VDU condensate back-up? 4

[MaNaTMoS]

Are there any possibilites to drain the condensate to the barometric drum and restore the vacuum in "dry" distillation column? (no precondenser, 3-stage ejector system, delta T for 1st condenser's gas/condensate outlet is +14C. Motive steam pressure is OK, temperature is 30C above design). Changing the level in barometric drum is not possible, pressure in the drum is 1,01 ata. What parameters should I check more? Thank you.

 

[25362]

Are the barometric pipes submerged or otherwise their ends sealed with water all the time ?

 

[MaNaTMoS]

Yes, the seal legs are asolutely submerged in water all the time. Design of barometric drum is such one it doesn't allow loss of water seal, even if I drain everything from the drum.
Now the 1st stage of ejectors is out of process, and vacuum is some 10-15mm Hg higher then in previous case. I'm almost 100% sure that we're suffering from condensate backup in the 1st stage aftercondenser. (This column is a vacuum bottoms stripper, the quantity of stripping steam used is 50% less then design. The first column has the same 3-stage ejector system, which has great performance all the time.)
Could it be possible that seal legs design is bad? It's a 4" line, not straight one (we have 4 pipe "knees" x 45 degrees between the condenser and the barometric drum). I'm asking this because everytime we start the VDU, the vacuum is very good in both columns for first several weeks, and then it's going worse by time.
What should I try further more?
Thank you.

 

[rmw]

Check the seal legs right at the surface line of the water leg that is pulled up into the seal leg for pin holes. Often, the wetting/drying of this area, as vacuum varies with temperature, etc, will cause corrosion in this area, and pin holes will develop, that will load up the downstream jets, robbing the primary jet of performance.

Also, if all seal legs are into the same hotwell, and there is no baffle or weir separating the last stage jet, if it is a direct exhaust into the hotwell, not through an aftercondenser, air bubbles from the agitated water in the hotwell can find their way into the seal legs of barometric drums of other stages, adding air load to those stages, and impeding the downward flow of the drains from the barometrics.

rmw

 

[25362]

From what I could understand, the vacuum malfunction appears on both towers ("main dry" and stripper) after weeks of normal operation (?)

There are several possible other reasons, mostly connected with the condenser's performance:

1. Jet surging
2. Increased gas load due to thermal cracking in the
heater.
3. Condensate back up, as by restricted condensate flow.
The condensate would become "subcooled".
4. Progressive fouling of the condenser.
5. Insufficient cooling water.
6. Higher cooling water temperatures.
7. Air release out of solution in the cooling water.

The absolute pressure would rise if condenser performance drops as listed above, or by putting one set of jets off commission.

Upon a pressure increase at suction a jet may start surging because the vapor velocity in the diffuser throat drops below sonic and its pressure ratio suffers.

As a result, the increased vacuum column pressure produces fewer overhead vapors which again improves the condenser performance, bringing down the pressure in the diffuser throat restoring above-sonic velocities, and normal operation.

This cycle would be repeated until the condenser is restored to normal operation, or the gas load from the VDUs is reduced.

This can be done on dropping the heater's oulet temperature by, say, 4-5^oC, a fact that would lessen thermal cracking. The resulting improved vacuum could compensate for the the lower temperatures maximizing the vacuum gas oil offtake.

One must check one by one of the above possible causes to be sure. Good luck.

 

[MaNaTMoS]

Thank You for your suggestions.

Reply to RMW:
All seal legs are in the same hotwell, but the vapour line after 3rd stage condenser goes to the seal drum before entering barometric drum. Off-gas from seal drum is directed to the process heater, condensate is draining to the barometric drum.
Regarding Your first suggestion, I think the only way we could check the seal legs corrosion is to shut off the unit and open the drum. That should happen in the near future.

Reply to 25362:
The vacuum has improved after shutting-off the 1st stage ejectors (2 parallel ejectors), which (in normal process conditions) has no sense at all. The only thing that crosses my mind is lack of condenser duty, due to tubes being submerged in condensate (condensate draining temperature is 28C, vapour outlet from condenser has temperature 42C - it should be 25-28C).
The vacuum in the first column is very good (745mm Hg); this column takes the biggest part of the vapour load with no problems at all. I have checked the temperature profiles accross the 1st column condensers (all 3 stages), and these profiles are OK - condensate is always some 1-2C hotter than vapour outlet from every condenser. Bottoms from this column are directed to column below (vacuum stripper), with use of stripping steam 50% less then design, as i mentioned before.
Are there any operational ways to improve the vacuum, especially in the second - vacuum stripper column?

