17 replies to this topic

[merac]

Hi
I have some doubts related to CONTROL OF A DISTILLATION COLUMN:
1) Pressure control:
The pressure of the top of the distillation column can be controlled to be a constant value but there are sometimes that it is prefered to minimize this pressure as much as possible in order to facilitate separation and reduce reboiler duty. In this situation, when atmospheric temperature decreases, the condensation capacity increases and top pressure decreases so the pressure control of the column let this pressure to decrease.
My question is: how do you decide which type of control is better: constant pressure or floating pressure? Which are the factors I should considered in order to decide it?

2.Flooding control
In order to control the flooding of the column, I can control differential pressure (top/bottom of the column) so if this differential pressure increases it would mean that flooding is increasing, but:
Which variable should be manipulated to reduce flooding?
Which is the control strategy? Can anybody show me a control scheme?

Thanks

[Padmakar Katre]

Dear,

First I will reply to your 1^st query; i.e., how to decide the column pressure:

Except when distilling heat-sensitive materials, the main consideration when selecting the
column operating-pressure will be to ensure that the dew point of the distillate is above
that which can be easily obtained with the plant cooling water. The maximum, summer,
temperature of cooling water is usually taken as 30 °C. If this means that high pressures
will be needed, the provision of refrigerated brine cooling should be considered. Vacuum
operation is used to reduce the column temperatures for the distillation of heat-sensitive
materials and where very high temperatures would otherwise be needed to distill relatively non-volatile materials. The basic fundamental to apply here is the dew point temperature of the O/H vapor at the column pressure and the second thing i.e. your cooling medium temp will be less than the dew point temp of the column overhead vapors otherwise it will result into increase in column pressure. For to decide the column pressure which you need to have is cooling medium conditions, heating medium conditions and the complete properties of the components you are going to handle.

As you told that the column pressure is always minimized to facilitate the separation the fact behind is your relative volatility goes on increasing as we start reducing the separation pressure but we can minimize to that extent that your overhead vapors dew point temp would be greater than your cooling medium temp as I mentioned above.

Now about your second query - i.e., Problem of the column flooding:

There are two types of flooding which results into high diff pressure across the trays or packing:

• the first one is entrainment flooding which is result of the excess vapor flow rate or the vapor velocity which will cause the resistance to the liquid flowing downward;
• or in other case if your liquid flow rates are too high then there you will find downcoming flooding.

Every column is designed for a certain L/D ratio if it gets upset by any of the above two means you will find flooding in the column. So normally first of all check your flows across the column. Apart from the above there may be some sort mechanical misalignment of the column internals which also lead to increase in diff pressure. In the case of distillation columns, if you have a high reboiling rate that causes excessive vapor generation, there will be carryover of the liquid to the tray above; or in the reverse case if your feed or reflux flow control fails and you have high liquid flows downward then also the same situation will occur. As I told you, each column is designed for certain L/D ratios because apart from flooding, there different parameters like weeping/coning and all. Any upset in this will cause your tray or packing efficiency to come down and result in an upset in the product specification.

I would like to share my experience about the EO (Ethylene Oxide) absorption by water. We used to maintain a certain flow of water to absorb EO in the inlet gas to absorber. In the event of water temperature goes high for which EO absorption is quite sensitive and that will result into Escaping EO which will increase flooding in the tower. So, we have general guideline that if all the liquid flow and the gas flow is constant and if there is an alarm of High Diff Press our first action used to increase the water flow. This may -or may not - be the case with the other column so while troubleshooting we have to check all the parameters and find which is culprit and get a rid of it which will solve your problem. I hope this will help you to understand a little about your query.

Waiting for your reply,

Regards,
Padmakar Katre

[merac]

Thank you for your reply, but I think I have not explained very well my questions.
My doubts are related to BASIC CONTROL OF A DISTILLATION COLUMN, not with the fundamentals of the process, with this I mean that I have doubts related to which is the controlled variable, manipulated variable, type of control (feedback, feedforward, cascade, etc); and my doubts are related to:
1.CONSTANT PRESSURE OF FLOATING PRESSURE CONTROL
How to decide between a constant pressure control or a floating pressure control?
Which factors should I considered in my process when thinking about which type of pressure control design?

