30 replies to this topic

[rabindra]

Hello

 

In one of the columns (Deethaniser) in my refinery (not commissioned) there is a butterfly valve at the inlet of the hydrocarbon line. The note for the valve says that

 

"Valve to allow operator to throttle the inlet pressure to the thermosyphon to achieve optimum performance. Downstream pressure gauge shall be readable from the valve".

 

As per my understanding thermosyphon operate by the difference in density of the inlet liquid hydrocarbon and outlet mixture of liquid and vapor hydrocarbon. The driving force for the liquid to flow through the reboiler is the liquid head from column to the inlet of reboiler and the density difference. Generally in other units I have seen that if there is a valve at the inlet of the thermosyphon reboiler, it is fully open to minimise pressure drop. Now my question is:

 

If the valve is throttle to control downstream pressure, how would it affect the performance of the reboiler? Instead the liquid level in the coulmn shall increase to provide more head for the hydrocarbon flow.

 

Please correct if my analysis is wrong. A copy of the P&ID is attached for reference.

 

regards

 

rabindra

Attached Files

•  Butterfly Valve in Reboiler Inlet.jpg   159.86KB   41 downloads

[AlertO]

hi rabindra

 

The circulation rate through thermosyphon reboiler is designed based on the exact static praasure and the pressure drop along the circulation loop.  However, the actual system is normally generate less pressure drop than design due to some margin provided. Therefore, based on this configuration (without the throttled valve) the fluid can be flown through the reboiler higher than the expected value (while the rundown flow is reduced). hence, the butterfly valve is used for throttling to creat some pressure drop to the circualtion loop. This aims to achieve the target flow rate in the circulation loop.

 

Hope this may help

[fallah]

rabindra,

 

The PID seems having "As Built" status, and the valve is there to regulate required gravity flow which couldn't be created by reboiler/column levelling as per performed hydraulic design, probably due to construction limitations...  

[Pilesar]

The column sump liquid is at its bubble point temperature at the top of the liquid level. At the bottom of the column, the pressure is higher due to the liquid static head and therefore the liquid is subcooled below its boiling point temperature. Increasing the pressure drop through the reboiler inlet piping reduces the pressure at the inlet to the reboiler tubes. The result is that there is less tube length required to heat the process liquid and the boiling regime begins lower in the exchanger which can improve the overall heat transfer coefficient. It is a common practice to design thermosiphon inlet pipes to approach 30% of the system pressure drop. Valves are designed into the inlet pipes of vertical thermosiphons when there is a wide range of possible operating conditions.

[Bobby Strain]

I expect that when you start up the system you will always leave the valve fully open. These systems tend to be self-regulating.

 

Bobby

[PingPong]

I suggest to leave the valve always fully open. This is the product of a design engineer who does not really understand what he is doing. Unfortunately there are many of those around nowadays.

 

When a process design engineer (like me) specifies a vaporization (say) of 20 wt% on a datasheet for a thermosyphon reboiler, then that is just a typical number which is known to be okay. Then the datasheet goes to an engineer who uses HTRI to size the reboiler, and the younger they are the more exact they try to match that (arbitrairy) 20 wt%, sometimes even proposing a valve or rerstriction orifice in the reboiler inlet line to reduce the liquid circulation. Personnally I don't care at all what the actual vaporization will be when the unit operates, and therefor I do not allow any restriction in the line.

 

The purpose of the reboiler is to provide a certain vapor flowrate (which corresponds to a certain duty) to the column. Let's say vapor is to be 10 t/h.

At 20 wt% vaporization that means that the hydraulics of the system should be such that 50 t/h of liquid enters the reboiler.

If the hydraulics turn out to be such that theoretically 67 t/h of liquid will flow to the reboiler, while it produces 10 t/h vapor, the vaporization will be 15 wt% instead of 20 wt%. So what? I don't lose any sleep over that.

 

And look at that P&ID: the operator has to adjust the valve while looking at a pressure gauge. The operating pressure of a deethanizer is usually in the range of 20 - 25 barg so that gauge probably has a scale from 0 - 50 barg. But 1 meter liquid LPG column is only 0.05 bar so how is anybody able to carefully adjust that valve? But even if that pressure gauge were super accurate: what pressure should the operator be aiming at to obtain the correct liquid flowrate? It is too silly for words really.

[aroon]

I agree with AlertO and Pilesar. Both have given good reasons and those seems to be practically true. Many time it is not possible to design any system (here it is inlet piping of the reboiler) with 100% perfection. This is because of the following reasons:

 

1. There are several margins involved during design phase (e.g. margin demanded by end user, etc.) 

2. In most of the designs, there are multiple cases involved (some number of operating cases, turn down, SOR, EOR) so the basis for the design is always the conservative case. This basis cannot optimized one for all cases.