Thank you again.

 

[25362]

MaNaTMoS, your analysis of having the condensate removed at a much lower temperature than the uncondensed gas is apparently correct. What is the normal temperature profile of the condenser, and of the cooling water, when the vacuum is OK ?

The strangeness of the troublesome condition is that it appears only after some weeks of normal operation. Thus one should look after those changes that may be taking place as time passes.

Is there a possibility of salt forming, precipitating, adhering and partly plugging the barometric legs' non-vertical sections ?

Is the condensate of this "lower" stripping tower top condenser removed with a pump or by gravity ? Pumps may develop inward air seal leaks with time. Leaks that may induce condensate back up and the covering of condenser tubes with an increase in the load to ejectors. If this is the case a temporary solution could be spraying water on the seals. This may stop air leaks by substituting water for air.

Sometimes, if H2S is present, an injection of NH3 may improve the vacuum by the formation of ammonium hydrogen sulfide which is soluble in water. Thus, the load on the ejector is reduced and vacuum can be restored.

Anyway, your problem may be solved by careful observation of the factors that develop over some time lapse of normal operation. When identified the culprit can be removed. Good luck.

 

[MaNaTMoS]

All condensate lines from both columns condenser systems are gravity-drained. The seal legs from the 1st column condensers have only 2 x 45 degrees elbows, just above the barometric drum. The 2nd column condenser system has 4 elbows.
Delta T for cooling water in the first condenser (vacuum-stripper, 2nd column) is only 21 - 18 = 3 C, which is very good indication of submerged tubes inside the condenser. Delta T for cooling water in the 1st column 1st condenser is 27 - 18 = 9 C.
Last year we had a problem with plugging of flame arrestor grid, on the vapour line from seal drum to the process heater. The effect was creating the back pressure to ejector system and great loss of vacuum in both columns. I've checked delta P accross the grid, and it's OK.
About your suggestion of NH3 injection... What is the best location for it? And what range of NH3 concentrations would fit? I have an idea to send a sample of barometric drum water (for H2S content) and ejector off-gas to analysis (to determine if there is air leakage into the system).

Best Regards,
MaNaTMoS

 

[25362]

Since at higher absolute pressures the HTC on the shell side should improve, the fact that you lose condensing efficiency points to waterlogging.

The question that bothers me is why all this happens after weeks of normal operation. Do you have a hint ?

Is the stripper operating on not-quenched vacuum bottoms ?
H₂S is not very soluble in water. I think a preliminary qualitative smell test would reveal its presence. I'd be grateful to learn of the results of your quantitative tests.

 

[rmw]

MaNa...,

I went back and re-read once again your original post. It doesn't hurt that I also attended a vacuum seminar in the interim, as well, and was reminded of a fact that I knew before, but failed to mention.

You state that your steam motive pressure is OK, but that the temperature is 30C above design. That indicates superheat in your steam, which changes the specific volume of the steam, making it larger, and reducing the amount of motive that the nozzle in the jet can pass, which reduces the capacity of the jet. High superheat is just like lowering the pressure, and moves the shock wave towards the nozzle end of the jet, into the inlet side of the difusser, and makes the jet break, which then enables backsteaming of both the motive, and process gas from the intercooler.

Second thing that I would mention in concert with the discussion of what has already gone on in this thread, based on a real problem at a refinery attending this seminar, is, what is the specific gravity of the liquid in the tail leg, and what is the height. Has the sp. gr. changed from the original design. Was enough height designed in for the sp. gr. of the condensate. Also, if the first jet is breaking, then the first intercondenser is seeing column vacuum, and the tail leg may have only been designed for the discharge pressure of the first jet, not the column vacuum, pulling condensate into the intercondenser which cannot drain fast enough when the jet finally remakes, and before it breaks again. Do you hear the characteristic "woosh, woosh" sound at the first jet??

