Jump to content



Featured Articles

Check out the latest featured articles.

File Library

Check out the latest downloads available in the File Library.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Gas Turbine Re-Ratings- posted in Ankur's blog

1

Stabilizer Column Control Scheme

high reflux rate

21 replies to this topic
Share this topic:
| More

#1 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 06 September 2018 - 11:21 AM

Hi

 

 

In our Refinery, we have NHTU unit which is Naphtha Hydrotreating Unit; it treats Combined Naphtha flow which is coming from CDU side. There is an issue regarding the sensitivity of NHTU Stabilizer control as following;

 

 

The NHTU Stabilizer is having a very large reflux to distillate ratio (more than 10 times). The reflux drum (V-8133) level controller (81LC2023) is manipulating the distillate flow (81FC2027), and the top tray temperature controller (81TC2035) is manipulating the reflux flow (81FC2025). A small percentage adjustment of reflux flow by the temperature controller will result in a large percentage change in the distillate flow, since the level control will need to adjust the distillate to compensate the change in reflux to control the level. This kind of control scheme is making frequent disturbance !

 

The reboiler which is used is a heater which is over designed. Normally panel operator prefers to stabilize the column through Overhead PC or reflux rate rather than touching reboiler during any upset. The heater is designed with four burners and two of them are online only. The heat input of the two burners is sufficient to achieve required bottom temp. 

 

So when the top temperature increases because of higher vapor flow comes with the feed, a sudden increase in the reflux rate to adjust top temp as well as drum level will drop because of high pressure. The reflux drum level controller will then need to reduce the flow which is going to GTU to bring the level up. After a few minutes the top temp will drop because of this high reflux and pressure will drop, so the the level at OVHD vessel will icrease suddenly and the flow to GTU will increase. This cycle will be repeated frequently during any small change in the feed. It will remains repeating until the panel operator break the Cascade control of the pressure and control it manually to stabilize the unit. In general when ever the pressure is taken in manual the effect of it to stabilize the unit is very clear. We tried to take reflux rate in manual during upset but it complicate the stabilization more.  We detuned PC but it still not capable to control any minor change in the feed. We tried to make the action of TC is a little bit slower but it never helped. It make disturbance in the unit when we tried to slower the tuning of TC. TC action is fast to stabilize the column currently!

This normally posts a big challenge to the control engineers especial when the system is not well understood. 

 

GTU feed is always fluctuate because of the current NHTU Stabilizer control scheme. A stable feed rate to the GTU should result in a better control of LPG product property. Thus, operations would be more confident in accepting higher GTU feed rate, instead of having large amount of hydrocarbon liquid being sent directly to the fuel gas system without any LPG recovery. More stable control of LPG recovery C3 mol % would be a result. 

 

Please avdise.

  

Attached Files



#2 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 06 September 2018 - 11:48 AM

The pressure of the column is 17 barg normally. The top tray temp is normally 108 C. There is no bypass for the finfan or seawater cooler to control the pressure of OVHD. Normally the pressure reading is taking before finfan to manipulate the PC which is controlling the flow to ATU. 



#3 Bobby Strain

Bobby Strain

    Gold Member

  • Members
  • 2,537 posts

Posted 06 September 2018 - 01:30 PM

You should get a book by Greg Shinskey. Try to attach a readable copy of your schematic.

 

Bobby


Edited by Bobby Strain, 06 September 2018 - 10:23 PM.


#4 PingPong

PingPong

    Gold Member

  • Members
  • 1,187 posts

Posted 07 September 2018 - 08:21 AM

As Bobby already said: we need a better flow scheme.  Moreover provide a flow scheme of the reboiler heater and its controls.  And a flow scheme of the stabilizer feed preheat (against bottoms, I expect).

What about DCS screen shots?

 

Now, it is not clear how the reboiler heater duty is normally controlled; and whether or how the feed is preheated.

 

What is your best estimate of duties of condenser, reboiler and feed preheat?  If condenser duty is small compared to preheat plus reboiler duties than that will also contribute to unstable operation.

 

Is the TC really on the top tray?  Or a few trays lower?

 

You seem to think that changes in stabilizer feed composition cause the imbalances, but did you ever analyze the feed?

But why would the feed composition change considerably?

I suspect that variations in reboiler duty (either caused by poor control system or poor operator interference) and/or feed preheat are a more likely cause for the problems in column overhead.

 

And what about the composition of the liquid distillate (GTU feed)? Is that routinely analyzed?

And if so is it always on spec with regard to C5 content?

