Dp15k

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

DESIGN PRACTICES Section

Page

XV-K

1 of 17

Date December, 1999 Changes shown by ➧

CONTENTS Section

Page

SCOPE ............................................................................................................................................................ 3 REFERENCES ................................................................................................................................................ 3 DESIGN PRACTICE ............................................................................................................................... 3 INTERNATIONAL PRACTICE ................................................................................................................ 3 OTHER LITERATURE ............................................................................................................................ 3 GENERAL ....................................................................................................................................................... 3 RISK FROM FLAMMABLE AND TOXIC MATERIAL RELEASES ................................................................ 3 TOXIC GAS DETECTORS.............................................................................................................................. 4 APPLICATION IN H2S DETECTION....................................................................................................... 5 APPLICATION IN HF DETECTION ........................................................................................................ 7 PERIMETER MONITORING................................................................................................................... 7 FLAMMABLE GAS DETECTORS .................................................................................................................. 8 TYPES OF FLAMMABLE GAS DETECTORS ........................................................................................ 8 LOCATING HYDROCARBON DETECTORS ......................................................................................... 8 ALTERNATIVES TO TOXIC AND FLAMMABLE GAS DETECTORS ......................................................... 10 FIRE DETECTORS ....................................................................................................................................... 10 SMOKE DETECTORS .......................................................................................................................... 11 HEAT DETECTORS ............................................................................................................................. 11 FLAME DETECTORS ........................................................................................................................... 11 FLAME DETECTOR PERFORMANCE................................................................................................. 12 DUAL SENSOR PERFORMANCE ....................................................................................................... 13 DETECTOR RESPONSE ..................................................................................................................... 13 WHEN TO INSTALL FIRE DETECTORS? ........................................................................................... 14 DESIGN GUIDELINES FOR DETECTORS .......................................................................................... 14 TABLES Table 1 Table 2

Substances Exhibiting Significant UV Absorption ............................................................. 12 Potential UV and IR Flame Detector False Alarms ........................................................... 13

FIGURES Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

Radiation Spectra from a Hydrocarbon Fuel Fire.............................................................. 15 Flame Detector vs Sunlight Transmission......................................................................... 15 Typical Flame Detector Response to Gasoline Pan Fire .................................................. 16 Flame Detection Distance vs. Fire Size ............................................................................ 16 Flame Detection Distance vs. Viewing Angle.................................................................... 17

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES Section

Page

XV-K

2 of 17

Date

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

December, 1999

Revision Memo 12/99 Page 3 Page 5 Pages 5 and 6

Page 7 Page 8

Page 9

Page 10 Page 11 Page 16

Page 17

Updated references. Added clarification to second bullet on availability of toxic gas detector technologies Added "Potential" to the definitions of the H2S Hazard Areas. Changed the upper limit of the Low Potential H2S Hazard Area and the lower limit of the Medium Potential H2S Hazard Area from 1000 vppm to 250 vppm along the lines of the EBSI proposal in report MR.35DQ.97. Clarified definitions of all H2S Hazard Areas. Added flowchart to aid in the determination of the potential H2S hazard areas. Added section on Application in HF Detection. Updated the perimeter monitoring technology. Updated the catalytic point detector technology. Updated the IR Absorption point detector technology. Updated the Open-Path IR detector technology. Removed the following incomplete and unnecessary sentence: "For a fixed size hole, the amount of gas released at 200 psig (14 barg) is roughly equal to the amount of vapor released from liquid butane at 100°F (38°C)." Added "higher risk" to definition of materials involved. Added fifth bullet: "Control valves operating above 200 psig (14 barg)" to the list of leak sources. Clarified the air monitoring recommendation for "Cooling tower exhaust' (5th bullet) and removed the following statement: "Hydrocarbon analyzers in the return water may be substituted for hydrocarbon vapor detectors. Added "toxic gas" to detectors that are not required. Added statements related to plastic tubes that may be used as alternatives to traditional fire-sensing devices. Added "Flame Detector" to title of Figure 3 for clarification and removed the "6 in. (150 mm) Diameter". Also corrected Figure 4 distance scale's lowest value to 10 from 1. Corrected the degree signs on the viewing angles shown.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

EXXON ENGINEERING

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

DESIGN PRACTICES Section

Page

XV-K

3 of 17

Date December, 1999

SCOPE This section describes the use of detection systems for flammable or toxic gas releases and for fires. Early detection is a key to controlling a release or fire and mitigating its effects.

REFERENCES

➧

DESIGN PRACTICE X-G

Pumps - Shaft Sealing

INTERNATIONAL PRACTICE IP 4-3-1, IP 4-3-3, IP 10-8-1, IP 15-8-3, IP 16-6-1,

Plant Buildings for Operation and Storage Shelters for Process Analyzers Combustion Gas Turbines Combustible and Toxic Vapor Detection Systems Substation Layout

OTHER LITERATURE Outdoor Installation of Flammable Gas Detectors, ER&E Report No. EE.77E.78. Guidelines on Safety Requirements for H2S Monitoring, ER&E Report No. EE.15E.85. Guidelines for the Use of Fire Detectors, ER&E Report No. EE.43E.87. Marketing Terminal Fire Detection Systems - Engineering Standard, ER&E Report No. EE.2M.90. National Fire Alarm Code, National Fire Protection Association - NFPA 72 HF Alkylation Consequence Reduction Technology; Best Proven Reasonable Design for Detection, Automation and Initiation, ER&E Document No. 99SR37. Hydrogen Sulfide Handling Practices at refineries and Chemical Plants: A Benchmarking Study, EBSI Report No. MR.35DQ.97. Pump Sealing Technology Manual, ER&E Report No. TMEE-023. Proof of Principle, Development and Field Test of a New Commercially Available Hydrogen Sulfide Open Path Monitor Using Tunable Diode Lasers, ER&E Report No. EE.125E.98.