2.FLOODING CONTROL
Which is the controlled variable (differential pressure between top and bottom of the column)?
Which is the manipulated variable to reduce flooding?
How is the control scheme?

[Padmakar Katre]

Dear,
Here I am uploading some articles related to distialltion control. I don't understand what you mean by floating pressure control which I have never seen.Floating and control I think these 2 words are quite contradictory here anyways just go through these papers and still you have a doubt you are welcome.Now your second query i.e. differentail pressure control I have never seen this type control system which controls the diff pressure by controlling a variable like your reboiling rate .vapor flow ,liquid flow because there are many reasons for flooding in the tower and I have never seen or heard anywhere about this concept of the diff pressure control i.e. closed loop what I have seen or heard is open loop of diff pressure.Anyways Even I will wait to see the reply if some one knows this kind of control system I am doubtful.

Regards,
Padmakar Katre

Attached Files

•  Distillation.zip   82.19KB   784 downloads

[merac]

Hi
After having thought a bit about how to control flooding in a distillation column, I have thought the following control strategy (that it is shown in the file I have attached), but I still have some doubts.

Control strategy (See control scheme attached):

-Controlled variables: 1) Differential pressure (DP) between the top-bottom of the column, in
order to control flooding. If the differential pressure exceeds a value
(setpoint of the DPC) this would mean that there is flooding in the column.
2) Bottom quality product. This variable is controlled by the
temperature of one of the last trays with a temperature control (TC).

-Manipulated variables: 1) Steam flow that enters the termoshyphon.
- If the bottom temperature of the column decreases, it would be necessary to increase steam flow that enters the reboiler.
- If flooding occurs, it is necessary to distinguish two types of flooding because the way in which the manipulated variable must change is quite the contrary:
-If flooding is due to an excessive liquid flow in trays-------steam that enters to termoshyphon should increase (direct action) in order to vaporizate the liquid of the tray and reduce flooding.
-But if flooding is due to an excessive vapor flow that causes that the liquid of one tray contaminated the liquid of the upper tray-----------it would be necessary to decrease steam flow (inverse action) that enters to termoshyphon in order to minimize vapor flows in the column


- Control strategy:CONTROL WITH RESTRICTIONS (OVERRIDE CONTROL)
While the differential pressure is less than the setpoint (SP), the steam flow that enters the termoshyphon is manipulated so as to control the bottom temperature of the column (the setpoint of the FC is the output of the TC), BUT if the differential pressure exceed the setpoint (SP), the important variable to control is the flooding of the column (differential pressure), not the bottom temperature, so the setpoint (SP) of the FC is the output (OP) of the DPC not the output of the TC (See control scheme attached):.

-My doubts are the following:
- how is flooding controlled in real life?
How can the DPC be defined if we want to control the two types of flooding? Is it possible?
Can a controller be defined with two types of actions: direct and inverse?

If anybody can help me , I would be really pleased. Thanks

Attached Files

•  flooding_control_in_a_distillation_column.doc   195KB   461 downloads

[JoeWong]

merac,

Do you want to control flooding but ignore product quality ?

DP is measure from top tray to bottom tray. Thus tray below feed will experience higher liquid loading and tends to flood ? How to fix the DP set point.

I guess the control strategy would be

- target on product quality
- design tray / packing for worst case
- monitor dP, T, P, flow, etc

Please ignore if not suit to your though.