Things to check.

rmw

 

[MaNaTMoS]

Thanks to everyone who took place in this discussion during last several weeks.
Vacuum Distillation Unit will be turned down, as a part of regular process equipmet inspection and revamping program.
If you have any useful suggestions or tips, please don't hesitate to notify me.

Wish you happy Easter holidays.


Regards,
Zauberberg

 

[MaNaTMoS]

After short turnaround of ADU and VDU, we have found deposits in the ejector system (draining out from the barometric drum), after flushing the system with full pressure steam procedure. Deposits contained oily fraction, water, and something which looked like pieces of thermal insulation or even more - like fine moss, equally distributed over and inside the black oily fraction. After thermal degradation of sample at 500°C without air, what was left lonely was like numberless small hairs with black dust and shiny, fine particles. The condensers were apparently clean before the start of the cleaning procedure. I’m waiting for the sample metal content laboratory results.
I read some time ago about microbial growth and corrosion in VDU condensibles draining system, along with condensate subcooling and back-up problems due to different causes of flow restrictions, and I wonder about that posibility in our vacuum distillation unit. The microbial growth inside non-vertical parts (four 45 degrees elbows) of condensate draining system (4’’ and 2’’ seal legs) and barometric drum could be the cause of progressive flow restriction and condensate back up. That could be an explanation of gradual loss of vacuum by time, as the growth thickness becomes higher. Also, I think I should count for pour point of trim gas oil which is draining down with subcooled water, and environment temperature influence on oil quantities being solidified. With condensate becoming subcooled, there's much higher possibility for drained oil fraction to stop flowing (higher viscosity, reaching the pour point?) and plug the drainage system. It’s very symptomatic that only 2nd column (vacuum bottoms stripper, with ejector system on the lower platform level - 13m from the ground) suffers from loss of vacuum by time. Now I’m thinking what could be the way to proove or reject these conclusions.
I also think about pieces of thermal insulation (glasswool) entrained somewhere from VDU system, totally neglecting ejector nozzles or/and condensers plugging possibility. That doesn't look like very possible thing to happen during VDU operation. Before the start of VDU, we have done the full vacuum test. It was OK.
I would appreciate any useful comment about this problem.

 

[25362]

A two-step vacuum distillation is not a frequent procedure. Can you tell us why has these system been adopted ? Is this to improve gas oil recovery ? There are other more conventional ways to do that. Please comment.

 

[MaNaTMoS]

Dear 25362,
Your question is the same as if You asked me "Why Hungary doesn't have a seaside, but has military navy forces?"
That reminds me of my life philosophy, so i liked Your question very much.
Regarding the VDU configuration, I don't know if there was anything different from current installation in my refinery before. I'm 28 and this is the only VDU I have experience with. Anyway, thank You for Your suggestion about HVGO.

I'm very interested to find out what operational moves can I make to improve vacuum in such conditions, as they were found in VDU before the turnaround. Maybe heating the condensate lines with steam hoses from outside, wraped around the pipes could help? Or treating the system with biocides occasionally, if we find out that microbial growth is the main problem? Changing the pH, for some bacteria species, also presents a part of sollution.
If You have any ideas or experiences with such VDU symptoms and problems in practice areas, I would be very appreciate to hear what they look like.

Regards,
Z.

 

[MaNaTMoS]

Hello to everyone again!

Yesterday I have received laboratory results for ejector system deposits metal content. The results are presented further in this text.

The "loss of weight" after thermal degradation of sample at 550°C (without presence of air) was 84.55%. The metal content of the residue is as follows:

Fe - 23.1%
Al - 3.83%
Si - 12.5%
Ca - 4.55%
Mg - 1.62%
Zn - 0.16%
Mn - 0.16%
Cr - 0.13%

After seeing this report, I decided to check thermal insulation ("mineral wool" which is used for equipment insulation) metal/metal oxides content. The following data are from product specification sheet:

SiO2 - 43%
Al2O3 - 15%
Fe2O3 - 7%
CaO - 15%
MgO - 8%
Na2O - 2.5%
K2O - 0.5%
TiO2 - 1%

The most confusing thing is that some of the bacteria species which could be present in such systems (vacuum condensers, seal legs, barometric drum and off-gas seal drum) also produce metabolic fibers with high content of Si, Ca, Mg.
Are there any alternate ways to determine the origin of ejector system deposit, and percentage of corrosion products in it? How can I prevent repeating of this kind of problem in the future? Have anyone had similar experiences with VDU before?