GENERAL While the goal in design and operation of refineries and chemical plants is a plant free of gas releases and fires, the risk of a release or fire is never zero. Some areas of the plant have a greater risk than others. The risk is influenced by the material contained, the conditions under which it is contained, and the equipment and procedures used to contain it. The risk from flammable or toxic gas releases or fires can be reduced by early detection and subsequent mitigating actions. Detection and response to a release or fire, is often effectively provided by the process operators, in conjunction with their daily activities. However, sometimes the nature of the event is such that a more rapid detection and response is required than can be expected from the operator. Additionally, the increased use of remote monitoring and automation can reduce the time operators spend on the unit, possibly increasing the time to detect a leak. When more rapid detection is necessary than can be provided by the process operator, an instrumented monitoring system can be used.

RISK FROM FLAMMABLE AND TOXIC MATERIAL RELEASES The risk from a flammable or toxic material release is the combination of the probability of a leak occurring and the potential consequences should the leak occur. In determining the probability of a leak, the type of operation, the operating conditions, and the equipment involved are all considered. In determining the consequences, the material contained, the size of the leak, the proximity of the leak to other equipment or the fence line, and speed with which the leak might be detected without instruments are all considerations. All releases occur as a result of some failure. These failures can generally be classified as follows:

• •

Human errors (e.g., failure to close a drain valve, incorrect valve opened).

•

Passive equipment failures (e.g., pipe gasket failure or pipe rupture).

Active equipment failures (e.g., pump seal or bearing failure).

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES Section

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS

Page

XV-K

4 of 17

Date

PROPRIETARY INFORMATION - For Authorized Company Use Only

December, 1999

EXXON ENGINEERING

RISK FROM FLAMMABLE AND TOXIC MATERIAL RELEASES (Cont) In addition to this general classification, data on typical refinery and chemical plant manufacturing indicate certain areas which have historically caused release problems. Areas of higher probability include:

•

Open sampling or drain points.

• •

Equipment opened for maintenance.

•

Atmospheric seals on sulfur units.

• •

Flange gaskets.

•

Small lines in vibrating service.

• •

Lines in corrosive/erosive service.

Seals on rotating equipment, i.e., pumps and compressors.

Tanks, due to overfilling.

Plate and frame heat exchangers.

When a leak does occur the consequences may range from inconsequential effect to personnel health effects, fire, or vapor cloud explosion. The severity of the consequences depends on the following factors:

• •

The material leaked, specifically its toxicity or flammability.

•

The proximity of the leak to in-plant personnel, ignition sources, and the property line.

• •

The ability to detect the leak and stop or reduce the rate.

The mass rate leaked which is a function of the material vapor pressure, the operating temperature and pressure, and the cross-sectional area of the opening.

Atmospheric conditions, especially wind velocity and direction.

Toxic gas, flammable gas, and fire detectors can reduce the risk of a release by reducing the severity. Severity is reduced because faster response is possible. The response may be carried out by the operator or through an automated response. Automated responses include stopping and/or isolating equipment and automated foam, water spray, or deluge systems. While vapor and fire detectors reduce risk by decreasing the severity of a release, other design changes and/or instrumentation may provide equivalent risk reduction by reducing the probability of a release. Designs which reduce the probability of a release may be considered as an alternative to detectors, e.g., pump seal leak detection in place of toxic or flammable vapor detectors. Also see ALTERNATIVES TO TOXIC AND FLAMMABLE GAS DETECTORS later in this section.

TOXIC GAS DETECTORS Toxic gas often is present in refinery or chemical plant streams. Hydrogen sulfide (H2S) is often present in refinery streams, but hydrogen fluoride (HF), chlorine (CI2), carbon monoxide (CO), anhydrous hydrogen chloride (HCI), bromine (Br2), and ammonia (NH3) may also be present at a site. Small releases of toxic gas are a threat to personnel working directly where the leak occurs. Medium releases may affect other personnel working on the unit but not directly at the release source. Large releases may produce toxic concentrations outside the unit or even outside the plant fence line. Toxic releases may be generally classified by the cause of the release as follows:

•

Releases resulting from venting or draining equipment, typically for sampling or as part of maintenance. Toxics are not intentionally vented but equipment may not be properly gas freed or the toxic concentration may not be properly identified.

•

Releases from "active" equipment. From previous experience, active equipment which is a likely release point includes: pump and compressor seals, sulfur plant seals, and H2S incinerators which have lost flame. Since the amount of active equipment is low, the number of likely release points is relatively few and known in advance.

•

Releases from "passive," higher integrity equipment, such as pipe flanges and pipe/vessel wall ruptures. These are lower probability release points than for active equipment. Unlike active equipment, the amount of passive equipment is very large and the most likely release points are not easily identified. Different detection methods are employed for each type of release.

•

Since releases which occur from opening equipment immediately expose personnel, the absence of toxics must be confirmed beforehand. Effective control of releases is by training, effective work permits including contractor monitoring, personal or portable gas detectors, and the use of personnel protection like breathing apparatus when appropriate. Sampling systems should be designed for closed transfers whenever possible.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

DESIGN PRACTICES Section

Page

XV-K

5 of 17

Date December, 1999

TOXIC GAS DETECTORS (Cont) ➧

➧ S

•

Releases from higher potential, active equipment should be monitored with toxic gas detectors. Toxic gas detectors available for use in refineries and chemical plants are based on a wide range of technologies including electrochemical sensors, solid state sensors, ion mobility sensors, and FTIR and laser diode open path systems. Detectors should be located close to the potential leak source. This will signal the problem to those in the general area or those who might enter the area.

•

Releases from higher integrity, passive equipment usually do not require monitoring. Because of the lower likelihood of a release and the very large area covered by passive equipment (e.g., pipe flanges) complete monitoring is not possible. Manifolds with a large number of valves and flanges, representing a concentration of leak sources, may be locally monitored if the toxic concentrations are high.

APPLICATION IN H2S DETECTION H2S is the toxic gas most commonly handled in refineries. Three hazard levels are defined below for H2S operation: high, medium, and low. A flowchart that aids in the determination of the potential hazard areas is also provided. High potential hazard areas contain BOTH a potential release source AND high H2S concentrations. Examples of potential release sources include floating roof tank seals, pump and compressor seals, sulfur plant seal legs, H2S incinerators, vents, drains and sampling connections. High H2S concentrations are defined as those where the H2S concentration in the vapor phase of the contained stream is above 2 volume%. If H2S is dissolved in a liquid, and an isenthalpic flash of the liquid at atmospheric pressure produces a vapor with more than 2 volume% H2S, the stream would also be considered high concentration. Typical sources with higher release potential and high H2S concentrations include, but are not limited to:

• •

Amine regenerator overhead pumps.

