26 replies to this topic

[CSNK]

Dear All,


I want to design a PSV for the interstage drum of the two stage centrifugal compressor.
Whats the best practice to know the relief rate through the PSV.

I have calculated max flow taking into consideration that the suction valve of second stage gets close - but in that flow rate comes out to be too high then rated(normal flow).
One more is that in case the 2nd stage antisurge valve gets fully open - then there could be very high flow in the interstage drum.?

Now my doubt is that in case of antisurge valve failure - what is the method to know the max. relief rate.?

Thanks.

[paulhorth]

Chiks,
What is the design pressure of the interstage drum, and the design pressure of the discharge?
Choose the design pressure of the interstage system carefully, to help with this problem.
If you make the interstage drum have the same design pressure as the discharge, you have no relief case at all (other than fire). Alternately, you can design the interstage system for the settle-out pressure, assuming this is higher than the normal interstage pressure. Again, this could eliminate the relief cases.
Failure of the antisurge valve should not result in high pressure in the drum because the machine will be sucking out all the flow that goes in.

Paul

[CSNK]

Interstage design pressure is 7.03 Kg/Cm2.
Discharge Pressure is 4.91 kg/cm2.
PSV set pressure is also 7.03 kg/cm2 with 10% accumalation.
PSV is actually designed for blocked outlet case:-
But i am not able find where blocked outlet case will result into overpressure and what will be the relief rate.
Actually i am checking adequacy of this PSV(interstage drum).

Thanks.

[fallah]

Interstage design pressure is 7.03 Kg/Cm2.
Discharge Pressure is 4.91 kg/cm2.
PSV set pressure is also 7.03 kg/cm2 with 10% accumalation.
PSV is actually designed for blocked outlet case:-
But i am not able find where blocked outlet case will result into overpressure and what will be the relief rate.
Actually i am checking adequacy of this PSV(interstage drum).

Thanks.

chiks,

Please attach a sketch of the system including all valves for submitting better assistance.

Fallah

[CSNK]

Pls find the attached sketch.
PSV-A/B/C/D.

Attached Files

•  compressor sketch.xlsx   67.28KB   99 downloads

[fallah]

chiks,

And what about design pressure of discharge line and settle out pressure?

Fallah

[CSNK]

Design pressure of line is 7.7 kg/cm2
Compressor operates at suction pressure of 1.1 kg/cm2 and 1st stage discharge(second stage suction) pressure is 4.9 kg/cm2 and final (2nd) stage discharge pressure is 15.7 kg/cm2.

Edited by chiks, 13 October 2011 - 02:06 AM.

[fallah]

Design pressure of line is 7.7 kg/cm2
Compressor operates at suction pressure of 1.1 kg/cm2 and 1st stage discharge(second stage suction) pressure is 4.9 kg/cm2 and final (2nd) stage discharge pressure is 15.7 kg/cm2.

chiks,

I would suppose design pressure of the drum and that of 1st stage discharge line are 7.03 and 7.7 kg/cm2, respectively. Hence my suggestion:

For blocked outlet case (when the suction valve of second stage gets is closed) PSV relief would be maximum flow of 1st stage compressor (with discharge pressure of 7.03 kg/cm2) minus the maximum flow would be recycled through anti surge valve in full open position.

IMO, simultaneous occurrence of anti surge valve failure and blocked outlet is a double jeopardy scenarios and as per API 521 isn't a basis for design.

Fallah

[CSNK]

Thanks Fallah - i got good insight on the topic - because i calculated max. flow of compressor - but did not deducted(minus) the flow from antisurge.
Now my doubt is that i have antisurge valve data sheet - but max flow through FV is not given in data sheet - how can i find this?

[paulhorth]

Chiks,

These design pressures are very low, the second stage discharge design pressure is within the 150# flange rating. Why don't you make the design pressure of the interstage system also 15.7 kg/cm2g, the same as the 2nd stage discharge, as I said in my first post? This will cost nothing, because the piping will be the same (150#) and the vessel weight is not going to be higher going from 7.7 to 15.7 because it will be at its minimum plate thickness.
Doing this will eliminate the relief case on the interstage drum ( except for fire). This is a simple solution at no cost and saves work, requires a smaller PSV, and is safer.

If these design pressures have already been fixed then this is an example of poor process design for the reasons given above. If it's not too late, they should be changed, in fact the discharge design pressure should be raised to about 18- 19 barg, up to the limit of 150# flanges. This will cost nothing.

You also have not addressed the question of the settle out pressure.

