primarily guided by the Deepwater Port Act as modified in 33 CFR 148 150 by the Maritime Safety and Security Act and other legislation and agency determinations.

March 2006

C3.2-2

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

These measures include: •

Training,

•

Formal Operational Procedures, and

•

Inspections.

2.1 Training Training requirements for crews of LNG carriers are specified in the IMO STCW Convention and those for the FSRU are detailed in 33 CFR 150. A wide variety of training is included for both, including marine firefighting, water survival, spill response and clean-up, emergency medical procedures, hazardous materials procedures, confined space entry, and training on operational procedures. Specifics are also included in the below summary of the Deepwater Port Operations Manual requirements. 2.2 Formal Operational Procedures Both the FSRU and the visiting LNG carriers would be required to have formal operating plans that cover an extensive array of operational practices and emergency procedures. LNG carriers are required by the IMO to meet the ISM Code, which addresses preparing for responding to emergency situations like fire and LNG releases. The LNG carrier’s navigational, pollution response, and some emergency procedures would also be covered in the Deepwater Port Operations manual, which addresses every aspect of the FSRU operations. The minimum contents of this manual are detailed in 33 CFR 150. This manual provides detailed requirements that cover contingencies and normal operations. The operations manual must meet all requirements set forth by the US Coast Guard, and be approved by that organization before operations begin. The operations manual is required by 33 CFR 150 to address the following areas: •

The DWP facilities must be clearly described physically and geographically, applicable codes for design and construction must be detailed, schematics of all systems must be included which show the positions of all operations and safety equipment. The communications system must be described and communications procedures laid out.

•

Procedures for the visiting LNG carriers are also required to be included. Operating hours must be set and sizes and types of tankers that may be received must be described. Navigation standards for the LNG carriers must be set forth, including operating limits for each type of carrier. Speed limits for the safety zone must be specified, as well as the means of tracking, communicating and giving routing instructions to the carriers. Required notices that carriers must give prior to arrival must be detailed. Rules for navigating in the safety zone and for mooring/unmooring must be detailed. Special equipment needed for mooring or navigating must be described. Procedures for clearing all carriers and support vessels away from the FSRU in the event of an emergency or for normal operations must be specified.

•

Weather forecasting and information dissemination procedures must be set forth. Specific weather limitations must be defined for carrier arrival, cessation of cargo transfer operations and departure of carriers from moorings in the event of adverse weather being forecasted or as it occurs unexpectedly. This includes defining conditions in which the FSRU would be secured and evacuated.

March 2006

C3.2-3

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

•

The manning requirements for all operational and emergency situations must be specifically described, with personnel in charge of major evolutions designated by name, in writing. The supervisors would be reviewed by the US Coast Guard to ensure they have the proper qualifications and training to perform their duties.

•

Procedures for major evolutions, such as cargo transfers, must be set forth in detail. Manning and training requirements, specific duties for watchstanders and supervisors and emergency shutdown system settings must be detailed. Special precautions and handling procedures for LNG must be included.

•

Maintenance program requirements and specific procedures are required to document the service and repair of cargo equipment, fire fighting systems, safety equipment and cranes.

•

Occupational Health and Safety training procedures and requirements must be detailed, including: housekeeping, illumination requirements, fall arrest equipment, personnel transfer systems, hazard communication, permissible exposure limits for hazardous substances, protective guards around machinery, electrical safety, lockout/tagout procedures, crane safety, sling usage, hearing conservation, hot work, warning sirens, and confined space entry.