•

Compressors for H2S rich gas.

• •

Claus plant atmospheric seals.

•

Burners of H2S combustors and incinerators.

Sour water stripper overhead pumps.

Air intakes in Claus plants.

•

Water seals on flare drums which often have a high H2S concentration. Medium potential hazard areas require BOTH a potential release source (see examples under High Potential Hazard Areas) AND contained H2S concentrations between a minimum of: a)

250 volume ppm, if potential exposure involves possible direct exposure to the vessel vapor space, such as sampling or opening vessels, or b) 500 volume ppm, for dispersed process exposure not involving possible direct exposure to the vessel vapor space and a maximum of 2.0 volume%. Typical examples for such areas include:

•

Sour water drawoff facilities.

• •

Sour crude tanks and slop tanks.

Asphalt and heavy fuel oil tanks. Low potential hazard areas contain EITHER a potential release source (see examples under High Potential Hazard Areas) with H2S concentrations of between a minimum of 10 vppm and a maximum of: a) 250 volume ppm, if potential exposure involves possible direct exposure to the vessel vapor space, such as sampling or opening vessels, or b) 500 volume ppm, for dispersed process exposure not involving possible direct exposure to the vessel vapor space, OR high H2S concentrations within passive, high integrity equipment. High H2S concentrations are defined as those where the H2S concentration in the vapor phase of the contained stream is above 2 volume%. If H2S is dissolved in a liquid, and an isenthalpic flash of the liquid at atmospheric pressure produces a vapor with more than 2 volume% H2S, the stream would also be considered high concentration. Examples of passive, high integrity equipment includes process vessels, piping, flanges, and other static equipment (no moving parts).

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES Section

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS

Page 6 of 17

XV-K Date

PROPRIETARY INFORMATION - For Authorized Company Use Only

December, 1999

EXXON ENGINEERING

TOXIC GAS DETECTORS (Cont) Typical examples of Low Potential Hazard areas include:

• •

Light ends pumps handling fluids with low H2S concentrations (e.g., untreated propane).

•

Heavy fuel loading.

•

Acid gas transfer lines (typically high H2S concentration but no potential release source).

Sour fuel gas at fired boilers as long as the gas contains less than 500 vppm H2S.

FLOWCHART TO DETERMINE H2S POTENTIAL HAZARD AREAS Start

Is H2S vapor concentration > 20,000 vppm? (Note 2)

No

Is there a potential release source? (Note 1)

Yes

Is H2S vapor concentration > 20,000 vppm? (Note 2)

Yes

High Potential Hazard Area

Yes

Medium Potential Hazard Area

No Is H2S vapor concentration > 500 vppm? (Note 2)

Yes

Yes

No

No

Is H2S vapor concentration > 250 vppm? (Note 2)

Yes

No

No

Negligible Potential Hazard Area

No

Is H2S vapor concentration > 10 vppm? (Note 2)

Note 1: Potential Release Sources are defined above, under "Application of H2S Detection" Note 2: H2S Concentrations are defined above, under "Application of H2S Detection"

Is direct exposure possible?

Yes

Low Potential Hazard Area

DP15Kf0

Additional information on H2S hazard levels may be obtained from the EBSI Report No. MR.35DQ.97, titled "Hydrogen Sulfide Handling Practices at refineries and Chemical Plants: A Benchmarking Study." Fixed H2S detectors are recommended for use within high potential hazard areas. The detection points should be located at each high potential release point. Ideally, the detectors should be located within 5 ft (1.5 m) of the leak point, and between the centerline and 2 ft (60 cm) below. Access may require some adjustment to the location. The detector should alarm between 10 and 20 ppm. When multiple sources are located together, e.g., pumps, a single monitor can service two sources, provided the sources are located within 10 ft (3 m) of each other. EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

DESIGN PRACTICES Section

Page

XV-K

7 of 17

Date December, 1999

TOXIC GAS DETECTORS (Cont) Fixed detectors are not recommended for medium and low potential hazard areas. While fixed detectors are not recommended for medium hazard areas, other protective measures may be specified by the local SOC. The area may be marked, personal H2S detectors may be required, and/or authorization before entering the area might be required. In the low hazard area, it is unlikely that a person will be exposed to dangerous concentrations as a result of a leak. Dispersion calculations for a 0.5 ft3/sec (14 liter/sec) leak of 1000 ppm H2S show that the concentration is reduced to 100 ppm in 3 ft (1 m) and 14 ppm in 10 ft (3 m). The alarm system for high potential hazard areas should allow for rapid identification of the leak source. Three options, which will allow rapid identification, are: 1. Locate all detector outputs in the control room on a panel or distributed control computer display. The panel/display should indicate all detectors in alarm, and the first detector which went into alarm. A local common alarm should sound in the process area. 2. Locate a panel with all alarms in the field and a common alarm to the control house. The panel should indicate all detectors in alarm and the first detector which went into alarm. A local common alarm should sound in the process area. 3. Provide each detection point with a flashing light which would indicate an alarm. All detectors in the area should be tied to a common audible area alarm and a common control room alarm. An H2S detector should also be placed in the air intake duct for normally occupied buildings in close proximity to high potential hazard H2S areas. Normally occupied buildings are those to which persons are regularly assigned and in which they perform the majority of their duties such as unit operating control rooms, office buildings, laboratories, and central maintenance workshops. The local Safe Operations Committee (SOC) will determine whether areas occupied for part of the day such as local operator shelters or change houses are classified as “normally occupied." Close proximity will be determined by the size and scope of potential releases. Where release scenarios and dispersion modeling have already been carried out, the results will indicate which buildings can be subject to high concentrations. When dispersion results are not available, nominally occupied buildings within 200 ft (60 m) should include a detector. Normally H2S detectors in air inlets should alarm at 10 ppm H2S. When the building is a multi-unit control house or designated safe haven, the air system should also shut down upon alarm. The shutdown concentration may be higher than 10 ppm, provided there is an alarm at 10 ppm. Higher shutdown levels should be based on discussion with the local hygienist. In addition to locating fixed detectors at potential leak sources, other toxic gas detectors may be used for additional protection of personnel from H2S leaks in high potential hazard areas. Personal monitors may be provided to those working in the high H2S area. Although H2S has a characteristic rotten egg odor, continued exposure at low levels or exposure in conjunction with hydrocarbons may deaden the sense of smell. The personal detector provides additional protection that H2S is detected. Personal monitors may be coupled with the use of escape devices. The escape device may be a small air cylinder or a respirator to ensure that the person is not overcome while exiting. (The effectiveness of respirators decreases as H2S concentration increases.) A decision to use personal monitors and escape packs is made by the local site based on their assessment of the risk reduction it would provide at the site. ➧ S