As I have said in several previous posts on this subject, it is not general practice to take credit for the antisurge system in defining a compressor blocked outlet relief case. This was certainly the view taken by a major operator on a recent gas plant project i was involved in.

Please now will you consider my two posts and give a reply?

Paul

[CSNK]

Dear Paul,
Thanks for your reply.
Your opinion is absolutely right.
But this is the existing system and i am doing adequacy checking based on new throughput.
So for the same reason i would like to know the max relief rate as PSV is being already designed based on blocked outlet case.
Thanks.

Edited by chiks, 13 October 2011 - 04:43 AM.

[paulhorth]

Chiks,
New throughput, eh? How? - has a larger motor been fitted, or new impellers? Has the speed been increased? Any of these changes will increase the blocked outlet relief flow at the discharge.
Do you have performance curves for the new operating case?
You need to check the discharge PSV capacity as well as the interstage PSV.
My view is that the blocked outlet relief case, on the discharge, as well as the interstage, should NOT take account of the flow back through the antisurge valves.
This is the philosophy adopted by a number of leading operating companies and I have posted several times on this point.

And I still ask you about the settle out pressure - have you checked it?

Paul

Edited by paulhorth, 13 October 2011 - 11:12 AM.

[CSNK]

Dear Paul,
I take your point into consideration.
But when we say its blocked outlet and also surge valve failure - isn't it double jeopardy?
Then what will be the best practice to know the relief rate.

Compressor operates at suction pressure of 1.1 kg/cm2 and 1st stage discharge(second stage suction) pressure is 4.9 kg/cm2 and final (2nd) stage discharge pressure is 15.7 kg/cm2.


I have not calculated settle out pressure.

Thanks.

[paulhorth]

Chiks,
Best practice in the industry has long moved on from the simple concept of "double jeopardy" meaning the occurrence of two unrelated faults.
First, the recommended practice API RP 14C, used in the offshore industry since the 1970s, calls for two independent means of protection against overpressure, the basis being that there can be a malfunction of one protective device as well as the malfunction in the control system or operator error giving rise to the hazard. That is, two unrelated malfunctions, or double jeopardy.
Second, the modern requirement is to assess the probability of the hazard by an estimate of the SIL (Safety Integrity Level) requirement. This is (to simplify) an estimate of the probability of the hazard occurring due to successive failures taken together. For example (to simplify again) if the probability of the discharge valve closing is 0.1 per year and the probability of the antisurge system malfunctioning is 0.01 per year, the combined probability is 0.001 per year or 1 in 1000. This is the threshold criterion used by many companies as an acceptable risk - depending on the consequences. Therefore, the protective device ( e.g.a PSV) should be installed to deal with an event of this frequency or worse,(to reduce the risk to below 1 in 1000), and that can be a "double jeopardy" scenario.

But the case in question may not even require two faults. It is possible that a single fault in the compressor control system could result in both a closed discharge ESDV and also a failure of the antisurge valve to open.

I also repeat my questions about the new throughput which you mentioned:

has a larger motor been fitted, or new impellers? Has the speed been increased? Any of these changes will increase the blocked outlet relief flow at the discharge.
Do you have performance curves for the new operating case?
You need to check the discharge PSV capacity as well as the interstage PSV.

Paul

[CSNK]

Thanks paul once again for valuable input.
It means for knowing the max flow i should consider the closing of second stage suction valve and calculate the max relief rate.(without considering the opening of antisurge valve)

Yes for throughput - Speed of compressor has been increased.
I have performance curve for new operating case.
Interstage PSV is designed as i mentioned earlier and discharge PSV is designed based on final receiver pressure considering blocked outlet.(vapor leaving the final receiver to Absorber.)

Paul - i more question i would like to ask - as this thread has given me a good explanation
I have other compressor (two stage centrifugal-motor driven) in which there is PSV at the immediate discharge of 1st stage and 2nd stage.
Also there is PSV at interstage drum and
I just want to know why such kind of protection is given?
(This motor driven cent. compressor is added in parallel with the one we are discussing)
Thanks.

[paulhorth]

Chiks,
Thanks for the reply.
First, I should have added that reasoning from the philosophy of API RP 14C would also support my earlier argument about the anitsurge valve, without bringing SIL into it - thus:
API RP 14C calls for two independent means of protection against overpressure due to any fault.
The fault (blocked outlet) can arise for a number of reasons, including operator error, loss of instrument air, faulty ESDV, a trip signal from another system, etc.
For this compressor, the first means of protection is the antisurge system. A second independent means of protection must be provided, and that is the PSV. Hence the PSV is sized assuming the antisurge system has failed.