3. Sometimes, unavailability of some data tend design engineers to use thumb rules.

 

There are several more reasons. So the given option of throttling the butterfly valve is achieve efficient operation in all the cases.

 

- Aroon

[TS1979]

I don't see any reason that requires a throttling valve there. Maybe that valve can only be used for isolation.

 

For the reboiler, the best performance shall be achieved when you maximize your recycle speed. High recycle speed has a few advantages for the reboiler operation:

1. Reduce the reboiler fouling

2.Avoid potential hydrocarbon coking inside reboiler

3. More stable gas flow rate.

 

The boiler duty should be controlled by heat medium flow (such as steam) to the reboiler based on column bottom temperature. Heat duty controls the vapor load inside the column.

[PingPong]

Therefore, based on this configuration (without the throttled valve) the fluid can be flown through the reboiler higher than the expected value (while the rundown flow is reduced)

That is not correct. The rundown is not affected by the liquid circulation over the reboiler.

 

.... therefore the liquid is subcooled below its boiling point temperature. Increasing the pressure drop through the reboiler inlet piping reduces the pressure at the inlet to the reboiler tubes. The result is that there is less tube length required to heat the process liquid and the boiling regime begins lower in the exchanger which can improve the overall heat transfer coefficient.

That heating of the process fluid till its boiling point is peanuts compared to the total duty of the reboiler.

[rabindra]

........Therefore, based on this configuration (without the throttled valve) the fluid can be flown through the reboiler higher than the expected value (while the rundown flow is reduced). ..........

 

I agree with PingPong, since the same amount of material is coming back to the column it wont affect the rundown flow rate.

 

 

 

1. There are several margins involved during design phase (e.g. margin demanded by end user, etc.) 

2. In most of the designs, there are multiple cases involved (some number of operating cases, turn down, SOR, EOR) so the basis for the design is always the conservative case. This basis cannot optimized one for all cases.

3. Sometimes, unavailability of some data tend design engineers to use thumb rules.

 

There are several more reasons. So the given option of throttling the butterfly valve is achieve efficient operation in all the cases.

 

- Aroon

 

In SOR and EOR case, since the liquid level will be same available head will be same. Only difference will be the rate of the heating media to the reboiler, which will control the circulation flow in the reboiler. Hence I do not think there is a need of the butterfly valve for control of the liquid flow during SOR/EOR case.

 

I would like to know your views.

[fallah]

rabindra,

 

I still await to know the status of the PID you are referring to. If the supportive documents of that PID is available, you can refer to the calculation note of reboiler hydraulic design and levelling with respect to the column. It might the elevation difference between reboiler and column not to be as per the design value due to construction limitations. Please let's know about the conclusion might be made...

[Bobby Strain]

Since this installation seems to be complete already, maybe Rabindra will give us some feedback whenever it is put into operation. Will be a good lesson for us all.

 

Bobby

[chemsac2]

I tend to agree with PingPong in that people seem to be giving too much credence to 20-25% vaporization and resort to all sorts of jugglery to ensure that kind of vaporization in all operating scenarios.

 

I would size thermosyphon system to ensure vaporization between 10-50%. Lower limit is to ensure stable thermosyphon and upper limit to avoid dry film heat transfer. As long as vaporization for all scenarios falls within this range, valve is not needed.

 

However, for turndown cases, too high a static head can lead to % vaporization dropping below 10% and hence unstable thermosyphon action. An inlet throttling valve can come handy in such cases to limit circulation rate and bring vaporization back to stable region. Too high a static head is typically when too much of a margin is applied in all of the sub-systems. 

 

Even HTRI is okay with installation of inlet throttling valve. HTRI has a provision for adding inlet valve pressure drop. Attached HTRI presentation also supports this.

 

% throttling of inlet valve can not be judged by pressure gauge at exchanger outlet, but would have to be based on operator observation of thermosyphon and subsequently column stable operation.

 

Regards,

 

Sachin 

Attached Files

•  Thermosyphon_Valve.jpg   51.68KB   43 downloads

Edited by chemsac2, 30 March 2014 - 01:30 AM.

[breizh]

Hi,

In addition of the answers above let you consider this paper ,

Hope this helps

 

Breizh

 

[aroon]

Rabindra,

 

 

In EOR and SOR, the composition of the column bottom liquid is always different than the normal operating case, so even if you maintain the constant level. It will not give guarantee of optimum static head for the circulation due to change in density of the liquid, which is based on composition.