APPLICATION IN HF DETECTION For detailed information on the latest technology on hydrogen fluoride (HF) detection refer to the ER&E Document No. 99SR37 titled "HF Alkylation Consequence Reduction Technology; Best Proven Reasonable Design for Detection, Automation and Initiation". The document includes information on detector types, range, selectivity, response time, stability, reliability, and deployment.

S

PERIMETER MONITORING

➧

Perimeter monitoring of the unit with a number of fixed monitors is usually not recommended. By placing detectors at potential high hazard leak sources, most leaks should be detected before reaching the perimeter. The remaining leaks will occur from higher integrity equipment and may still not be detected by fixed perimeter monitoring. Certain environmental conditions may allow a narrow plume to pass between adjacent detectors, or an elevated leak may not be detected by monitors at grade. Under these conditions, perimeter monitoring might even provide a false sense of security. Perimeter monitoring might be considered in special circumstances such as close proximity to a fence line or occupied buildings. If a site determines that perimeter monitoring is required, an open path detection system should be considered. These systems can provide more effective coverage than a series of point detectors and depending on the number of point detectors that would be needed, may also be more cost effective. In open path detection systems, as described in EE.125E.98, a light beam traverses a relatively long path using a transmitter/receiver system and an average concentration along the path is based on the amount of light absorbed by the component being monitored. Open path detection systems utilizing tunable diode lasers have recently been demonstrated for H2S and HF. Paths up to 330 ft (100 m) can be monitored.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES Section

Page

XV-K

8 of 17

Date December, 1999

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

EXXON ENGINEERING

FLAMMABLE GAS DETECTORS The detection of flammable vapors is similar to that for toxic vapors. The most likely release points are similar, as is the behavior of the vapor as it disperses. The concentration of flammables required for ignition is generally much higher than the concentration of toxic gas required for an acute health threat. However, the use of liquefied hydrocarbons above their normal boiling points can lead to much larger leaks and higher concentrations. Although detection is similar, the consequences of flammable hydrocarbon and toxic releases can be quite different. Since detectors are installed to assist in mitigation of releases, the different consequences of flammable gas and toxic gas releases can result in different deployment strategies.

TYPES OF FLAMMABLE GAS DETECTORS

➧

➧

➧

S

There are three distinct types of flammable gas detectors for use in refineries and chemical plants: catalytic point detector, short path infrared point detector, and infrared open path detector. There are advantages and disadvantages for each. The most common gas detector is a catalytic point detector. These detectors operate by sensing the resistance change of a heated catalytic element caused by the catalytic oxidation of flammable gases on the surface of the element. They have proven to be mechanically robust, highly specific, and are low cost. As with all point detectors, they have the disadvantage of monitoring only the surrounding air. As the area to be monitored increases, additional detectors are required. A second disadvantage of the catalytic point detector is that the catalytic element can be attacked or coated resulting in a loss of sensitivity. Halogens, hydrogen sulfide, lead compounds, and silicones can all reduce the sensitivity. To ensure that sensitivity is maintained, monthly functional checks and full calibration checks on a quarterly to semi-annually basis are typically performed. Finally, if the flammable level exceeds the upper explosive level (UEL), the reduction in oxygen reduces the catalytic conversion. When the UEL is exceeded, there may be an apparent reduction in the concentration, even registering below the lower explosive limit (LEL), due to the lack of complete catalytic conversion. A second type of point detector uses infrared (IR) absorption to detect hydrocarbons. These detectors operate by passing an infrared beam through a short path sample cell and measuring the transmitted light at the hydrocarbon absorption wavelength and at a non-absorbing reference wavelength. The ratio of the reference signal to the hydrocarbon signal is proportional to the hydrocarbon content in the beam path. The reference wavelength can also be used to continuously monitor detector performance, signaling when maintenance is needed. While the cost of the IR detector is 1.5 to 2 times greater than the catalytic detector, internal self-checking and negating the effect of contaminate gases reduces the scheduled maintenance while providing the same reliability. Unlike catalytic detectors, there is no loss in sensitivity at concentrations higher than the UEL. However, the point detector still samples only the surrounding environment, and multiple detectors are required for an area. Larger areas can be covered by the use of an open-path IR detector. These detectors use the same detection concept as the IR point detector. However, instead of using a short path, the beam is projected over 100 to 300 ft (30 to 90 m) and can be reflected using mirrors. The open path detector provides information about the average flammable concentration along the path. Therefore, the detector will produce the same response for a 5 ft (1.5 m) wide plume of 50% LEL as it will for a 25 ft (8 m) wide plume of 10% LEL. Its cost is considerably more than point detectors and therefore will prove cost effective when it replaces a number of point detectors, including their associated maintenance cost. All types of flammable gas detectors are calibrated for one composition and their response to other compositions may vary. For example, a catalytic detector calibrated to the LEL of methane will read 65% of the LEL when propane is present at its LEL. Many gases do not exhibit as pronounced a difference in sensitivity. IR detectors will have similar sensitivity issues. Where multiple gases are present, the calibration standard should be selected to allow sufficient detection for as many other potential gases as possible. Detailed requirements for calibration of flammable gas detectors are given in IP 15-8-3.