Thank you in advance.

Regards,
Z.

 

[MaNaTMoS]

Additional observations

Relative density of deposits
After an ordinary field procedure (throwing the piece of untreated deposit sample in a glass of water), we have noticed the following: at first, sample sank to the bottom of glass; along with time, it was slowly breaking into smaller pieces which found they way to water surface, floating.

I've exposed a piece of insulation to same thermal procedure (550C, no air) as it was for deposits sample, and the results are - no changes at all, even the colour remains practically the same. Concerning the composition and purpose of mineral wool, it could be expected.

 

[INBCPE]

Steam jet systems are funny. The problem you describe is remarkably similar to a problem I had with a 3-stage about 5 years ago in a soup plant. The system was used to pull vacuum on the soup kettles to help cool them down.

When I showed up to troubleshoot it, it had the classic signs of fouling. The X stage was breaking, the water temp rises where pretty small, the heads on the other stages were very hot, interstage pressures were high. I also got better vacuum when the X stage was off. I talked to one of the operators, and knew within two minutes of that conversation what was going on. The operator said one of the batches burped into the jet system. There was chili in the x stage, chili in the z stage, chili on the roof, chili in the boiler. They had tried valiantly to clean the system, but once we opened the unit up, we saw a uniform coating of chili sauce on the tubes on both condensers. The condensers were fouled. There was even product fouling some of the condenser drains.

The maintenance department contracted with an outside pressure washing company and did a swell job on cleaning the unit. The next day the unit started up and ran like it was brand new.

I used to be traveling steam jet repair man. I don't pretend to know what exactly is causing your problem (but my experience is telling me it's a gradual fouling problem), but I can offer you advice that proved itself time and time again for me around the world: be thorough, be methodical, never assume that something is OK (steam quality better than 98%, nozzles are clean, tube bundles aren't fouled, no coolant leaks from water side tubes to vapor side, SG of coolant hasn't changed, no leaks coming from condensate pumping system, no hardness in the water, etc, etc). You need to measure and check everything. Don't guess or assume. You will waste time. The vulcan mind meld does not work with inanimate objects like steam jets. If you can't open the system up to inspect it, the best you can do is modify conditions (increase/decrease flows and pressures) and collect data. The manufacturer should have a good data sheet to let you know how the system should perform.

As far as the material you found, it could be anthing. Don't limit the possibilities to one side of the system or the other. I once found a brick inside a large barometric condenser, and I never figured out how it got there. Try not to get hung up on it, but if you can find out where it came from, great. Also, the constituents of the material you found look like more than one thing. I couldn't venture a guess as to what it is, though.

One other thing. I had tail legs flood because it had some really high molecuar weight hydrocarbons in them. The stuff wouldn't drain out because they had a low SG, but it wouldn't boil off either. We just kept getting a thicker and thicker layer of this stuff in the tail leg until it backed up into the condenser and shot the system to hell. And when the tail legs drained, the stink was unbearable.

 

[INBCPE]

Oh, I also had one job where the water bonnets on the heat exchangers were fouled with zebra mussles, and it produced the same effect (breaking X stage and low water delta T).

 

[MaNaTMoS]

Thank You, INBCPE, your comments deserve at least 3 stars.
The unbearable smell you mentioned in your last troubleshooting story was exactly the same thing in our case. In combination with fine hairs being uniformly dispersed and large quantities of Si, Ca, Al in the ejector deposits, that was the main reason for asumption of microbial growth inside condensate system is to be made.
Concerning the high SG of hydrocarbons (and "burped soup" case), I suspected this was happening because of the following: every time when we would increase the HVGO pumparound rate (or lower its return temperature in order to maintain design 80C design column top temperature, by trimming the bypass flow through TWS heat exchangers - the vacuum would start to fall (our current top temperature is 115C, and vacuum goes up to 740mmHg). In such case, as You previously well described, the possibility of more subcooled high SG liquid entrainment to the ejector system looked like another cause of poor performance, rather than higher crack-gas load in this bottoms' stripper column.

Regards,
MaNaTMoS