The security plan is part of the operations manual and is covered in detail in the below security section. An environmental monitoring program also must be included, which describes procedures for monitoring the effects of the port on its surroundings. This must include periodic re-examination of the physical, chemical and biological factors examined in the Environmental Impact Statement, as well as air and water monitoring proscribed by other statutes and state law. Detailed studies are required in the event of a spill. 2.3 Inspections The US Coast Guard has the authority and jurisdiction to perform inspections of Project vessels in U.S. waters, or on the high seas after a vessel states intent to moor at the DWP. Additional inspections may be carried out on LNG carriers by their flag states, by classification societies, and by the owners. Per 33 CFR 150, the US Coast Guard also may inspect the FSRU at any time, with or without notice, for safety, security, and compliance with applicable U.S. laws and regulations. 33 CFR 150 mandates that the FSRU be self inspected every 12 months by the owner or operator to ensure compliance with applicable safety and security laws and regulations. The results must be reported to the US Coast Guard COTP within 30 days of completion, and may be checked for accuracy by a Coast Guard inspection at any time. This report must include descriptions of any failure, and the scope of repairs subsequently made. Any classification society certification or interim class certificate must be reported to the COTP as well. The US Coast Guard has marine inspection programs for ships, Outer Continental Shelf structures, DWP Facilities and waterfront facilities. US Coast Guard Officers and Petty Officers receive very detailed training on applicable regulations and inspection techniques. For this project, the most applicable Safety programs include the Port State Control program and 33

March 2006

C3.2-4

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

CFR 160 for the inspection and routing of visiting ships, and the DWP inspection program specified by 33 CFR 150. Ports State Control of visiting vessels occurs by means of a US Coast Guard Boarding, targeted at determining the vessels compliance to international IMO standards for safety, pollution control, loading, and watch stander qualification, training and procedures. Vessel safety, sanitation and cargo handling equipment is inspected, emergency drills and procedures may be ran in order to determine crew proficiency, navigation practices are examined, and all pertinent plans, safety management systems and other required documents are examined. The required 96-hour Notice of Arrival for these vessels allows the Coast Guard ample time to determine which vessels to board, whether to conduct the boarding in port or at sea, or even if entry is denied pending an inspection. The COTP decides which vessels are at highest risk for non-compliance with IMO conventions through a process by which the following factors are considered: The owner, Flag State and classification society of the vessel - some owners, flag states and classification societies have a history of poor inspection and regulation of their vessels; how many times and how recently a vessel has been boarded or detained for violations previously; and the type of cargo the vessel is carrying. The vessels having the most factors of concern are boarded immediately, while others may be boarded on subsequent entries into the U.S. Vessels found to be in non-compliance with IMO standards may be recommended for further flag state or classification society audit, detained in port until their discrepancies are fixed, ordered to anchorage for the same purpose, or forbidden to enter U.S. waters. 33 CFR 160 gives authority to each US Coast Guard District Commander or Captain of the Port to order a vessel to operate or anchor in the manner directed when there is a suspected violation of law or treaty, there is a failure to satisfy the cargo transfer provisions of 33 CFR160.113, or if justified by weather, visibility, port congestion or condition of the vessel. 33 CFR 160.113 Gives COTP the authority to prohibit a vessel from transferring cargo or operating on the navigable waters of the US if the vessel’s history of accidents, pollution incidents, or serious repair problems creates reason to believe that the vessel may be unsafe or pose a threat to the marine environment. It also allows these restrictions for other reasons: The vessel is in violation of a law or regulation, has discharged oil or other hazardous substance in violation of US law or treaty, fails to comply with Vessel Traffic Service requirements, or does not have at least one licensed deck officer on the navigation bridge that speaks English. One of the relevant results of this inspection regimen is that every Project vessel and the FSRU would be inspected at least yearly for compliance to all applicable IMO standards and U.S. laws. Equipment, training, qualifications, operating and emergency procedures, administrative controls, and most every other aspect leading to safe operation of the FSRU and project vessels would be checked by the owners, the flag states (for vessels) and the United States for compliance. 3 Security Measures that Help Prevent Release Incidents Due to Deliberate Attacks Regulation and operational procedures play a vital role in the prevention of terrorist acts. In fact, much of what prevents or mitigates an accident will do the same for a terrorist act (double hulls, fire suppression systems, etc). However, potential deliberate acts of terrorism expose the Project to new threats, many of which cannot easily be prevented, though mitigative actions may be nearly the same after the incident occurs. March 2006