LOCATING HYDROCARBON DETECTORS Several factors must be considered when determining the need for flammable gas detectors. They include: the potential for vapor cloud formation upon release, the potential for equipment to develop a leak, and the proximity of the potential leak point to ignition sources. Contained materials with higher potential for vapor cloud formation upon release include:

•

Liquid with an atmospheric boiling point below ambient temperature, (C4 and lighter).

•

Liquid which is contained at a temperature above its atmospheric boiling point such that 25% or more vaporizes upon release to the atmosphere.

• ➧ ➧

Gas at high pressure, above 200 psig (14 barg). Materials which are contained above their autoignition temperature are not likely to form a flammable vapor cloud, since they would most likely ignite immediately upon release. Therefore monitoring is not appropriate. When equipment contains higher risk materials, i.e. materials with a higher potential for vapor cloud formation, and the equipment has a higher potential for leaks, fixed detectors are recommended. The following equipment should be considered:

•

Pump seals.

•

Compressor seals. EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

DESIGN PRACTICES Section

Page

XV-K

9 of 17

Date December, 1999

FLAMMABLE GAS DETECTORS (Cont)

➧

For pumps, a detector may be located at each seal. One detector may be used between two seals provided the detector is within 5 ft (1.5 m) of each seal. The detector should ideally be located at or below the centerline of the pump and at least 18 in. above the ground. Some changes to location may be necessary for accessibility. For compressors, a detector may be located as close as possible to each seal, at or below the centerline for heavier than air gases and above the centerline for lighter than air gases. There are other lower probability sources of hydrocarbon leaks. The leak point from these sources is not as localized as with pump and compressor seals. Generally hydrocarbon detectors are not deployed. However, when they are in close proximity to ignition sources and contain the higher risk materials described above, detectors may be considered. Lower probability sources include:

• •

Small pipes and connections subject to failure due to vibration.

•

Pipes subject to external corrosion (i.e., cooling tower overspray).

• •

Drain and sample points.

Manifolds with a large number of flanges.

Control valves operating above 200 psig (14 barg).

Detectors may be considered if these potential leak points are within 100 ft (30 m) of ignition sources such as:

➧

•

Furnaces.

• •

Hot, uninsulated metal surfaces above 600°F (315°C).

•

Other internal combustion engines.

• •

Electrical substations.

•

Areas with high levels of construction or maintenance.

Vehicle traffic.

Railroads.

• Designated smoking areas. Hydrocarbon detectors should also be used to monitor the air of enclosures to prevent the buildup of an explosive atmosphere. These detectors may also be coupled with an automated shutdown system to prevent a possible confined space explosion. Air monitoring of the following is recommended: •

Air intake to control rooms and "safe haven" buildings, per IP 4-3-1.

• •

Gas turbine acoustic enclosures, per IP 10-8-1.

•

Hydrocarbon process analyzer buildings, per IP 4-3-3.

• •

Cooling tower exhaust, monitoring for presence of hydrocarbons from a cooling water exchanger tube leak.

Gasoline octane analyzer buildings, per IP 4-3-3.

Compressor shelters, which do not provide free ventilation as per IP 4-3-1.

Detectors should also be considered for monitoring of offsite pressurized storage. The offsite location results in reduced surveillance by process operators and leaks are less likely to be detected. The very large volume of pressurized liquids could result in severe consequences. The need for flammable gas monitoring would be further increased when the pressurized storage is relatively close to the fence line and developed areas. If the storage area is diked, one detector may be placed within the dike along each wall. This will take advantage of the wall which will tend to contain the hydrocarbon vapors (excluding LNG). For sites with prevailing wind patterns, fewer detectors may be required, as determined by the local SOC. Open-path detectors may also be used to monitor the perimeter, especially when there is no dike. Hydrocarbon detectors may be set between 20 and 60% of the LEL. A setting of 40% is nominal. In some cases H2S detectors can also provide adequate warning of a hydrocarbon leak. Where H2S is present in hydrocarbon vapor streams at greater than 0.5 mole%; a 10 ppm detection of H2S typically also provides a warning of approximately 10% of the LEL. As with toxic gas detectors, the alarm system should also help the process operator to identify the source of the leak. The use of control room display, local panel, or individual flashing lights, as described in the toxic gas section, are all options.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES Section

Page

XV-K

10 of 17

Date December, 1999

S

➧

S

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

EXXON ENGINEERING

ALTERNATIVES TO TOXIC AND FLAMMABLE GAS DETECTORS There are alternatives to toxic and flammable gas detectors which might be considered in specific cases. These alternatives monitor the conditions which may lead to a release. By anticipating the release, these systems reduce the probability of a release occurring. Since they prevent, rather than mitigate a release, these systems should be applied first, whenever possible. Level detectors on atmospheric and pressurized storage are simple instruments which are used to prevent a leak or spill. Rather than overfill a tank and detect the released hydrocarbon, level instruments detect overfilling before it occurs. When used as a safety device, this level instrument should be an independent alarm, separate from the normal tank indicator level. Although small bore piping, especially when used in vibrating service, is considered a higher leak risk, proper reinforcing of the connections reduces the risk of failure. With proper support, toxic or flammable gas detection is not required. If there are a large number of small bore connections in a concentrated area, some gas detection may still be advantageous. This would be determined only on a case-by-case basis. Leaks from pump and compressor seals will also be reduced with proper seal leak detection. On pumps, a leak detection system utilizes primary and secondary seals and monitors the space between the seals for leakage. Leakage can be detected by a rise in pressure, a change in liquid level when a barrier liquid is used between the two seals, or a change in temperature when cryogenic fluids are pumped. For pumps in high hazard H2S service or C4 and lighter liquid service, hydrocarbon or toxic gas detectors are not required when dual seals with leak detection (control level 5 or 6 as defined in Section X-G) are installed. Section X-G also recommends, but does not require, bearing vibration or bearing temperature monitors. Experience has shown that the majority of pump seal failures which result in significant flammable releases or fires are the result of a bearing failure. Detecting the possibility of a bearing failure before it occurs could prevent seal leaks and possible resulting fires or vapor clouds. Therefore bearing monitors are also recommended, but not required, when gas detectors are not used. If seal leak detection and not gas detection is used, it is increasingly important to ensure proper support of all small bore piping around the pump. If the small bore piping were to fail, there would be no gas detection, so it is important to reduce the risk of failure by proper support. The use of bearing monitoring will also reduce the risk of small bore pipe failure by warning of excessive vibration. On compressors, differential pressure on seal buffer gas will warn of a potential leak. Provided that the International Practices on compressors are followed to reduce and monitor vibration, flammable gas detectors would not be required on the seals. However, toxic gas detection is still recommended for compressors because of the much lower level of leakage which is tolerable and the additional benefit provided by detecting other small leaks around the compressor.

