Forums
Blogs
Articles
File Library
FAQ
Search
FB
Hello all. I am a senior refinery operator employed with the national oil company of Trinidad & Tobago. I work on a vacuum distillation unit where recently a new primary ejector was installed. Before this the unit was operated without any primary ejectors. Typical top pressure was in the range 40-55mmHg with a net bottoms yield in the range 35-50% depending on the charge stock ( we routinely charge as much as 10 different crudes to the refinery ). With the new ejector commissioned the top pressure achievable can be as low as 15mmHg and the bottoms yield has been reduced ( although this has not yet been quantified ). The operating instruction is to hold a top pressure of 23-27mmHg, a wash oil pump around flow of 10-12000 b/d and a wash oil down pipe temperature of 716F. There is a steam line connected via an auto valve to the column's vapor line which could be used to control the top pressure. However the present control philosophy is to control the top pressure by using the column's off gas auto to hold a back pressure on the column. As a result of this (I suspect) the column pressure control is very unsteady with pressure swinging between 15-50mmHg. Are there any better pressure control options? Thanks very much for any information.
|
|
Vacuum Tower Pressure Controls
Started by Guest_srojack_*, May 21 2008 11:11 AM
7 replies to this topic
Share this topic:
#2
Zauberberg
-
- ChE Plus Subscriber
-
- 2,728 posts
Gold Member
Posted 21 May 2008 - 12:08 PM
Spillback pressure control in vacuum overhead systems is a common design practice for maintaining stable operation. However, based on my operational experience, this arrangement actually works in a very small percentage of cases, and possible explanations are: uneven flow of noncondensables from vacuum heater (which makes controlling the spillback flow very difficult), inadequately designed pressure control loop - including improperly sized control valves, bad performance of vacuum condensers (unsteady supply/temperature of cooling water), poor arrangement of parallel ejector systems etc.
I believe your equipment is designed to withstand full vacuum (both tower and overhead system). In such case, switching PCV to manual (or complete closing and isolation with gate valves) can give you a clue what is going on: whether the problem originates from vacuum heater excessive cracking or bad heat balance control in main VDU column, or the problem is in pressure control loop itself. Sometimes, it is possible to flood LVGO bed with excessive pumparound and cause liquid carryover into the overhead system which adversely affects pressure control. Also, I have seen excessive back-pressure taking place in VDU because of plugged flame arrestors in the off-gas line to the vacuum heater. Integrated analysis and field troubleshooting is something which should be conducted - with process engineers and shift personnel participating together.
You didn't tell us what is the type of your VDU - is it dry or wet unit (with or without precondenser), with coil steam injection or not, and how do you control these parameters. How many ejectors operating in parallel? Based on your input data and pressure profile, I'm guessing this is a 3-stage ejector system, but some more data from your side would definitely make things more clear.
Best of luck,
I believe your equipment is designed to withstand full vacuum (both tower and overhead system). In such case, switching PCV to manual (or complete closing and isolation with gate valves) can give you a clue what is going on: whether the problem originates from vacuum heater excessive cracking or bad heat balance control in main VDU column, or the problem is in pressure control loop itself. Sometimes, it is possible to flood LVGO bed with excessive pumparound and cause liquid carryover into the overhead system which adversely affects pressure control. Also, I have seen excessive back-pressure taking place in VDU because of plugged flame arrestors in the off-gas line to the vacuum heater. Integrated analysis and field troubleshooting is something which should be conducted - with process engineers and shift personnel participating together.
You didn't tell us what is the type of your VDU - is it dry or wet unit (with or without precondenser), with coil steam injection or not, and how do you control these parameters. How many ejectors operating in parallel? Based on your input data and pressure profile, I'm guessing this is a 3-stage ejector system, but some more data from your side would definitely make things more clear.