JoeWong

[Padmakar Katre]

Dear Merac,
I really appriciate your efforts.But the problem as I stated you earlier is that there may other reasons to have increased DP.If we consider for a while that your reboiling rate is the main culprit to get higher DPs and rest of the parameters are ok the liquid flows are quite steady and within the operating range still we should not play with your reboiling because once the temp profile gets upset you have no other option than to divert your products to the crude tanks from where you will reprocess after stabilization of the column.So you have to be more concerned about the product quality.By having DP in a closed loop I think this is an addition of a reason to upset the column.In a normal running or a steady-state opertaion we don't have much variations in the steam flows to the reboiler now if your steam flow is going to be governed by two controller i.e. your control tary temp closed loop output i.e. TIC and Diff Press Closed loop PDIC output as set point to the FIC of steam flow.variations in steam flow will cause unsteady-state in the steam headers to a certain extent.Now you have to consider the following cases
1.If there is minor leak in the impulse tube or PT(LP side) manifold which will cause the DP to go up and this will also cause the steam flow to reduce and upset in your column profile and the subsequent operations downstream of this column.
2.In case of the column bottom LIC sticks close you will find the potential hammering in the column.

I have consulted today so many process engineers and senior process as well instrumentaion engineers and senior instru personals everyone has the same opinion let the DCS operator or the plant process engineer to troubleshoot the reasons for the increased DP in the column and to take corrective actions that means to have DP in the open loop and not in the closed loop.
The second question I asked them is have you ever seen any where in any distillation column PD in closed loop nobody was able to give some concrete example.
I have a counter opinion since my first reply to your query If I am wrong or making a mistake I am so sorry ,if this is so ignore my all replies.

[suhas]

We can control the flooding in the column using advance process control technique (Multivariable control techniques) In most of the cases normally column is either limited by a liquid or vapor capacity. In this technique model is built with flooding factor and the reflux flow/ steam flow and this controller will assure that the reflux flow/ steam flow will not go beyond the limits that resulting in flood. This limits are determined by the model.

[djack77494]

merac,
In partial response to your queries, I think the simplest answer, alluded to by JoeWong, is that you do not normally control column flooding. You are normally concerned with product flowrates and qualities, and these set your column operating conditions. As long as no part of the column is flooding, that is fine and you achieve optimum steady state control. If you start to flood, then "things go to hell in a handbasket" as they say. Your operation will not be satisfactory, and you must take steps to move out of the flooding regime. (P.S.-Best separation is achieved just before you start flooding.)
Doug

[merac]

Hi djack77494!
OK, you don't normally control column flooding, but in case someday the column starst to flood, what is it done?
Is there any control strategy (using advanced conventional control but not multivariable control) prepared in the control system of the unit? or is this operation let to the control operators who work in control room, to solve the problem?

Hi suhas,
Is it only possible to control column flooding by multivariable control?
Can it be controlled with the restriction control strategy that I proposed some posts before (Posted Oct 21 2007, 02:02 PM)?

Thanks

[suhas]

You can definitely control using the restriction control method provided that you should know the reason why column is limiting. To control the flood first thing u should know is the reason to flood. Whether it is vapor limited or liquid limited, and which section of column. Also I think it is not required to have tighter control on flood by keeping steam flow on DP, if it’s a limited by vapor. But instead of that it can be done easily by the operator and their experience. Also you can make a flood factor tag on the DCS which will be a good indicator of flood rather than DP. In flood factor correlations you can simply place max liquid is reflux + feed and maximum vapor flow as reflux + distillate, using design properties you can simply make a DCS tag of flood factor which you can control. But as it is an approximate value you have to find out at which value your column is actually flooding. I will prefer to have this tag on DCS and according to that operator will take some actions manually rather than just only seeing a DP.

[pawan]

Dear Mr. Katre
I am currently working on debottlenecking of our MEG plant. I would like to have ur help on EO absorber. As U already mentioned in this and my some other post that it is very sensitive to temp.

I found that our EO loop is limiting on pressure and for increased production (~10% rise), columns are going to be in flooding zone. So our team suggested going from Tray to Packed tower which is quite an expensive job. May be ~ Rs. 30-40 Million of investment.

My problem is that I am not confident on the data of dilute EO solution data (Activity coeff etc.). Currently our Cycle water is at ~37°C in peak summer at the inlet of absorber. I would like to have some idea about improvement if I reduce the temp say upto 25°C or so at th einlet by chilling. (Surplus Chilled Water is available from waste heat).

This iwll reduce the water circulation, may reduce steam requirement in stripper (I am not sure) and will avoid conversion of towers from tray to packings.