FIRE DETECTORS Flammable gas detectors may not be appropriate for all types of hydrocarbon releases. In operations where gases may occasionally be present at 10% of LEL or more as part of normal operation, detectors could result in numerous false alarms, e.g., well head areas of offshore platforms and loading racks. In other cases, the release may readily ignite, e.g., material above its autoignition point, or high concentrations of hydrogen. In these cases, fire detection systems may be used. Early detection of fire will provide the best chance for control. Fire detectors may also be used even if hydrocarbon or alternative leak detection is already in use, when it is desirable to further reduce risk. The combination of fire detection with automated actuation of a water deluge or spray will mitigate the effect of a fire and reduce risk. Pump retrofits which might require deviation to spacing standards may be able to offset the increased risk of reduced spacing with an automatic deluge system. There are also fires which do not develop from flammable vapors which may also benefit from the use of fire detection as follows: 1. Electrical equipment fires. 2. Mechanical equipment fires as a result of failures and increased friction and heat. 3. Fires from flammable dusts or solids, as in exhaust ducts or filter baghouses. Fire detectors can be classified as one of three broad types: 1. Smoke Sensing - These detectors respond to the presence of smoke particles. These detectors are primarily used indoors and in enclosed spaces. 2. Heat Sensing - These detectors respond when the sensing device becomes heated to a predetermined level. 3. Radiant Energy Sensing - These detectors respond to the radiant energy produced by burning substances. Flame detectors sense the radiant energy from open flames with background sunlight or ambient light. Spark/ember detectors sense sparks or embers in dark environments such as ductwork.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

DESIGN PRACTICES Section

Page

XV-K

11 of 17

Date December, 1999

FIRE DETECTORS (Cont) S

SMOKE DETECTORS Smoke detectors are usually employed only in confined spaces such as control house buildings or electrical substations. There have been numerous fires in electrical substations, despite proper electrical design. All new substations should have smoke detection, tied to a central control house alarm. Smoke detection should also be considered for retrofit of any existing substations. The large amount of electrical equipment in control house buildings is a potential source of fire too. Smoke detection is not employed in outdoor process areas because wind currents make detection unreliable. There are two types of smoke detectors, ionization and photoelectric. Ionization detectors are usually preferred for electrical or high energy, flaming fires which generate small smoke particles. These are specified for control rooms, IP 4-3-1, and electrical substations, IP 16-6-1. Photoelectric detectors are usually preferred for low energy, smoldering fires. Additional detail on smoke detectors is available in NFPA 72.

S

➧

HEAT DETECTORS Heat sensing detectors should be employed when the potential source of the fire is well known. Temperature sensitive devices should be placed near each fire source. These heat sensing devices are usually not subject to false alarms. This makes them ideal for initiating automated response to fires such as water deluge sprays. There are three types of heat detectors: temperature sensing point detector, rate-of-rise point detector, and temperature sensing line detector. Temperature sensing point detectors use a fusible link or plug which melts at a pre-determined temperature to detect a fire. Rate-of-rise detectors respond to a sudden increase in temperature, typically 12-15°F/min (6-8°C/min). The rate of rise is typically detected by the rapid expansion of gas within one or more detecting heads. Line detectors depend on localized heating of a temperature sensitive line. In common line detectors, heating will cause the insulation between two wires within the line to melt creating a short. Other line detectors may measure an increase in electrical resistance of wires or a decrease in optical transmission of optical fibers to detect heating. Point detectors can be applied to specific locations where the likelihood of fire is higher. Pumps with flammable liquids or liquids operating above their autoignition points and exchangers operating above the process material auto ignition point are possible locations. In certain cases air-pressurized, long-lasting, UV-resistant plastic tubes may be used as alternatives to traditional fire-sensing devices. The tubes are wrapped around the equipment they protect and, when they melt during a fire, the loss of their contained air pressure initiates an alarm. This type of fire-sensing devices may be used, for example, in remote or congested areas, in pumps with liquids above their autoignition temperature, or in pumps utilizing dual sealing systems and meeting the criteria for possible monitoring due to fire potential per the Pump Sealing Technology Manual (TMEE-023). Plastic tube-type detectors have been used in the past also in LPG and in low flash service pumps, as well as in the diked areas of spheres, on the spheres themselves, and on vapor recovery lines containing flammable air hydrocarbon mixtures. While heat sensing point detectors are typically not used outdoors, the process pressure and flammability of liquids and vapors in refineries and chemical plants often make for intense fires which can be easily sensed. Temperature sensing detectors are more often employed than rate of rise detectors. Line detectors may be useful for equipment which covers a large area and in which a fire is likely to develop anywhere in the area with equal probability. Equipment such as long conveyor belts and switch gear are possible locations for line detectors. Because temperature sensing point detectors monitor a small local area and have been demonstrated to have low false alarm rates, they have been coupled with automatic actuation of a deluge system. Two systems which have been used are:

•

Pilot heads with a fusible plug are typically located 4-6 ft (1.2-1.8 m) from possible fire sources such as higher risk pumps or exchangers. The fusible plug can be selected to melt from 135 to 580°F (57 to 304°C). The pilot heads are part of a piping system, which contains pressurized air so that the melted plug releases the air, which actuates the deluge valve.(1)

•

A similar system replaces the pilot heads with nylon tubing. The tubing, instead of a fusible plug, melts and actuates the emergency response. This is similar to the plastic tubing described above for point detectors. Note: (1)

S

Baytown refinery and chemical plant have large number of deluge water systems coupled with pilot heads which have demonstrated response to fires and few false alarms. Maintenance costs are low.