C3.2-5

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

The foundation for the FSRU and visiting LNG carriers' security would be the requirements for a security plan outlined in 33 CFR 150. This plan would address security issues including, but not limited to: •

Access control for people, goods and material;

•

Monitoring and alerting vessels that approach or enter the ports security zone;

•

Identifying risks and measures to deter terrorist activity;

•

Internal and external notification requirements and responses in the event of a perceived threat or attack on the port;

•

Designating a Port Security Officer; providing identification means for port personnel; security training requirements;

•

Actions and procedures that are scalable to the threat; emergency procedures such as evacuation; special operations procedures (re-manning, refueling, diving, support vessel operations and logistical concerns);

•

Recordkeeping for maintenance; and

•

Tests and operations outlined in the operations manual.

In addition, radar monitoring of the security zone is a required when any vessel approaches or enters the zone. Such vessels must be identified and warned off via radio. 3.1 Requirements to meet IMO’s International Ship and Port Facilities Security Code (ISPS) Code IMO’s ISPS code has the following additional requirements: • Security levels; •

Ship security plans;

•

Ship security alarm systems;

•

Automatic identification systems;

•

Port security plans;

•

Declarations of security; and

•

Facility security plans.

For the U.S., these IMO requirements are addressed in 33 CFR Subchapter H—Maritime Security. 3.1.1 Security levels For the U.S., security levels are covered in 33 CFR 101, which ties the three tiered Maritime Security (MARSEC) level to the five level Department of Homeland Security's Homeland Security Advisory System as the below table depicts.

March 2006

C3.2-6

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Homeland security advisory system (HSAS) threat condition

Equivalent maritime security (MARSEC) level

Low: Elevated: Guarded:

Green Blue Yellow

MARSEC Level 1.

High:

Orange

MARSEC Level 2.

Severe:

Red

MARSEC Level 3.

Specific actions would be required of Project personnel at each level, and would be detailed in the security plan for the FSRU and the Ship Security plans. Changes in MARSEC level is communicated by the COTP via Broadcast NTM, and all who are required to have a security plan (facilities, vessels must report attainment of measures in their plan that correspond to the new MARSEC level to the appropriate Coast Guard District Commander. When the USCG determines it is necessary to enact additional measures to counter a maritime threat, the USCG Commandant (or delegate) may issue a directive to those required to have a security plan (or portions of, as needed) to take additional security measures to counter the threat. Reporting of attainment of the measure or its approved equivalent is carried out in the same way as a change in MARSEC, but within a time period specified by the directive. 3.1.2 Vessel security plans 33 CFR 104 requires every vessel owner or operator who operates in U.S. waters to develop and submit to USCG a vessel security plan. The regulations provide the format and requirements for the plan. Vessel security plan implementation must be evaluated by an onboard verification by the flag state or a security organization recognized by the flag state before an International Ship Security Certificate (ISSC) can be issued for that vessel. These plans must include provisions for access to the ship by ship personnel, passengers, visitors, etc; restricted areas on the ship; handling of cargo; delivery of ship’s stores; handling unaccompanied baggage; and monitoring the security of the ship. These measures are intended to prevent deliberate destructive act on board a vessel and the possible hijacking of the vessel for use as a weapon (ramming other vessels, bridges, blocking channels, releasing cargo near shore, etc). Control and compliance measures for those vessels in violation of this requirement include the vessel’s inspection, delay or detention. Vessel operations may be restricted, port entry into the U.S. denied, or the vessel may be expulsed from a U.S. port. Lesser administrative or corrective actions may be taken. The vessel’s security plan is subject to USCG approval, which may be withdrawn, which would make it illegal for the vessel to operate in, on, under or adjacent to U.S. waters.