Can U give me some idea about how much reduction is possible in water flow, currently we are operating at 3% EO at the scrubber outlet in the water @14.5 bar (abs) pressure.

[Padmakar Katre]

Dear Pawan,
The thread is something with the different concern and you asked me some what different anyways
Can you give me the terminal conditions across the your absorber/scrubber.
1.gas flow and its compositions with its temp, pressure
2.liquid flow temp/pressure
3.your Cycle Gas flow to the EO Reactor and if possible your ethylene and oxygen flows with their purity.
4.MEG content in Cycle Water
5.Dia and height of scrubber
6.No of trays

if you can't share this data here you can mail me.My mail id is
pady_chem_iisc@inbox.com

[djack77494]

Dear Pawan & Padmakar S Katre,
I know your queries and responses are very important. However, the last two posts really address a new topic and are best treated that way. I would suggest that they be redone (reposted) as a new topic.

In addressing merac's question, I confess to not being sufficiently versed in modern process control to provide an adequate answer. I would think that by monitoring and alarming certain process parameters - e.g. high delta P across a column - incipient flooding could be detected (or inferred). Then actions could be taken to move in the direction away from flooding. Either vapor or liquid traffic of both in the column would have to be reduced, depending on what type of flooding is likely. You really want to avoid flooding altogether, because if you wait until you're sure flooding has occurred, then you probably have a major process upset on your hands. There are probably sophisticated modern controls that could better infer the approach to flooding than my clumsy delta p method.
Regards,
Doug

[JoeWong]

merac,

I guess this has gone into Advance Process Control strategy...

You may look into
Distillation Column Flooding Predictor and DCFP Brief

You may consider to contact
George E. Dzyacky...
M. Geoff Mphahlele, University of Texas at Austin,
R. Bruce Eldridge, University of Texas at Austin

JoeWong

[samritikalra1]

JoeWong:

Dear,

IN all the process systems we should first consider the process safety, then equipment constraints and at last the product quality. we should not ignore dP controlliing product quality.

Regards,

[Hansson]

Dear Merac,

You have asked two very interesting questions. I will answer your second question first:

I have designed a DP controller that works very well. It is implemented in a main fractionator where naphtha, kerosene and diesel are separated from extra heavy gasoil. Any diesel we can extract from the low value bottom product increase over profit. Hence, separation between bottoms and diesel is to be maximised. Increasing feed inlet temperature increase diesel yield but at the same time loads up the tower. We have experimentally checked where our flood point is (DP value at incipient flood) and have then the differential pressure of the section of the tower that is most prone to flooding as the target DP (DP set-point close to but BELOW the DP that corresponds to incipient flooding) and the feed temperature as the variable. This control strategy has saved us more than 10 M$/year compared to the previous control strategy.

In a simple tower such as a C3/C4 splitter, the economical incentive is less, since you do not get extra pay for extra high quality. You just stay on specification with a minimum effort. So in short, a DP-controller can be a good idea when any incremental separation adds to your profit, otherwise there is no use for it.

As for your first query – should pressure be controlled or allowed to float?

If your pressure swings are great, your tower might go from dumping to flooding at constant reboiler duty, none of which are good. If pressure swings are moderate, you may still be in the range where tray efficiency is pretty good and separation is not that much affected. However, if you use temperature control to achieve quality control, then you have to pressure compensate your TC-setpoint in order to get constant quality. I personally do not like to have floating pressure control but would recommend you to use a fixed pressure as long as possible. If pressure control can not be achieved at your current set-point due to too high ambient temperature for example, I prefer to increase the pressure set-point so that you remain in the control range and change the TC-set points so as to keep constant product qualities.

Regards
Hansson

[gigigawa]

Dear Merac,

if you are in jet flooding conditions, in order you should:

- increase top tower pressure

- reduce tower reflux

- reduce reboiler duty

- reduce the tower feed

Shortly you should see that the top or bottom flow will increase at the same tower feed rate then the average hour flow.

When it came back to original flow you are out of flooding conditions.

Regards

Marco