FLAME DETECTORS Radiant energy (flame) detectors should be used when there are multiple fire sources in an area. The flame detector covers a wide area and may be more economic than point sources. Flame detectors also respond faster than heat sensing or smoke detectors. When rapid spread of the fire is expected, and automated response is required, the flame detector will provide faster response than the heat sensing device.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES Section

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS

Page

XV-K

12 of 17

Date

PROPRIETARY INFORMATION - For Authorized Company Use Only

December, 1999

EXXON ENGINEERING

FIRE DETECTORS (Cont) Flame detectors respond to the radiant energy, which is emitted from all open flames. Flames emit energy in wavelengths from the ultraviolet (UV) through the infrared (IR) wavelengths, about 0.18 to 5 microns. To discriminate between fire and other sources of radiation, particularly sunlight, narrow wavelengths are monitored. The wavelengths are typically emitted from burning material but absent from other sources. Based on the wavelength monitored, there are two types of flame detectors, UV and IR. UV detectors monitor radiant energy from about 0.18 to 0.27 microns. All flames emit radiation in this region. At these wavelengths, radiation from sunlight is absorbed by the atmosphere so that there is no background solar radiation, and the detector is not affected by the sun. IR detectors monitor radiant energy from about 4.1 to 4.6 microns. All hydrocarbon flames produce CO2 which has an emission spike at this band, see Figure 1. The amount of CO2 determines the intensity of the radiation. Radiation from the sun is partially absorbed by the atmosphere in this region, see Figure 2. The partial absorption is sufficient to allow the detector sensitivity to be adjusted to zero out the effect of the sun.

S

FLAME DETECTOR PERFORMANCE Both UV and IR detectors have reduced ability to detect fires under certain conditions. UV radiation is more easily absorbed than IR. Dense smoke or high concentrations of dirt or oil on the lens reduce UV sensing more than IR. UV absorbing vapors, see Table 1, which might be present will reduce UV sensing but have no effect on IR. At least two UV instruments provide internal reflectance checks to warn that the glass is dirty. IR, unlike UV, does not have the ability to detect all types of fires. If the fire does not generate sufficient CO2, e.g., hydrogen or ammonia, an IR sensor will not detect it. IR radiation can also be absorbed by ice or water which might cover the lens. TABLE 1 SUBSTANCES EXHIBITING SIGNIFICANT UV ABSORPTION Acetaldhyde

Cumene

2-Nitropropane

Acetone

Cyclopentadiene

2-Pentanone

Acrylonitrile

O-Dichlorobenzene

Phenol

Alpha-Methylstryene

P-Dichlorobenzene

Phenyl Clycide Ether

Ammonia

Ethanol

Pyridine

Aniline

Ethyl Acrylate

Styrene

Benzene

Hydrogen Sulfide

Tetrachloroethylene

1,3 Butadiene

Methyl Methacrylate

Toluene

2-Butanone

Naphthalene

Trichloroethylene

Butylamine

Nitroethane

Vinyl Toluene

Chlorobenzene

Nitrobenzene

Xylene

1-Chloro-1-Nitropropane

Nitromethane

Chloroprene

1-Nitropropane

Both UV and IR detectors also are subject to false alarms due to the presence of non-fire UV or IR sources within the facility. Potential false alarms for each type of detector are given in Table 2.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

Section

Page

XV-K

13 of 17

Date December, 1999

FIRE DETECTORS (Cont) TABLE 2 POTENTIAL UV AND IR FLAME DETECTOR FALSE ALARMS POTENTIAL FALSE SIGNAL

UV

IR

Welding

Positioning and Aiming Voting with multiple detectors Bypass during hot work

No effect

Direct sunlight

No effect

Positioning and aiming to avoid direct beam, e.g., sunrise

Lightning

Time Delay of 3 seconds eliminates

No effect

Radiation from x-rays

Bypass during inspection work

No effect

"Friendly fires," flares, furnaces, etc.

Positioning and aiming

Positioning and aiming

UV lamps for fluorescent penetrant and magnetic particle exams

Shield UV lamps from detector sight Bypass during inspection

No effect

Flashing lights: engine timers, emergency vehicles, flash cameras

Positioning and Aiming Bypass during use

Positioning and Aiming Bypass during use

High pressure sodium, mercury vapor, or quartz-hydrogen lamps

Glass (not plastic) shields over lamps to absorb UV

No effect

Hot surfaces

No effect

Flicker circuit tests for flicker from flames which is absent from hot surfaces with steady radiation levels. However, steam clouds passing between the hot surface and the detector may simulate flicker.

Reflected sunlight

No effect

Positioning and aiming. Reflected sunlight can cause false alarms.

Although IR detectors claim to be unaffected by solar radiation, false alarms within Exxon have been attributed to the sun. Because the IR detector zeros out background solar radiation, direct or reflected sunlight, e.g., glare at sunrise or reflection off water may trigger a false alarm. False alarms are more likely when flames with low amounts of CO2 are to be detected, for example in hydrogen rich POWERFORMING streams. For these streams, the detector sensitivity is usually increased, increasing the potential for false alarm. The selection of a detector will be site dependent. It should consider the potential non-detection of the flame due to interference by dirt or vapors for UV or due to the type of flame for IR. In addition the frequency of false alarm conditions at the site and the difficulty of designing around them should be considered.

S

DUAL SENSOR PERFORMANCE The field experience with simple UV and IR devices has led to the development of dual wavelength sensor to reduce the incidence of false alarms. IR-IR detectors monitor a second narrow band typically near 3.9 microns. Hydrocarbon fires produce less radiation here than at the CO2 peak. The two bands are compared to increase the rejection of false signals due to flickering radiation from flashing lights, reflected sunlight or heated equipment with flickering. UV-IR detectors combine standard single band UV and IR sensors. Fire must be sensed by each device to provide an alarm. Since the false alarms from UV and IR sensors are generally mutually exclusive, the number of false alarms is reduced. Generally dual sensors type detectors are preferred as field experience has shown that they do reduce the number of false alarms.