 

In addition, heating media is to maintain a temperature of column bottom as per process requirement and not to have desired circulation. Circulation rate only can be achieved by adjusting static heads and circulation rate is very vital to have stable column operation.

 

Let me also know your views on this.

 

- Aroon

[AlertO]

Hi

 

Seems to be a long story. However, we have both designs (with and without the valve) around the world. The valve is there because there is some benefit, isn't it?  Just go back to see the need of that valve.

 

The valve is needed in case you want to adjust the flow; therefore,...

 

if your operation is quite constant in a narrow range, of cource, that valve is useless but,

if you have a number of operating ranges e.g. EOR, SOR or turndown, the valve becomes useful.

 

Just see and understand the plant operation, the answer is there.

 

Hope this may help you

[katmar]

In my personal experience I have only once seen an installation that required a valve in the reboiler circulation line.  This ocurred when a vastly oversized reboiler was used.  In operation it was unstable.  It was so oversized and generated so much vapor that that the pressure drop in the return line caused the circulation to stall. With no fresh liquid coming into the reboiler, heat transfer then also stopped and of course with no more vapor being generated the pressure drop decreased and then a surge of liquid would rush into the reboiler tubes and the cycle repeated.  The operators described it as "burping". Putting a butterfly valve into the inlet line solved the problem and I guess we were probably only using the bottom 25% of the tube length.

 

I see that Kister (Distillation Operation pg 446) refers to Lieberman (Process Design for Reliable Operation) who gives an example of an installation where restricting the reboiler inlet line almost tripled the reboiler capacity.  However, I would agree with PingPong that provided you get your design close to the correct values then it will not matter if the fraction vaporized varies a little from the design case.  But it seems that if you get the design hopelessly wrong and the circulation is much too high then you can get into the situation described by Pilesar and seemingly confirmed by Lieberman. 

 

I suppose that the margin for error will also increase as the overall size goes down.  If your design calls for a 4.5" inch ID inlet pipe to provide the required restriction to get the correct flowrate and you are forced (by standard sizes) to use a 6" pipe then the restriction will be significantly less than what you wanted it to be.  If your design results in a requirement for a 20.5" line and you have to install 22" then the difference is much smaller.

[Bobby Strain]

Lieberman probably fail to mention that the reboiler was cleaned, too.

 

Bobby

\

[PingPong]

It was so oversized and generated so much vapor that that the pressure drop in the return line caused the circulation to stall.

Normally the reboiler duty, and therefor the vapor generation, is controlled by a TC a few trays higher on the column that affects the heating medium in some way. In such case an oversized reboiler is counteracted by reducing the flow of heating medium, or backflooding the exchanger with steam condensate, or ... (whatever). Apparently that was not the case with this reboiler so it could produce as much vapor as it liked?

 

 

Putting a butterfly valve into the inlet line solved the problem and I guess we were probably only using the bottom 25% of the tube length.

Does this mean that the operators allow just that amount of liquid to enter the reboiler that they want to vaporize in it? So it is not a thermosiphon anymore, but a total vaporizer? What about fouling in the long run? Or accumulation of heavies?

Apparently it works good enough in this particular case, but this is a fix, not a solution. The solution here would have been to change the duty control of the heating medium for the reboiler, or add one if there was none.

 

I don't have Lieberman's book, so I have no idea what the real problem there was.

I doubt that it had to do with heating length for the subcooling liquid feed. All liquid from the last tray is passed through the reboiler, plus an amount of recirculated liquid from the reboiler outlet. Only the liquid from the last tray is subcooled, not the liquid from the reboiler, which is at boiling point. More liquid recycle should not increase the required area for heating the subcooled tray liquid as it does not increase the heating duty. One could even argue that more recirculation of reboiler outlet liquid should reduce the required area of this section due to increased velocity and increased heat transfer coefficient. 

Edited by PingPong, 31 March 2014 - 01:05 PM.

[katmar]

@PingPong - this was a long time ago and the controls were rudimentary. You are absolutely right that this was a fix and not a solution. In a production environment this is often how things go.

 

I have not done the calculations but I can envisage a possible situation where the increase in HTC due to the increased liquid velocity is less than the increase that would be obtained by having a higher fraction vaporized (which would induce more turbulence than the high liquid velocity).  As I say, I can't be sure of it but conceptually it is a possibility.  I also don't have the Lieberman book so all I have is Kister's reference to it. 

[aroon]

 

It was so oversized and generated so much vapor that that the pressure drop in the return line caused the circulation to stall.

Normally the reboiler duty, and therefor the vapor generation, is controlled by a TC a few trays higher on the column that affects the heating medium in some way. In such case an oversized reboiler is counteracted by reducing the flow of heating medium, or backflooding the exchanger with steam condensate, or ... (whatever). Apparently that was not the case with this reboiler so it could produce as much vapor as it liked?