Now you have new performance curves, you need to check the new maximum flow through the machine at relieving conditions.

For your second compressor, can you send a diagram, please? I would assume that if this is in parallel with the existing machine, the design conditons will be the same? I would guess that there is a block valve between the first stage discharge and the interstage drum. This means that the blocked outlet relief has to be upstream of this block valve, while the drum itself needs a fire case PSV as well. Just delete the block valve, and one PSV can serve both cases.

Paul

[CSNK]

Existing comp. is steam driven
Parrallel comp. is motor driven (installed later for revised throughput).

Parallel comp runs at fix load while existing is variable (as per requirement).
Thanks.

Edited by chiks, 14 October 2011 - 11:22 PM.

[paulhorth]

Chiks,
Thank you for this diagram.
I do not know why the two compressors have different PSV provision. Now that you have supplied this diagram, it looks to me as if the existing compressor needs to have a PSV immediately at the discharge of each stage, just like the new compressor.
All of our previous posts have been about the blocked outlet case for the PSV on the interstage drum - but you have now shown two block valves upstream of there. How can this PSV protect against blocked outlet if the block is between the compressor and the PSV?
You would have to delete these block valves or lock them open - but one is an ESDV so it can't be locked open.

Is there a design pressure break at the block valve?

Paul

[CSNK]

Dear Paul
All this is existing system.
PSV's (1st and 2nd stage) are in vendor scope - still i am too confused why PSV's are not in the existing compressor?
And yes u r right that - Block valve is Locked open - but one is ESDV.
Design pressure of line remains same till the blocked valve and also till the interstage drum remains same - infact remains same till the suction of second stage.
PSV at 1st and 2nd stage discharge are set at 8.0 and 39.0 kg/cm2 respectively.

This PSV's (1st and 2nd stage immediate discharge) are being designed based on blocked outlet - which i assume that the ESDV malfunctions and gets close - where as compressor remains running ? - am i right.
If this is the case then i have to calculate the max. flow based on max. suction pressure as discussed earlier.?

[paulhorth]

Chiks,

This PSV's (1st and 2nd stage immediate discharge) are being designed based on blocked outlet - which i assume that the ESDV malfunctions and gets close - where as compressor remains running ? - am i right.

Yes, that's right.

f this is the case then i have to calculate the max. flow based on max. suction pressure as discussed earlier.?

Yes.

PSV at 1st and 2nd stage discharge are set at 8.0 and 39.0 kg/cm2 respectively.

I would expect from this that the design pressure of the discharge piping changes from 8.0 to 4.9, and from 39.0 to 15.7, at the downstream of the last block valve on the discharge of each stage - is this right?

Paul

[CSNK]

For information.
Design pressure of piping remains same till the second stage suction (7.7 kg/cm2)
And desing pressure of piping at second stage discharge is 16.7 kg/cm2.

[paulhorth]

Chiks,
There is something here that I don't understand. You have said

PSV at 1st and 2nd stage discharge are set at 8.0 and 39.0 kg/cm2 respectively.

And now you say

Design pressure of piping remains same till the second stage suction (7.7 kg/cm2)
And desing pressure of piping at second stage discharge is 16.7 kg/cm2.

How can the PSV set at 8.0 protect a system designed for 7.7? ( The 4.8 kg/cm2g in my earlier post was a mistake).
And, how can a PSV set at 39.0 protect a system designed for 15.7 kg/cm2g?
The PSV set points should not be higher than the design pressures.

Paul

Edited by paulhorth, 18 October 2011 - 02:47 PM.

[CSNK]

Dear Paul,
I have checked once again - but the pressure is the same.
Discharge line PSV's are located close to compressor (casing of the compressor is designed at 44 kg/cm2) and are in scope of vendor.
This PSV's are basically to protect Compressor.

[paulhorth]

Chiks,
It seems to me that there is some piping and valves in your system which can be overpressured. Please can you have a look at the attached sketch, which is for the second stage discharge. There is a similar problem with the first stage discharge.
Is my sketch correct?
What is the design pressure of the piping and valves between the compressor and the second block valve?

Paul

Attached Files

•  PSV and design pressure sketch.pdf   69.74KB   61 downloads

[CSNK]

Dear Paul,

Pressure is being marked in the file.
Your sketch is right.
From this view point it looks as if the pipeline pressure needs re-rated? AM i Right

As i mentioned earlier - this is the existing system.

Attached Files

•  PSV and design pressure sketch.pdf   213.05KB   43 downloads

Edited by chiks, 19 October 2011 - 06:50 AM.