March 2006

C3.2-7

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

3.1.3 Ship security alarm systems Ship security alarm systems are required by the ISPS code for Project LNG carriers. These systems are manually operated by the crew in the event of a terrorist destructive act or attempted takeover. An alarm does not sound on the vessel, but does automatically send a signal to appropriate authorities, such as the Coast Guard. 3.1.4 Automatic identification systems (AISs) As described in the above vessel collision avoidance section, an AIS provides augmented data to radar users, which aid in the identification of vessels. The traffic controllers onboard the FSRU, the VTS and USCG responders would be able to locate and identify vessels outfitted with AIS more quickly and accurately, thus decreasing confusion and response time to an emergency, including security alarm activations. 3.1.5 Port security plans The ISPS Code requires ports to have a port facility security officer and to develop a port facility security plan which must interface with the individual vessel security plans. In the United States, 33 CFR 103 mandates an Area Maritime Security plan which applies to all vessels and facilities located in, on, under, or adjacent to waters subject to U.S. jurisdiction. This regulation empowers the COTP to set up counsels to advise on port security, write and exercise the area security plan and defines required elements of the plan. (ex. Plan must address actions to be taken for a change of MARSEC, what to do if a vessel security alert system is activated, estimated response and timeframe for a Transportation Security incident, etc) 3.1.6 Declarations of security Declarations of security are required by 33 CFR 101 for ports across the US, and are intended to serve as the formal means by which the security actions of the vessel and port are agreed upon during mooring and cargo transfer operations. This declaration must be signed by the vessel and facility security officer prior to commencement of offloading. 3.1.7 Facility security plans Under the USCG maritime security regulations (33 CFR 105 Subpart D), LNG facilities that receive LNG carriers will have to develop a facility security plan. Like the vessel security plans that have to meet the ISPS Code, the USCG regulations define areas the facility security plans have to address, including: •

Security administration and organization of the facility;

•

Personnel training;

•

Drills and exercises;

•

Records and documentation;

•

Response to change in security level;

•

Procedures for interfacing with vessels;

•

Declaration of Security;

•

Communications;

•

Security systems and equipment maintenance;

•

Security measures for access control, restricted areas, handling cargo, delivery of vessel stores and bunkers, and monitoring;

March 2006

C3.2-8

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

•

Security incident procedures; and

•

Audits and security plan amendments.

Like ship security plans, USCG must approve facility security plans. If the COTP deems a waterfront facility unsafe or insecure in any way by, vessels may be prevented from docking there, or be moved if already docked. Other, control and compliance measures for facilities for violations of these requirements include restriction on facility access, conditions being put on facility operations, suspension of operations, or revocation of approval for the facility’s security plan which makes it illegal for the facility to operate. 3.1.8 Coast Guard operational measures applicable to security of the Project The USCG, in addition to its inspection duties, is also an active enforcer of all applicable national and international law on the high seas and within the waters of the United States. The USCG's enforcement of these laws will significantly add to the security of any nearby facility. These actions may include: •

Enforcement of 96-hour Notice of Arrival (NOA) requirements, including vetting crew and passenger lists against terrorist and criminal databases.

•

Conducting regular patrols with aircraft and armed surface vessels to support Maritime Domain Awareness (knowing what vessels are within or near U.S. waters).

•

Conducting Right of Approach questioning of any vessel to determine county of registry, last port of call, crew nationality and other useful data.

•

Conducting background intelligence checks on sighted vessels and like checks on the crews of boarded vessels.

•

Monitoring all vessel traffic over 300 GWT with 25 NM of Pt. Fermin Light as part of VTS LA/LB (Note: this area is approx 5nm from the FSRU and covers approaches from the West).

•

Conducting armed escorts of vessels deemed to be High Risk.