 

In any case you should detune the action of the reflux drum level controller to smooth the variations in GTU feed rate.

 

Note also that C3 % recovery is determined by actual T and P at condenser outlet, not by control system.



#5 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 07 September 2018 - 02:52 PM

1dc8776ed782d0d5a80e981d0b7d83d2-full.jp



#6 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 07 September 2018 - 03:17 PM

Thanks all for your comment.

 

As per the latest scheme posted, the bottom flow is heated by splitting it to two streams; One stream is heated using F-8131 (two burners are running out of four) and the second stream is heated through convection zone at F-8130. So you can imagine the first stream temp increased much more than the second stream but both of them combined together before returned back to the column.  I had a discussion with the panel operator regarding controlling the heat input to the column, he doesn't prefer to touch the heater as it would make a change in temp by +- 3 C which he doesn't advise. Truly, I didn't try with them to control the disturbances through the heater heat input because they are convinced that it will disturb their unit based on their experience.  However, you can see that there is a temp controller at the outlet of the heater to control Fuel gas pressure in order to achieve the temp set value. For the heater F-8131, the BTM stabilizer stream temp will be controlled well but for the second heater F-8130 the BTM stabilizer stream temp at convection zone will not be controlled well because this heater is built for the reactor inlet temp and stabilizer BTM stream is just recovering the heat at convection zone.  So, in case feed increased, the stabilizer column will take time to reach steady state and normalized.  Any minor change will upset the column.  We run a tuning prog to set the best tuning figures for PC, TC, LC, FC of stabilizer column.  It helps yes but for a limit.  We are still unable to get a stabilized process around this column during any change in the feed.  CDU is using the same crude but from time to time we might change the cut point of naphtha slightly or sometimes we are getting lighter crude components slightly.  All these minor things affect on the column. TC is the top trays which is the most sensitive area to the LV traffic change.  We have a daily analysis for Naphtha components and generally they are normally change once per week. There is no analysis for the feed to GTU.  The level of reflux drum is de-tuned. 



#7 PingPong

PingPong

    Gold Member

  • Members
  • 1,187 posts

Posted 08 September 2018 - 03:24 AM

The flow scheme is still so small that I can't properly read the text and numbers on it.

 

As I said before: you have a heat balance problem with this column. Look up the process datasheets for the finfan and water condensers, for the reboiler heater F-8131 and for the reboiler convection section in F-8130 and for the feed preheat exchanger.

 

Another problem is the mixing of the two reboiler streams (from heater F-8131 and convection section in F-8130) upstream the TC that controls the firing. Unless there is a very long pipe with many bends between the mixing point and the TC there will never be a proper equilibrium established between vapor and liquid entering the column, so vapor quantity and composition going up into the column will not be constant but fluctuate despite constant value of TC. Take a look at the mixing length and piping diameter.

 

Moreover you have two temperature controllers on the column, one on the bottom stage (reboiler outlet) and one near the top stage. That cannot work, you can never fix two temperatures on a distillation column, no matter what books, articles or "experts" claim.

 

In your position I would break the cascade between the TC near the top tray and the FC of the reflux, and run the reflux on FC only. Set the reflux FC at the highest flowrate the pump and condensers can handle and monitor the C5 content of the distillate in some way. If C5 content is too high then slightly reduce the setpoint of the TC at reboiler outlet.

 

There is no analysis for the feed to GTU
What exactly happens in the GTU? Surely there must be a spec on the C5 content in LPG or butane product that is routinely verified?

#8 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 08 September 2018 - 04:07 AM

Tomorrow, I will take a snapshot for this scheme from DCS and post it here. 

 

Thanks for this highlight regarding heat balance around this column, but how would it help if I recognize there is an issue in this balance with current configuration?? 

The OVHD Finfan and sewater cooler dont have a bypass controller to control condensing rate or even the pressure at OVHD. Normally at summer they are running with four fans while at winter they use three fans. They dont regulate the seawater flow at the cooler and they dont regulate the speed of fan. So the major controller for OVHD part is is pressure. PC is taking input from OVHD stream before Finfan and accordingly it will take an action to relief this pressure through the flowcontroller to ATU.

 

Regarding mixing the two streams from reboilers, I will check and revert back.

You recommend to have one temp controller at this column. In the normal practice, operator will concern about this top tray temp TC. He believes this TC will ensure that the quality of flow to GTU is met. It is an indicator to limit C5 quantity and H2S to be sent to GTU. It will ensure that most of C5 will be remained at the bottom.

 

The flow which is sent to GTU is normally not tested because of sampling issue currently. I requested to correct this sampling point, so we can have analysis for the flow going to GTU.

