Sour water strippers: design and operation Simpler, more cost-effective sour water stripper designs can outperform more generally generally accepted designs in the refining industry NORMAN LIEBERMAN Process Improvement Engineering
I
n 1965, while working as a three designs reects correct control due to chlorides. 1 One process design engineer with process engineering principles. hundred ppm of NH3 is excesthe American Oil Renery in The two sour water strippers sive for desalter operation. Whiting, Indiana, I designed a that I was troubleshooting in When sour water stripper sour water stripper to remove India both reected common bottoms are used as make-up ammonia, hydrogen sulphide but incorrect process engineer- to the hydrotreater efuent and phenols from coker and ing design. wash water, an NH 3 content of a few hundred ppm is acceptFCC efuent sour water. Many years later, in 2012, I was Purpose of sour water strippers able. After all, the recirculated water has over working in two reneries in Renery sour water originates wash India, where I was trouble- largely from delayed cokers, 10 000 ppm of NH 3 , so the NH 3 shooting sour water strippers hydrodesulphuriser reactor content of the make-up water in both plants. None of these efuents, uid catalytic crack - is irrelevant, as long as over three strippers had worked ing units and visbreaker 90% of the NH3 in the sour properly, in that the residual fractionators. The main water stripper column is NH3 in the stripped sour water contaminants are NH3 and H2S. stripped out. was excessive. All three strip - Sour water stripper bottoms pers suffered from low tray are reused in two places: Modern conventional efciency, as well as reboiler Wash water for the crude stripper design and feed preheater exchanger desalter on a once-through Figure 1 shows the current fouling. While conventional, all basis, to remove chloride salts conventional design used at three stripper designs were that promote HCl evolution in most locations. The operating needlessly complex. This exces - the crude tower overhead conditions shown are typical sive complexity contributed to condensers of many reneries. Feed is their poor performance in the Make-up water for hydrohydro - heated by exchange with the eld. As a result of these de - treater efuent recycle wash stripper bottoms in E-1. The ciencies, the stripper bottoms water to remove ammonia reboiler duty (E-3) is required were diverted to the efuent sulphide salts that plug the for three purposes: • Heat feed from 180°F (82°C) water treatment plant rather downstream condensers. than being reused for their When sour water stripper to the 250°F (120°C) bottoms intended purpose as crude unit bottoms are used in the crude temperature desalter wash water. desalter, the NH3 content • Generate internal reux at There are three general ways should be about 10 to 20 ppm. tray 32 to design renery sour water Higher NH3 levels interfere • Break the chemical bonds strippers, but only one of these with crude unit corrosion between water, NH 3 and H2S.
www.digitalrefining.com/article/1000799
PTQ Q2 2013 1
Older conventional stripper Figure 2 is based on older Ammonia gas to sulphur plant conventional sour water stripper designs. In this version, there are only 16 trays rather 13 PSIG 190ºF than 40. Field observations Reflux 140ºF indicate that 15 or so stripping (40) trays are sufcient to remove E-2 over 99% of the NH3 from the (33) 200ºF E-1 sour water and an even greater Sour 90ºF 180ºF water (32) percentage of the H 2S. Note P-2 feed that reducing the reux drum temperature much below 165-170°F (74-77°C) results in salt plugging. The 190°F (88°C) reux drum temperature is (1) likely conservative (high) and results in increased water in the sulphur plant feed and E-3 155ºF 250ºF Steam thus more condensation in the sulphur plant feed NH 3 gas knock-out drum (and thus water that has to be recycled back to the sour water stripper). P-1 The design shown in Figure To desalter and 2 was common in the 1960s. hydrotreaters However, it also suffers from the same heat balance drawFigure 1 Conventional sour water stripper design used in newer units backs and needless The reux is generated in the efciency for trays 30-32? A complications as seen in E-2 circulating cooler. The single eld measurement is Figure 1. In other words, a lot reboiler duty is adjusted so as worth a thousand computer of equipment is added to to control the moisture content simulation calculations. I generate reux when no frac(by temperature) of the ammo- conducted a eld test by shut- tionation is required between nia-rich gas owing to the ting off the circulating reux the feed and overhead prodsulphur plant thermal reactor. (E-2) and bypassing the feed uct. Again, the only purpose This design, while conven- preheat exchanger (E-1) to of the tower is to strip out the tional, is not logical. Why maintain a constant reboiler NH3 and the H2S. preheat the feed onto tray 32 duty and a constant tower top with E-1, and then remove this temperature. The effect on the Correct stripper design heat in E-2? It would have the NH3 in the stripper bottoms In 1969, while working for the same effect on the stripping was quite negligible. And why now vanished Amoco trays to: should it change? Will not the International Oil Company in • Bring in the feed colder onto stripping factor (the ratio of the UK, I designed a sour tray 32 (see Figure 1) vapour to liquid) for the strip- water stripper that eliminated • Reduce the heat extracted in ping trays remain unaltered? the unnecessary features of the E-2, the circulating cooler, to The conclusion is that the unit shown in Figure 2. offset the colder feed. preheat exchanger (E-1), the Figure 3 shows the essentials What would the effect be of circulating reux loop (E-2) and of a correct sour water stripper eliminating the E-1 feed trays 33-40 serve no actual engi- design. Feed is brought in at preheater on the stripping neering purpose. ambient conditions (70-100°F,