•

Placing Armed Sea Marshals on board High Risk vessels (Note: the determination to provide escort or Sea Marshals for any Project vessel is at the discretion of COTP).

•

Conducting searches of vessels suspected of violating immigration, customs and narcotics laws.

•

Inspecting the safety gear of all U.S. flagged and state registered pleasure craft and commercial vessels.

•

Conducting searches of foreign vessels with flag state or Master's consent for evidence of violation of applicable laws.

•

Acting in accordance with the U.S. Military Standing Rules of Engagement to protect U.S. citizens and property.

•

Patrolling, warning and boarding vessels to enforce security zones.

March 2006

C3.2-9

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Should the threat level or other circumstance dictate, the USCG and other military branches would take measures to provide for the security of the Project. The nearby presence of military vessels and aircraft conducting operations and surveillance of the Point Mugu Sea Range would also augment Maritime Domain Awareness, and would periodically result in the presence of armed warships within relatively close proximity to the FSRU. All of these vessels could be hailed on frequencies available in the FSRU communications centers, and all are allowed by the rules of engagement to protect themselves, other U.S. military units, U.S. Citizens and property if being attacked. The COTP may restrict anyone, or anything from entering a waterfront facility subject to U.S. jurisdiction or boarding a vessel subject to U.S. jurisdiction deemed necessary for safety or security. Further, to prevent damage or injury to vessels or facilities or safeguard ports, territory, or waters of the U.S., COTP may establish a security zone, consisting of whatever sections of water and land deemed necessary. No person or vessel may enter this zone or leave any article on a vessel or facility in this zone without COTP (or designee) approval. Any vessel, facility or person in this zone may be inspected or searched, and items or persons may be removed from the zone as deemed necessary. Guards may be posted on any vessel or anywhere in a security zone deemed necessary. Movements of vessels may be controlled as necessary, and within the territorial seas of the U.S., the COTP may enlist the aid and cooperation of Federal, State, county, municipal, and private agencies to assist. Licenses and required documentation may be required by the COTP for personnel entering a waterfront facility, who may revoke/not approve such based on deciding that the person is a security risk. An appeals process is set up, as is a board to hear such consisting of a Coast Guard Officer and members from company management and a labor representative.

March 2006

C3.2-10 Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

C3-3 DESIGN AND SAFETY STANDARDS APPLICABLE TO NATURAL GAS PROJECTS

Cabrillo Port LNG Deepwater Port

March 2006

DESIGN AND SAFETY STANDARDS APPLICABLE TO NATURAL GAS PROJECTS Documents Incorporated by Reference into Title 49 CFR Part 192, Appendix A, as amended through June 14, 2004

Title (applicable edition)

A.

American Gas Association (AGA)

(1)

AGA Pipeline Research Committee, Project PR-3-805

B.

American Petroleum Institute (API)

(1)

API Specification 5L

Specification for Line Pipe (42nd edition, 2000).

(2)

API Recommended Practice 5L1

Recommended Practice for Railroad Transportation of Line Pipe (4th edition, 1990).

(3)

API Recommended Practice 5LW

Transportation of Line Pipe on Barges and Marine Vessels (2nd edition, 1996)

(4)

API Specification 6D

Specification for Pipeline Valves (Gate, Plug, Ball, and Check Valves) (21st edition, 1994).

(5)

API Standard 1104

Welding of Pipelines and Related Facilities (19th edition, 1999, including its October 31, 2001 errata).

C.

American Society for Testing and Materials (ASTM)

(1)

ASTM Designation A 53/A53M-99b

Standard Specification for Pipe, Steel, Black and Hot-Dipped, Zinc-Coated, Welded and Seamless (ASTM A53/A53M-99b).

(2)

ASTM Designation A 106

Standard Specification for Seamless Carbon Steel Pipe for High-Temperature Service (ASTM A10699).

(3)

ASTM Designation A 333/A 333M

Standard Specification for Seamless and Welded Steel Pipe for Low-Temperature Service (ASTM A333/A333M-99).