 

In the other hand, there is an issue with GTU products; LPG & Reformate from its Depro column. It is a copper strip failure for LPG product from GTU. They believe the issue is because of NHTU OVHD temp. They believe that when OVHD temp at OVHD vessel drops to 80 C, they will suffer from copper strip failure at GTU. There is no final investigation report for this failure so far and this is what operator believes on. Whenever I requested to increase the pressure at GTU Depro column to increase C4 recovery at LPG, panel opreator refuses because he has a fear from copper strip failure. He dosent believe that NHTU operator stripped all H2S from the process before send it to GTU.

 

So if I try to break TC of NHTU top tray stabilizer and keep running reflux on auto, people will refuse because of the importance of this top temp for their products quality flow as they believe.

 

Please advise.



#9 PingPong

PingPong

    Gold Member

  • Members
  • 1,187 posts

Posted 08 September 2018 - 07:35 AM

Once again: you cannot control (fix) two temperatures on different column stages.

So you have to let one go free. I suggest you disconnect the TC near the top and maximize the reflux on FC.

Alternatively one could leave the reflux on TC but disconnect the reboiler outlet TC and fix the total reboiler duty, however in your design that is not possible due to the configuration of two parallel reboiler heat sources.

.... how would it help if I recognize there is an issue in this balance with current configuration??

Insight always helps. In this case you can forget that feed composition changes are the (main) cause of the column problems.

You need to realize that you have a lousy design with a lousy control system, so the best you can expect is a lousy operation.

The OVHD Finfan and sewater cooler dont have a bypass controller to control condensing rate or even the pressure at OVHD. Normally at summer they are running with four fans while at winter they use three fans. They dont regulate the seawater flow at the cooler and they dont regulate the speed of fan. So the major controller for OVHD part is is pressure. PC is taking input from OVHD stream before Finfan and accordingly it will take an action to relief this pressure through the flowcontroller to ATU

As there will always be a vapor distillate a PCV in the vapor product flow is OK for pressure control. Only if there was only liquid distillate you would need condenser control to control column pressure.

Controlling the sea water condenser outlet temperature in some way would allow adjusting C3 recovery as well as amount of dissolved H2S in GTU feed. But if it is too much trouble for the operators to adjust the sea water flow every now and then you can only sit back and watch.

He believes this TC will ensure that the quality of flow to GTU is met. It is an indicator to limit C5 quantity and H2S to be sent to GTU. It will ensure that most of C5 will be remained at the bottom.

First of all: it will not assure that most of C5 remains in bottom because the TC setting on the reboiler outlet also influences that. Most of the time those two TC's are fighting each other: the top one trying to send more liquid back to the bottom, the bottom one trying to blow more vapor up to the top.

That top TC adjusts the reflux but can never make the reflux larger than the max capacity of the pump and condenser. Hence my earlier suggestion to disconnect the TC and run max achievable reflux by FC setpoint.

 

I still don't know what exactly happens with the liquid distillate in the GTU.

Is it amine treating or not?

Is there a deethanizer and a depropanizer?

Or what?

 

If there is in the GTU no amine treatment of the liquid distillate from the NHTU then the stabilizer condenser outlet temperature and pressure will impact the amount of H2S in the GTU feed: the higher the stabilizer condenser pressure or the lower the condenser temperature the higher the H2S concentration in the GTU feed. The TC setting on the stabilizer top tray does not affect that.

Same story for the condenser conditions in deethanizer and depropanizer: high P and/or low T at condenser outlet can have impact on H2S content of column bottoms.

 

I wonder who is in charge in your plant.

Normally operators follow instructions from the responsible process engineer, I suppose you or your supervisor.

Operators do not have a chemical engineering degree and do not really understand what is going on in the plant, even if they think they do. They know from experience that if they increase one thing then somewhere else another thing will increase or decrease, they may have a theory in their mind why, but usually that theory is incomplete or simply wrong.

Anyway, that is your problem, or that of your supervisor. I can't do anything about that.


Edited by PingPong, 08 September 2018 - 07:37 AM.


#10 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 08 September 2018 - 08:40 AM

PingPong

 

I used to be RFCC engineer. I moved to a new refinery where most of process engineers are new and the old ones were resigned. I started the work in this refinery this year but as a process control engineer. I came to face this issue when I ran APC at GTU and couldnt achieve the target C3 recovery at LPG because Operator has the fear to increase the high allowable limit of pressure at Deproponizer column at GTU. If GTU LPG went off in copper strip © , the sphere at offsite will be sent to flare totally to get rid of this off-spec product. There is no reprocess facility to re-process this LPG again at process units. This is why the panel operator doesnt like to take any further increment at pressure or temp at GTU Depro. They always claim NHTU operating profile is causing this off-spec.