Best of luck,
#3
Zauberberg
-
- ChE Plus Subscriber
-
- 2,728 posts
Gold Member
Posted 22 May 2008 - 09:58 AM
Also look at Sloley's famous article, figures 3, 4, 5 and 6: there is an explanation of each pressure control loop with recycling gas to ejector suction, including advantages and disadvantages, applicability ranges etc.
http://www.clarkson....gn/distnprs.pdf
This is very interesting topic and the best thing would be if you can provide additional information (type of feed, configuration, coil/stripping steam details...), including system sketch as well. Ejector systems are beauties, "the only process equipment with a personality", as one of ChE legends used to say.
http://www.clarkson....gn/distnprs.pdf
This is very interesting topic and the best thing would be if you can provide additional information (type of feed, configuration, coil/stripping steam details...), including system sketch as well. Ejector systems are beauties, "the only process equipment with a personality", as one of ChE legends used to say.
#4
Guest_srojack_*
Guest_srojack_*
-
- guestGuests
Posted 24 May 2008 - 12:38 AM
Hello again. Thank you for this info. It gives some insight that was previously missing. You are right in that the equipment is designed to withstand full vacuum. The process engineers are however concerned about operating at "too low a vacuum". They say that too low a vacuum may have the effect of increasing the vapor velocity up the column so much that the wash circuit will have no effect, thus causing a poor quality HVGO (high carbon residue, metals, poor color and such through entrainment). Please comment on this. With respect to further info on unit design, the column is a wet one with both coil steam injection and stripping steam. The coil injection is at the convection inlet to the vacuum heater (injecting into eight parallel inlet passes). The source of steam is superheated 65psi. for both coil injection and stripping. The temperature of this steam ranges from 720-800F dependent on the flow rate through the coil and the duty on the vacuum heater. The coil injection rate is currently held at 6000 pounds per hour regardless of the heater charge rate and the stripping rate varies from 9000-11000 pounds per hour depending on the charge rate to the column (minimum charge to the column is set at 40000b/d). The overhead system is in three stages. There is one primary ejector with a rated capacity (i think) of circa 26000pph. exhausting into two primary condensers (in parallel). The condensates from these drain into the hotwell. There are three second stage ejectors pulling in parallel from these condensers and exhausting into one second stage condenser separated into two halves. Three third stage condensers pull from this condenser and exhausts into one condenser. Liquids drain into the hotwell and the uncondensibles arerouted via a knockout drum to the heater to be burnt (inline flame arrestors) and the excess to flare via a PCV (this back pressure is also used to "control the column pressure"). The motive steam for all ejectors is 175psi., the condensers are surface condensers and the cooling water is salt water supplied from off site at 20-30psig.. The unit is designed for a throughput of 90000b/d and consists of a Gas Oil Tower operation upstream the vacuum tower. The net distillate from the GOT amounts to circa 20000b/d. The GOT heaters (2) operate at a transfer of 745F while the vacuum heater has a transfer of 750F. The only ironclad controls on the GOT are an overhead temperature not greater than 500F and a top reflux temperature of not less than 300F. The column pressure is operated as low as possible (normally 9-15psig.)so as to promote a heavy gas oil draw off.
#5
Zauberberg
-
- ChE Plus Subscriber
-
- 2,728 posts
Gold Member
Posted 24 May 2008 - 11:01 AM
Heater COT can be considered low for vacuum service, and for most types of feed the cracking temperature is well above this value. I assume you have good heater design, with balanced heat flux in all 8 passes, so thermal degradation of feed really should not be an issue in your case.
Excessive vacuum can cause problems with HVGO quality - in some instances - due to high superficial velocity of upflowing vapors, but I believe this should not be the problem in your tower since it is designed for 23mm Hg, and as far as I understood you also operate some times with 15mm Hg overhead pressure without any problems related to distillate quality. Do you recycle slop wax to heater inlet, or it is lost together with short residue, by overflowing through the downpipe? For achieving high HVGO yields, recycling slop wax back to the heater is a must, since this fraction contains at least 35% of HVGO; but this is a design issue.
I wouldn't be so much concerned when switching overhead PC to manual mode and keep it open 10, 20 or 50%, depending on normal operating range for a given throughput and operating regime of VDU. This is the best way - and you certainly know this better than me - if plant can be run steadily with control loops operating in manual mode, and upsets start to happen when control loops are switched to automatic mode, then the problem is most probably in inadequate control system design. You can try it, by keeping PCV open as much as you consider to be safe, and watch for the results.