(4)

ASTM Designation A 372/A 372M

Standard Specification for Carbon and Alloy Steel Forgings for Thin-Walled Pressure Vessels (ASTM A372/A372M-1999).

(5)

ASTM Designation A 381

Standard Specification for Metal-Arc-Welded Steel Pipe for Use With High-Pressure Transmission Systems (ASTM A381-1996).

(6)

ASTM Designation A 671

Standard Specification for Electric-Fusion-Welded Steel Pipe for Atmospheric and Lower Temperatures (ASTM A671-1996).

(7)

ASTM Designation A 672

Standard Specification for Electric-Fusion-Welded Steel Pipe for High-Pressure Service at Moderate Temperatures (ASTM A672-1996).

(8)

ASTM Designation A 691

Standard Specification for Carbon and Alloy Steel Pipe, Electric-Fusion-Welded for High-Pressure Service at High Temperatures (ASTM A691-1998).

(9)

ASTM Designation D638

Standard Test Method for Tensile Properties of Plastics (ASTM D638-1999).

March 2006

C3.3-1

A Modified Criterion for Evaluating the Remaining Strength of Corroded Pipe (AGA-PR3-805-1989).

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Documents Incorporated by Reference into Title 49 CFR Part 192, Appendix A, as amended through June 14, 2004

Title (applicable edition)

(10) ASTM Designation D2513-1987 applies to §192.283(a)(1)

Standard Specification for Thermoplastic Gas Pressure Pipe, Tubing and Fittings (ASTM D25131987).

(11) ASTM Designation D2513-1999

Standard Specification for Thermoplastic Gas Pressure Pipe, Tubing and Fittings (ASTM D25131999).

(12) ASTM Designation D 2517

Standard Specification for Reinforced Epoxy Resin Gas Pressure Pipe and Fittings (D 2517-2000).

(13) ASTM Designation F1055

Standard Specification for Electrofusion Type Polyethylene Fittings for Outside Diameter Controlled Polyethylene Pipe and Tubing (F10551998).

D.

The American Society of Mechanical Engineers, International (ASME) and American National Standards Institute (ANSI)

(1)

ASME/ANSI B16.1

Cast Iron Pipe Flanges and Flanged Fittings (ASME B16.1-1998).

(2)

ASME/ANSI B16.5

Pipe Flanges and Flanged Fittings (ASME/ANSI B16.5-1996, including ASME B16.5a-1998 Addenda).

(3)

ASME/ANSI B31G

Manual for Determining the Remaining Strength of Corroded Pipelines (ASME/ANSI B31G-1991).

(4)

ASME/ANSI B31.8

Gas Transmission and Distribution Piping Systems (ASME/ANSI B31.8-1995).

(5)

ASME/ANSI B31.8S

Supplement to B31.8 on Managing System Integrity of Gas Pipelines (ASME/ANSI B31.8S-2002)

(6)

ASME Boiler and Pressure Vessel Code, Section I

Rules for Construction of Power Boilers (ASME Section I-1998).

(7)

ASME Boiler and Pressure Vessel Code, Section VIII, Division 1

Rules for Construction of Pressure Vessels (ASME Section VIII, Division 1-2001).

(8)

ASME Boiler and Pressure Vessel Code, Section VIII, Division 2

Rules for Construction of Pressure Vessels: Alternative Rules (ASME Section VIII Division 2-2001).

(9)

ASME Boiler and Pressure Vessel Code, Section IX

Welding and Brazing Qualifications (ASME Section IX-2001).

E.

Manufacturers Standardization Society of the Valve and Fittings Industry, Inc. (MSS)

(1)

MSS SP44-96

F.

National Fire Protection Association (NFPA)

(1)

NFPA 30

Flammable and Combustible Liquids Code (NFPA 30-1996).