There is no amine or caustic treating units between NHTU and GTU . No De-ethenizer column. There is no investigation was conducted for this issue, So I started now to play the rule of process engineer by asking to correct all sampling points and test the units as well as to simulate the current composition to get the best P &  T for NHTU. I dont know the history of these reboilers but it seems I will start digging in there design data.

 

BTW, During any upset TC MV is fully opened 100% usually. I told you at the begining, I tried once to stabilize the column during upset by disconnecting TC and run it on auto, yet the column never stabilize unless the pressure controller was taken on manual. 



#11 PingPong

PingPong

    Gold Member

  • Members
  • 1,187 posts

Posted 08 September 2018 - 12:28 PM

Can you post a flow scheme of the GTU, including controls ?



#12 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 08 September 2018 - 09:53 PM

832cf7974b8df6923c36915646f8c0df-full.jp



#13 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 08 September 2018 - 10:38 PM

I didnt know how to upload the file at this website. Plz check following link; it has a word file which shows DCS PrtScn in details.

 

https://files.fm/f/6mxsyjkf

 

I would like to correct one thing; The btm flow splits to two streams; one if fed at the convection zone and radiation zone of F-8131. The second stream is heated through Splitter reboiler at radiation zone of F-8240 


Edited by Fatema1987, 08 September 2018 - 11:01 PM.


#14 PingPong

PingPong

    Gold Member

  • Members
  • 1,187 posts

Posted 09 September 2018 - 03:56 AM

To upload here click on the button [More Reply Options] at bottom right of reply screen.

Then scroll down and click on [Browse....]

Then click [Attach This File]

And when your message text is complete click [Add Reply]

 

I have attached your file here:

Attached Files



#15 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 09 September 2018 - 06:27 AM

plz find attach

Attached Files



#16 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 10 September 2018 - 11:53 AM

I am wondering if building PCT would help in this case or not?



#17 Bobby Strain

Bobby Strain

    Gold Member

  • Members
  • 2,537 posts

Posted 10 September 2018 - 04:07 PM

?

Bobby



#18 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 11 September 2018 - 12:23 AM

?

Bobby

I already attached the DCS screen shots in word file. Hopefully you can see it. Please advise.



#19 PingPong

PingPong

    Gold Member

  • Members
  • 1,187 posts

Posted 11 September 2018 - 02:03 AM

I am wondering if building PCT would help in this case or not?
What do you mean by "building PCT" ?

#20 Fatema1987

Fatema1987

    Junior Member

  • Members
  • 14 posts

Posted 11 September 2018 - 05:47 AM

I am wondering if building PCT would help in this case or not?

What do you mean by "building PCT" ?

Having pressure compensated temp PCT at Ovhd to set the top temp. I am thinking in this way to minimize loses of C3 sent into ATU instead of sending it to GTU. So PCT will control reflux while the ratio of reflux/feed will control the heater temp TC.

What do you see?

#21 PingPong

PingPong

    Gold Member

  • Members
  • 1,187 posts

Posted 11 September 2018 - 07:03 AM

You have a PC on the system that should maintain constant pressure, so I don't see the point of using PCT for top tray TC.

Besides: then why not also use PCT for the TC at reboiler outlet?

 

C3 losses from stabilizer reflux drum to ATU are determined by seawater condenser outlet temperature and pressure.

There is a TI at the finfan outlet measuring 67 oC but that number is not important. However there is no TI where it really matters: at the outlet of the seawater condenser.

 

...while the ratio of reflux/feed will control the heater temp TC.
I don't know how exactly that would look on paper.

 

Try to find the datasheets with the design duties of the exchangers and reboilers around the stabilizer. As I wrote before: I suspect that the condenser duty is small compared to those of reboilers and feed preheat.



#22 Bobby Strain

Bobby Strain

    Gold Member

  • Members
  • 2,537 posts

Posted 11 September 2018 - 09:12 AM

We are all guessing what causes such instability. You will need to enlist the services of a fractionator control specialist to resolve your issues. And the first step will be to simulate the operation with a reliable process simulator. Steady state values can allow you to check where the temperature should be controlled. There must be hundreds of installations like yours without operating problems.

 

Bobby


Edited by Bobby Strain, 11 September 2018 - 01:03 PM.





Similar Topics