2 PTQ Q2 2013
www.digitalrefining.com/article/1000799
21-38°C) from the sour water feed tank. To heat the feed from 90°F (32°C) to 250°F (120°C) requires about 16 wt% steam ow, or about 1.3-1.4lb of steam per gallon of stripper bottoms, which is close to a typical design stripping steam ratio for sour water strippers. The E-1 feed preheater, reux pump (P-2) and the reux cooler (E-2) shown in Figure 2 are all eliminated. How, then, does one know that the design shown in Figure 3 will work? Because it was built this way (at the Amoco renery in Milford Haven, Wales, UK) in 1970, where it worked just as well as the conventional design shown in Figure 2. Two-stage sour water stripper Figure 4 shows a sour water stripper with a side draw-off. The partly stripped sour water is extracted from tray 8 and directed to the hydrotreaters for use as make-up water in the salt (NH4HS) removal step of the reactor efuent. Completely stripped water from the sour water stripper bottoms is sent to the crude desalter. While apparently non-conventional, this design is just a direct application of the “lean, semi-lean amine regeneration design” that is often used in the natural gas industry to save energy (reboiler steam). 2 The design of such two-stage rich amine regenerators is fairly conventional technology. The objective is to produce a very low residual H2S content in the lean amine, at the expense of a higher residual H2S content in the semi-lean amine. Tray efficiency For one of the sour water
www.digitalrefining.com/article/1000799
E-2
13 PSIG 190ºF
Ammonia gas to sulphur plant
200ºF
(16)
Sour water feed
E-1 90ºF
(13)
180ºF
P-2
(12)
(1)
E-3 155ºF
250ºF
Steam
P-1
To desalter and hydrotreaters
Figure 2 Conventional sour water stripper design used in older units Ammonia gas to sulphur plant
13 PSIG 190ºF
Sour water feed
90ºF (15)
(1)
E-1
Steam
To desalter and hydrotreaters
250ºF
P-1
Figure 3 Correct sour water stripper design without feed preheat or external heat extraction
PTQ Q2 2013 3
Ammonia gas to sulphur plant
13 PSIG 190ºF
Sour water feed
90ºF (15)
(8)
Medium content NH3 water to hydrotreaters
(1)
E-1
Steam
Low NH3 water to desalters
250ºF
P-1
Figure 4 Two-stage sour water stripper design without feed preheat
strippers in India, discussed earlier, the NH 3 in the bottoms product was about 200 ppm, even though the NH 3 in the feed was only 1000–2000 ppm. Renery personnel felt, and correctly so, that the 200 ppm of NH3 precluded the use of these sour water stripper bottoms as crude unit desalter wash water. Thus, the water owed directly into the renery efuent treatment plant. The problem was not a lack of stripping steam or the excessive pH of the stripper bottoms, as renery operators suspected. It is quite normal for the stripper bottoms to have a higher pH than the feed if stripping efciency is bad. That is, the acidic H 2S is stripped out of the sour water more easily than the basic
4 PTQ Q2 2013
ammonia. The problem of excessive NH 3 in the stripped water efuent was low tray efciency. Field observations To troubleshoot the stripping deciency, I made the following eld measurements: • The delta P across the bottom 32 trays was zero, meaning no intact trays • The delta P across the top eight trays was 7 psig (or about 16 ft of water, which equalled the vertical height of the tower above the chimney tray shown in Figure 1), meaning trays 33-40 are totally ooded. I then advised as follows: • Do not use valve or oat trays. The valve caps stick to the tray deck and promote high delta P and ooding. Use
a grid tray with a half-inch cap lift. Pro-valves are a good option that I have successfully used in fouling services • Bolt the trays to the tray ring or use back-to-back trays with shear clips3 to avoid tray failure due to pressure surges caused by high bottom liquid levels. One acceptable tray design is sieve trays with half-inch holes, which seem to work ne. Avoid packed towers. They are subject to vapour-liquid channelling and poor fractionation efciency due to sloppy instal lation, fouling or poor liquid feed distribution. When calculating the required hole area for the trays, do not forget that the vapour loads and hence the required tray hole area substantially diminish as vapour ows up the column from the reboiler to the feed tray. A recent publication4 discussed the various methods available to calculate tray efciency in sour water strippers: • Murphree vapour efciency • Overall efciency • Mass transfer efciency. This article suggests that the reboiler steam rate may be reduced by roughly 20% if the number of stripping trays is increased from 18 to 32, based on a computer simulation using mass transfer tray efciency. However, in reality, the tray efciency for perforated tray decks (valves, sieves or grid trays) is mostly a function of: • Tray deck levelness • Weir levelness. Decks that are out of level promote vapour channelling through the tray deck. Weirs that are out of level promote