(2)

ANSI/NFPA 58

Standard for the Storage and Handling of Liquefied Petroleum Gases (NFPA 58-1998).

(3)

ANSI/NFPA 59

Standard for the Storage and Handling of Liquefied Petroleum Gases at Utility Gas Plants (NFPA 59-1998).

March 2006

Steel Pipe Line Flanges (MSS SP-44-1996 including 1996 errata).

C3.3-2

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Documents Incorporated by Reference into Title 49 CFR Part 192, Appendix A, as amended through June 14, 2004

Title (applicable edition)

(4)

ANSI/NFPA 70

National Electrical Code (NFPA 70-1996).

G.

Plastics Pipe Institute (PPI)

(1)

PPI TR-3/2000

H.

National Association of Corrosion Engineers International (NACE)

(1)

NACE Standard RP-0502-2002

I.

Gas Technology Institute (formerly Gas Research Institute (GRI)

(1)

GRI 02-0057

Policies and Procedures for Developing Hydrostatic Design Bases (HDB), Pressure Design Bases (PDB), and Minimum Required Strength (MRS) Ratings for Thermoplastic Piping Materials (PPI TR-3/2000-Part E only, “Policy for Determining Long Term Strength (LTHS) by Temperature Interpolation.” Pipeline External Corrosion Direct Assessment Methodology (NACE RP-0502-2002). Internal Corrosion Direct Assessment of Gas Transmission Pipelines—Methodology (GRI 02/0057-2002).

DETERMINATION OF HIGH CONSEQUENCE AREAS HCAs must be determined using one of two allowable methods described in 49 CFR 192.903, using the process for identification described in 49 CFR 192.905 and guidance provided in an advisory bulletin (68 FR 42456, July 17, 2003). The length of the pipeline subject to pipeline integrity assessments and mitigation actions – the pipeline segment encompassed by the HCA – is also shown in these figures. Where a potential impact circle is calculated using either Method 1 or Method 2 to establish an HCA, the length of the HCA extends axially along the length of the pipeline from the outermost edge of the first potential impact circle that contains either an identified site or 20 or more buildings intended for human occupancy to the outermost edge of the last contiguous potential impact circle that contains either an identified site or 20 or more buildings intended for human occupancy. The regulations also allow operators to prorate the number of buildings within an impact circle until 2006. This exemption was intended to relieve the data collection burden on operators of existing pipelines but should not be applied to the new pipeline construction proposed for this Project. Pipeline operators are not required to use the same method along the entire length of any pipeline.

March 2006

C3.3-3

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR

Method 1. HCAs are defined in 49 CFR 192.903 as an area defined as: (i)

A Class 3 location, or (ii) A Class 4 location, or

(iii)

Any area in a Class 1 or Class 2 location where the potential impact radius is greater than 660 feet (200 meters), and the area within a potential impact circle contains 20 or more buildings intended for human occupancy (unless the exception in paragraph 4 applies), or

(iv)

The area within a potential impact circle containing an identified site.

PIR > 660 ft (200 m) PIR < 660 ft (200 m)

660 ft (200 m)

Class 4 Location

Class 3 Location

Pipeline

Cluster of > 20 homes

Class 1 or 2 Locations

Example of High Consequence Areas using Method 1

Method 2. The area within a potential impact circle containing: (i)

(ii)

20 or more buildings intended for human occupancy, unless the exception in paragraph (4) applies; or An identified site.

Paragraph (4) Exception: If the radius > 660 feet (200 m), the HCA may be identified based on a prorated number of buildings intended for human occupancy within 660 ft from the centerline of the pipeline until December 17, 2006. This exception was not intended for use for new pipelines. HCA

HCA Cluster of > 20 homes

PIR Cluster of > 20 homes

Pipeline

Example of High Consequence Areas using Method 2

March 2006

C3.3-4

Cabrillo Port Liquefied Natural Gas Deepwater Port Revised Draft EIR