Do you experience ejector surging? This happens when you exceed - for any reason - maximum discharge pressure of an ejector, followed by rapid loss of vacuum. It happens usually in cycles, and the most frequent reasons are excessive suction load or insufficient condensing capacity of downstream vacuum condensers. You can track this on your DCS history module and look for rapid losses of vacuum. Condensate back-up in vacuum condensers (due to problems in gravity flow of condensed oil and steam to the hotwell - especially if it is high pour-point oil fraction) can also cause cycling in tower operation. Check this by touching vacuum condensers on their external surface, and if there is a significant temperature gradient between condensate outlet and vapor outlet (2nd stage suction), then you are suffering from condensate back-up. Properly designed self-regulating heat tracing system can solve this problem.
Wet VDU towers are somewhat specific, in the sense that achieving highest vacuum does not correspond with maintaining tower top temperature as low as possible, which is the case in dry towers. Is your tower top temperature stable? And pressure drop across the trays/packings/sections, how does it behave during these upsets?
Excessive vacuum can cause problems with HVGO quality - in some instances - due to high superficial velocity of upflowing vapors, but I believe this should not be the problem in your tower since it is designed for 23mm Hg, and as far as I understood you also operate some times with 15mm Hg overhead pressure without any problems related to distillate quality. Do you recycle slop wax to heater inlet, or it is lost together with short residue, by overflowing through the downpipe? For achieving high HVGO yields, recycling slop wax back to the heater is a must, since this fraction contains at least 35% of HVGO; but this is a design issue.
I wouldn't be so much concerned when switching overhead PC to manual mode and keep it open 10, 20 or 50%, depending on normal operating range for a given throughput and operating regime of VDU. This is the best way - and you certainly know this better than me - if plant can be run steadily with control loops operating in manual mode, and upsets start to happen when control loops are switched to automatic mode, then the problem is most probably in inadequate control system design. You can try it, by keeping PCV open as much as you consider to be safe, and watch for the results.
Do you experience ejector surging? This happens when you exceed - for any reason - maximum discharge pressure of an ejector, followed by rapid loss of vacuum. It happens usually in cycles, and the most frequent reasons are excessive suction load or insufficient condensing capacity of downstream vacuum condensers. You can track this on your DCS history module and look for rapid losses of vacuum. Condensate back-up in vacuum condensers (due to problems in gravity flow of condensed oil and steam to the hotwell - especially if it is high pour-point oil fraction) can also cause cycling in tower operation. Check this by touching vacuum condensers on their external surface, and if there is a significant temperature gradient between condensate outlet and vapor outlet (2nd stage suction), then you are suffering from condensate back-up. Properly designed self-regulating heat tracing system can solve this problem.
Wet VDU towers are somewhat specific, in the sense that achieving highest vacuum does not correspond with maintaining tower top temperature as low as possible, which is the case in dry towers. Is your tower top temperature stable? And pressure drop across the trays/packings/sections, how does it behave during these upsets?
#6
Zauberberg
-
- ChE Plus Subscriber
-
- 2,728 posts
Gold Member
Posted 29 May 2008 - 10:36 AM
Hi Srojack,
No feedback from you in the past few days. How are the things going in your VDU?
Best regards,
No feedback from you in the past few days. How are the things going in your VDU?
Best regards,
#7
Guest_srojack_*
Guest_srojack_*
-
- guestGuests
Posted 29 May 2008 - 08:34 PM
Hey Zaubs,
Nice to hear from you again. As usual your info was very pertinent and coincided with my views. Thanks for the endorsement.
From my own observations it would seem that the additional vacuum capacity (provided by the addition of this new primary ejector) is more than enough to achieve 25 mmHg top pressure. The process engineering personnel tried to offset this by increasing stripping steam into the column and by increasing the back pressure on the column by throttling the of gas auto (in auto mode and increasing the set point). The stripping steam proved to be unnecessary and the back pressure control resulted in the column pressure surging.