www.digitalrefining.com/article/1000799
liquid channelling across the the crude unit desalter had will increase the stripper tray deck. Either type of chan- been reduced enough so that bottoms ow by 10-12%. But nelling causes loss of brine carryover was no longer this will be offset by the diminvapour-liquid contacting and a problem. ished fresh water make-up rate thus a severe reduction in tray Caustic addition to the sour to the desalter or hydrotreater efciency. Experience has water stripper feed will efuent wash. shown that the 15 stripping increase NH3 stripping efAbout 50% of reneries use trays shown in Figure 3 are ciency by reducing the open stripping steam and have adequate if sufcient steam is solubility of NH3 in water. This discarded their reboilers. Or used to heat the bottoms efu - may be good practice if the they are operating with ent to the boiling point of stripper bottoms is owing into massively leaking reboiler water at the stripper bottoms the renery efuent treatment tubes, which is much the same pressure, provided that the basin. But, it is better to have as open stripping steam. trays are properly installed and 10-20 ppm NH3 in the sour Referring to Figure 3, for designed correctly to prevent water stripper bottoms than to many renery operations, the dumping and channelling.1,3 contaminate the desalter with circulating thermosyphon NaOH, which promotes the reboiler shown can be elimiformation of brine-crude oil nated and replaced with the Use of caustic to improve emulsions. same quantity of stripping stripping I was working on a smaller steam, thus further simplifying crude unit in Alabama during Reboiler corrosion the design of the sour water 2012 to improve desalter ef- One of the questions that I am stripper installation. ciency. Their problem was the often asked is what can be periodic carryover of emulsi - done to mitigate reboiler tube Conclusions ed brine. The depth of the failures in sour water strippers Generally accepted sour water emulsion between the crude and the consequent steam leaks stripper designs in the rening and the water phases in their into the stripped sour water? industry are not necessarily desalter was excessive. The pH One suggestion is to avoid correct. The application of basic of the wash water from the the use of kettle reboilers or chemical engineering principles sour water stripper was 8.5 to once-through thermosyphon such as the stripping factor can 9, even though the NH 3 content reboilers in favour of circulat- often lead to simpler, more of the stripper bottoms was ing thermosyphon reboilers. cost-effective designs. Avoid zero. Discussions with the That is, feed the reboiler the use of utter cap-type operators revealed that they directly from the bottom of the valve trays in fouling services. added NaOH to the sour water stripper and design for about a The caps will stick to the tray stripper feed to diminish the 10-20% vaporisation rate (see deck and cause ooding. The NH3 content of the stripped Figure 3). This will reduce, but high pH of stripped sour water water. unfortunately not eliminate, is a symptom of poor stripping Caustic contamination of the reboiler tube failures. efciency, not the cause. desalter wash water is known There is a more helpful Preheating sour water stripper to promote high emulsion suggestion if one can meet the feed normally serves no levels. There are many refer - following two criteria: purpose. Avoid the use of ences to this problem in NPRA • Stripped water ows to the caustic if the stripped water is Q&A sessions. With this source crude unit desalter and/or the to be reused as crude unit as support, I asked the opera- hydrotreater efuent wash desalter wash water. A renery tor to stop the NaOH addition water make-up sour water stripper should to his sour water stripper feed. • External water is required to have about 20 trays, not 40. It Two hours later, at the same supplement stripped sour is best to use grid or sieve trays reboiler duty, the NH 3 content water at the above units. rather than a random type of the stripped sour water had The suggestion is to use open (rings) in the tower. If high gone from 0 ppm to about 5 steam in the stripper bottoms turndown rates are required, ppm and the emulsion layer in and abandon the reboiler. This use bubble caps rather than
www.digitalrefining.com/article/1000799
PTQ Q2 2013 5
valve tray decks. For many locations, the bottoms reboiler may be replaced with open stripping steam. References 1 Lieberman N, Troubleshooting Process Operations, PennWell, 4th ed. 2 Lieberman N, Troubleshooting Natural Gas Processing, PennWell. 3 Lieberman N and E, A Working Guide to Process Equipment , McGraw Hill, 3rd ed. 4 Hatcher N, Weiland R, Reliable design
6 PTQ Q2 2013
of sour water strippers, PTQ, Q3 2012, 83-91.
Norman Lieberman is a Chemical Engineer who specialises in troubleshooting refinery noncatalytic process equipment and in retrofit process design. He has taught his Troubleshooting Seminar since 1983 to over 18 000 engineers and operators. The first of his eight books, Troubleshooting Process Operations, has
been continuously in print for 32 years. He graduated from Cooper Union in New York City in 1964. Email:
[email protected]
LINKS More articles from the following categories: Mass Transfer & Separation Water/Wastewater Management
www.digitalrefining.com/article/1000799