 

PingPong,

 

I suspect that the vapor generation is only controlled by TC a few trays higher on the column. Reboiler outlet line is always mounted with TC, which operates the control valve located on the heating medium line to achieve the desired bottom temperature. Now the case where circulation rate is higher than design (with no butterfly valve), reboiler will try to achieve the reboiler outlet temperature with this high circulation rate, which subsequently will have higher vapor at same outlet temperature due to high circulation rate (as quality of the two phase will remain same at particular temperature). Now this additional vapors will have some/great impact on the column environment based on amount of duty added by the reboiler. Some times few trays of the column starts running dry, which is many times due to any abnormality in Reboiler operation.

 

[PingPong]

I have not done the calculations but I can envisage a possible situation where the increase in HTC due to the increased liquid velocity is less than the increase that would be obtained by having a higher fraction vaporized (which would induce more turbulence than the high liquid velocity).

Correct, but I was referring to the preheat part of the reboiler where the subcooled liquid is heated to its boiling point. It was in response to:

But it seems that if you get the design hopelessly wrong and the circulation is much too high then you can get into the situation described by Pilesar and seemingly confirmed by Lieberman.

I don't believe in the theory of more preheat area required for more circulation, for reasons I explained earlier.

 

In the mean time I found out what Lieberman was referring to. It was about a toluene reboiler which was retubed with high-flux tubing, while the 400psig heating steam was replaced by 100psig steam. After start-up the heat transfer rate was only one third of theoretical expected and the fix was to reduce the circulation across the high-flux tubes as it evidently interfered with nucleate boiling.

[katmar]

+1 for PingPong for digging up the reference. I am disappointed that Kister would refer to such a non-standard situation without pointing out the specifics. Lieberma's example, and the extreme over-design case I referred to above, are both real outliers and certainly should not be used as a justification for installing a valve in the circulation line on a routine basis.

 

Getting back to Rabindra's original question - unless the designer can present calculations that show the need for the valve I would not regard it as necessary.  On the other hand it will probably not do any harm if it is installed and left (locked) fully open, and removing it is probably not an option in terms of the contractors guarantees.

[xavio]

 

In addition, heating media is to maintain a temperature of column bottom as per process requirement and not to have desired circulation. Circulation rate only can be achieved by adjusting static heads and circulation rate is very vital to have stable column operation.

 

Let me also know your views on this.

 

- Aroon

 

In my opinion, regulating the flow of heating medium (steam) will have impact on circulation rate under the same static head. Less steam means less heat input, less vapor generated, and heavier density of the liquid/vapor mix; this will alter the whole thermal-hydraulic balance of the system. As bobby says, the system will self-regulate and move to new balance point.

 

 

That heating of the process fluid till its boiling point is peanuts compared to the total duty of the reboiler.

 

 

Not always I would say.

In low pressure applications, subcooling degree is not insignificant (as you can read from vapor pressure curve).

In very low pressure applications (vacuum), a small static head might be enough to cause a deep subcooling, which, in turn, results in very long heating zone (50% or more) and consumes large portion of heat duty (30% or more).

 

 

 Increasing the pressure drop through the reboiler inlet piping reduces the pressure at the inlet to the reboiler tubes. The result is that there is less tube length required to heat the process liquid and the boiling regime begins lower in the exchanger which can improve the overall heat transfer coefficient.

I think this is not a correct concept.

Putting a restriction (i.e., the butterfly valve) on inlet pipe will ADD pressure loss to the hydraulic loop.

Imagine closing a control valve in centrifugal pump outlet which causes flow to reduce.

The addition of obstacle in the thermosyphon loop will force the hydraulic balance to self-adjust to a new, lower circulation flow.

This lower circulation flow should translate to higher vaporization fraction and shorter heating zone as pointed by Pilesar.

 

I believe the valve is put as a corrective measure, either to accomodate very wide operating conditions or to compensate for any uncertainty during design/installation.

 

Thanks.

 

xavio

[paulhorth]

This discussion has turned out to be most interesting.

 

I have put in butterfly valves on the inlets to thermosyphon reboilers in several cases, with the intention of giving the operator the option of dealing with unstable circulation. In particular this can arise on turndown, as explained by Sachin. In a number of projects the client has specified a turndown to 25% of capacity as part of the basis of design. Given that the reboiler is always oversurfaced when clean, and more so if the client also specifies a capacity margin of say 10%, I considered that unstable surging could be a real problem and the valves were a simple remedy. But I accept the reasoning of Ping Pong and others that in practice the unit can work over a wide range of vaporisation.

 

Paul