The solution appeared to be very simple (I am still trying to convince the process engineer!). First I put the vacuum off gas auto on manual and opened it up to achieve a pressure just above atmosphere (15-16 psia). This ensured that there was no restriction in off gas flow out of the column, but also dropped the top pressure too low (<25mmHg). To offset this, one second stage ejector was taken off fully and the vapor inlet to another second stage was throttled (and sometimes even shut off) to maintain the column pressure at 25 mmHg. In other words the vacuum capacity was manipulated depending on the vapor load. The column operation has steadied as a result but the process engineer is still hesitant to adopt this method (maybe too simple to work??) even though the evidence is there to show. Can you comment?
Interestingly enough, the column design prior to the primary ejector installation recycled the slop wax to the bottom of the upstream Gas Oil Tower. Despite many reservations and outright objections by myself, this recycle circuit was engineered out and the slop wax now overflows via downpipe to the column bottom. How can HVGO optimization take place under these conditions without compromising quality? The vacuum gas oil is charged to a hydrotreater operating in mild hydrocracking mode (currently 735 F WABT). Anything to consider as a result? I have not had the chance to look at the DCS history to investigate ejector surging. I'll comment as soon as I do. Hope to hear from you soon.
Have a good day
Nice to hear from you again. As usual your info was very pertinent and coincided with my views. Thanks for the endorsement.
From my own observations it would seem that the additional vacuum capacity (provided by the addition of this new primary ejector) is more than enough to achieve 25 mmHg top pressure. The process engineering personnel tried to offset this by increasing stripping steam into the column and by increasing the back pressure on the column by throttling the of gas auto (in auto mode and increasing the set point). The stripping steam proved to be unnecessary and the back pressure control resulted in the column pressure surging.
The solution appeared to be very simple (I am still trying to convince the process engineer!). First I put the vacuum off gas auto on manual and opened it up to achieve a pressure just above atmosphere (15-16 psia). This ensured that there was no restriction in off gas flow out of the column, but also dropped the top pressure too low (<25mmHg). To offset this, one second stage ejector was taken off fully and the vapor inlet to another second stage was throttled (and sometimes even shut off) to maintain the column pressure at 25 mmHg. In other words the vacuum capacity was manipulated depending on the vapor load. The column operation has steadied as a result but the process engineer is still hesitant to adopt this method (maybe too simple to work??) even though the evidence is there to show. Can you comment?
Interestingly enough, the column design prior to the primary ejector installation recycled the slop wax to the bottom of the upstream Gas Oil Tower. Despite many reservations and outright objections by myself, this recycle circuit was engineered out and the slop wax now overflows via downpipe to the column bottom. How can HVGO optimization take place under these conditions without compromising quality? The vacuum gas oil is charged to a hydrotreater operating in mild hydrocracking mode (currently 735 F WABT). Anything to consider as a result? I have not had the chance to look at the DCS history to investigate ejector surging. I'll comment as soon as I do. Hope to hear from you soon.
Have a good day
#8
Zauberberg
-
- ChE Plus Subscriber
-
- 2,728 posts
Gold Member
Posted 31 May 2008 - 10:45 AM
Srojack,
Using stripping steam for column back-pressure control, as you also have noticed, cannot be considered as a serious and proper way to do it. In addition, if stripping steam incremental flow does not yield in higher HVGO recovery then it is, in simple words, wasting of money. If you can convince your process engineers that they are much smarter than you, they will accept any solution you propose. This is old trick. Engineers usually don't like clever senior operators, in spite of the fact they can learn so much of you. I am aware of this truth, since I spent almost two years working as ADU/VDU process engineer and plant manager, amongst other assignments I had in petroleum refinery. These days were, perhaps, the best I had in my career.
I would definitely close and isolate spillback PCV, and leave off-gas flow unrestricted. This is how your overhead system gains the maximum capacity. The only concern I would have in such circumstances, is whether I can maintain sufficient wash oil flow without flooding the wash oil bed, and keep tower heat balance under control due to increased distillate flow from the flash zone. If you have sufficient capacity in pumparound circuits, then there are no mistakes. Or at least you can conduct plant performance test, with different crudes you process, and see what is the optimum operating envelope of the tower at higher vacuums.