S

DETECTOR RESPONSE The response of a flame detector will affect its placement. For a given detector, the response time will be a function of the size of the fire and the distance to the fire. Response time usually quoted by vendors based on detection of a 1 ft2 (0.1 m2) gasoline fire in a shallow pan. Figure 3 shows a typical response time to a small gasoline pan fire (EE.43E.87). Each detector has a minimum response time below which reducing the distance to the flame has no effect.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES Section

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS

Page

XV-K

14 of 17

Date December, 1999

PROPRIETARY INFORMATION - For Authorized Company Use Only

EXXON ENGINEERING

FIRE DETECTORS (Cont) As the size of the fire increases, it can be detected with no change in response time at increasing distances. Quadrupling the diameter of the fire roughly doubles the distance at which it can be detected, at the same response time, see Figure 4. Flame detectors should be able to detect a 1 ft2 (0.1 m2) pan fire at 35 ft (10 m) in ten seconds (EE.43E.87). Most commercial units are capable of this. Detector response is also effected by the position of the fire within the detector field of view. Response times or distances are quoted for fires in the direct line of sight of the detector. However, detectors have a field of view of at least 90°, and up to 120°. Sensitivity is reduced as the fire moves away from the central axis, as shown in Figure 5 for a typical detector. In Figure 5, the detection distance of the "standard" fire at 45° is half that of the straight-ahead distance (EE.43E.87).

S

WHEN TO INSTALL FIRE DETECTORS? The need for flame or heat sensing detectors is determined by reviewing the potential hazards, risk exposure, and appropriate response to fire. The following conditions would support the installation of fire detectors:

•

High fire risk facility; quick response is needed to avoid rapid escalation and endangerment of personnel or equipment.

• •

Sufficient personnel not available for regular surveillance.

•

Capability exists for prompt response to alarm: shutdown, isolation, and fire suppression.

Early warning by other means not feasible: gas detectors, seal leak detectors, protective equipment systems, process instrumentation, etc.

•

Ongoing maintenance support available in the case of UV and IR detectors. The primary use of flame detectors within Exxon has been for truck loading racks for flammable products such as gasoline. When the loading operation is attended only by the driver, the detection system actuates a fire suppression system to automatically control and extinguish the fire. Detectors have also been applied in a limited way for special situations in the refinery. Equipment containing hydrocarbons above their autoignition point, which historically was subject to periodic leaks and fires, have been placed under surveillance. In this case, the detectors serve an alarm function only. Heat sensing devices in combination with deluge systems have been successfully used with low maintenance costs and essentially zero false alarms both onsite and offsite.

S

DESIGN GUIDELINES FOR DETECTORS When a fire detection system is needed, the following guidelines should be followed to ensure acceptable performance:

• •

Review possible fire scenarios: what fuels are involved, where might the fire start, how fast might it spread.

•

When a flame detector is used, a dual sensor IR-IR or UV-IR flame detector, is preferred to reduce the potential for false alarm and is required when the detector will automatically activate a suppression system.

•

IR flame detectors are preferred for hydrocarbons. When the fuel contains little or no carbon, a single UV detector or heat detector is preferred. Heat sensing devices are viable alternatives in either case provided the potential flame location is well known and the sensing device can be located nearby.

•

Flame detectors should be located no greater than 35 ft (10 m) from possible fire sources. At 35 ft (10 m), the detector should respond in ten seconds to a 1 ft2 (0.1 m2) pan fire of the expected material on fire.

•

Flame detectors should be positioned to see the base of the fire not just the flames above it.

•

Enough flame detectors must be deployed to avoid blind spots and to account for loss in sensitivity away from the detector's central axis.

•

To avoid false alarms from sources outside the risk area, flame detectors should not have a view of the horizon.

Where the rapid spread of the fire is likely, automatic actuation of protective systems should be specified.

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

DESIGN PRACTICES

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS

EXXON ENGINEERING

Section

Page 15 of 17

XV-K Date

PROPRIETARY INFORMATION - For Authorized Company Use Only

December, 1999

FIGURE 1 RADIATION SPECTRA FROM A HYDROCARBON FUEL FIRE

Watts Micron x Steradian

2

Single Frequency Infrared Sensor Response

Intensity

Spectra Radiant

1

0

0

1

2

DP15KF01

3

4

5

6

7

8

Wavelength (Microns)

FIGURE 2 FLAME DETECTOR VS SUNLIGHT TRANSMISSION

Ultraviolet

Visible

Infrared

100 75 50 25 0 0.2 Ultraviolet Sensor Response

0.3

0.4

0.5 0.6 0.7 0.8 0.9 1.0 Wavelength (Microns)

Solar Radiation Reaching The Earth

1.5

2.0

3.0 Infrared Sensor Response

DP15KF02

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

4.0

5.0

DESIGN PRACTICES Section

SAFETY IN PLANT DESIGN

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS

Page

XV-K

16 of 17

Date

PROPRIETARY INFORMATION - For Authorized Company Use Only

December, 1999

➧

FIGURE 3 TYPICAL FLAME DETECTOR RESPONSE TO GASOLINE PAN FIRE

Distance to Fire, ft. (1 ft = 0.305 m)

30

20

10

0

0

2

4

6

Response Time, Sec

DP15KF03

FIGURE 4 FLAME DETECTION DISTANCE VS. FIRE SIZE

Distance, ft (1 ft = 0.305 m)

1000

100

10

1

DP15KF04

10

100

Fire Diameter, ft (1 ft = 0.305 m)

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

EXXON ENGINEERING

SAFETY IN PLANT DESIGN EXXON ENGINEERING

FLAMMABLE GAS, TOXIC GAS, AND FIRE DETECTION SYSTEMS PROPRIETARY INFORMATION - For Authorized Company Use Only

DESIGN PRACTICES Section

Viewing Angle 0° 100%

75%

50%

Normalized Detection Distance

30°

45°

17 of 17

Date

FIGURE 5 FLAME DETECTION DISTANCE VS. VIEWING ANGLE

15°

Page

XV-K

25%

DP15KF05

EXXON RESEARCH AND ENGINEERING COMPANY - FLORHAM PARK, N.J.

December, 1999