HVGO recovery: without design changes, there is no extra space for improvements. One thing that coincides with maximum vacuum operation is, definitely, higher HVGO recovery. Make your flash zone pressure as low as possible, and flash zone temperature as high as possible - off course bearing in mind what I have said before - at the point where you can still maintain HVGO quality and tower heat balance. Some incremental HVGO yield is also possible at the expense of higher Concarbon and metal content - this should be defined based on downstream Hydrocracker performance and catalyst life cycle.
Here are some additional good links related to vacuum systems, from Graham-MFG website. I'm sure you'll find plenty of interesting information.
Keep us informed, and good luck!
http://www.graham-mf...ownloads/22.pdf
http://www.graham-mf...wnloads/210.pdf
http://www.graham-mf...wnloads/211.pdf
http://www.graham-mf...wnloads/214.pdf
http://www.graham-mf...wnloads/216.pdf
http://www.graham-mf...wnloads/220.pdf
http://www.graham-mf...wnloads/221.pdf
http://www.graham-mf...wnloads/232.pdf
Using stripping steam for column back-pressure control, as you also have noticed, cannot be considered as a serious and proper way to do it. In addition, if stripping steam incremental flow does not yield in higher HVGO recovery then it is, in simple words, wasting of money. If you can convince your process engineers that they are much smarter than you, they will accept any solution you propose. This is old trick. Engineers usually don't like clever senior operators, in spite of the fact they can learn so much of you. I am aware of this truth, since I spent almost two years working as ADU/VDU process engineer and plant manager, amongst other assignments I had in petroleum refinery. These days were, perhaps, the best I had in my career.
I would definitely close and isolate spillback PCV, and leave off-gas flow unrestricted. This is how your overhead system gains the maximum capacity. The only concern I would have in such circumstances, is whether I can maintain sufficient wash oil flow without flooding the wash oil bed, and keep tower heat balance under control due to increased distillate flow from the flash zone. If you have sufficient capacity in pumparound circuits, then there are no mistakes. Or at least you can conduct plant performance test, with different crudes you process, and see what is the optimum operating envelope of the tower at higher vacuums.
HVGO recovery: without design changes, there is no extra space for improvements. One thing that coincides with maximum vacuum operation is, definitely, higher HVGO recovery. Make your flash zone pressure as low as possible, and flash zone temperature as high as possible - off course bearing in mind what I have said before - at the point where you can still maintain HVGO quality and tower heat balance. Some incremental HVGO yield is also possible at the expense of higher Concarbon and metal content - this should be defined based on downstream Hydrocracker performance and catalyst life cycle.
Here are some additional good links related to vacuum systems, from Graham-MFG website. I'm sure you'll find plenty of interesting information.
Keep us informed, and good luck!
http://www.graham-mf...ownloads/22.pdf
http://www.graham-mf...wnloads/210.pdf
http://www.graham-mf...wnloads/211.pdf
http://www.graham-mf...wnloads/214.pdf
http://www.graham-mf...wnloads/216.pdf
http://www.graham-mf...wnloads/220.pdf
http://www.graham-mf...wnloads/221.pdf
http://www.graham-mf...wnloads/232.pdf
Similar Topics
Critical Pressure For Choke Valve SizingStarted by Guest_Sherif Morsi_* , 07 Nov 2017 |
|
|
||
Steam Pressure In Heat ExchangerStarted by Guest_mvanrijnbach_* , 15 Apr 2025 |
|
|
||
Steam Carrying Liquid From The Sour Water Stripping TowerStarted by Guest_kaidlut_* , 12 Sep 2024 |
|
|
||
Vacuum RecoveryStarted by Guest_Paper_* , 25 Feb 2025 |
|
|
||
Valve Cavity - Pressure Relief ValveStarted by Guest_CS10_* , 20 Feb 2025 |
|
|
