9 replies to this topic

[aju_1807]

Hi All,

I have very interesting query here. The basis of compressor discharge relief valve sizing for blocked outlet case is really debatable and I am really not able to find any convincing method for sizing the relief valve.

1. Some companies (good companies) advise to size the relif valve for rated flow rate of compressor only (probably taking credit of certain controls).

2. Some companies (like ours) ask to size the relief valve to size for stone-wall conditions (probably the safest approach to do??)

Now, my query is when we are sizing relief valve based on stone wall conditions (assuming anti-surge won't open and it will move towards the stone-wall point and the suction pressure increases). Now, in this condition the following is a method we are using but I am not sure how correct this method is for sizing. Appreciate expert opinion on this issue.

a) we are calculating the pressure ratio from compressor performance for all operating cases and identify the one with the max pressure ratio at stone-wall condtiions.

b) fix the compressor discharge pressure at relieving pressure of relief valve

c) based on pressure ratio (which is fixed for machine) calculate the suction pressure considering dischare pressure as calculated in b)

d) Get the comp performance curve from vendor for that suction pressure got from c)

e) calculate the vol flow rate at the stone-wall condt for this curve.

This is the max flow which compressor can deliver. In case of variable speed, we have consider maximum rpm of compressor (without any limitation of compressor motor as the argument was in winter motor can deliver almost 30% more power.even SHELL DEP recommends not to take any credit of motor limit (power limit) to reduce the flow rate (i.e., below stone-wall point).

Is that approach correct or is there any other basis to calculate the same? I am really very confused as different company is following different method.

Can anybody throw more light on this subject and if possible send me one typical calcs on my email is (ajaysharma_2001"rediffmail.com)

regards,

ajay

[aju_1807]

Could anybody share more views on this issue?

Appreciate some expert opinion.

Regards,

Ajay

[aju_1807]

Could anybody share more views on this issue?

Appreciate some expert opinion.

Regards,

Ajay

SO FAR I HAVE NOT RECEIVED ANY FEEDBACK/OPINION ON THIS ISSUE.

APPRECIATE EXPERT VIE ON THIS ISSUE.

REGARDS,

AJU

[sheiko]

Could you please define what is the so-called "stone-wall" condition?

[aju_1807]

Could you please define what is the so-called "stone-wall" condition?

Its end of curve conditions (maximum flow which compressor can deliver).

Regards,

Ajay Sharma

[djack77494]

Is that approach correct or is there any other basis to calculate the same? I am really very confused as different company is following different method.

ajay,
I find your query to be a bit confusing and I'm not so sure of all your points. I am responding based more on hoping to help you avoid frustrations rather than providing a high quality response. Hopefully others can "fill in the holes".

If you are sizing a PSV for a blocked outlet case, I fail to understand why you are introducing stonewall flow conditions. That implies the maximum possible flowrate through the machine; your PSV should be focused on providing (at least) the minimum requirement of the machine. Your methods sound a bit convoluted to me, but I am not so familiar with doing calcs of this type. I'd see if some "go-bys" or experienced engineers could be found to confirm your methods (or talk to the compressor manufacturer). Try to identify a small number of cases that bracket the solution; if you can identify (say) two extremes, then you may need to do only two sets of calcs. I take no issue with the Shell DEP, and favor NOT considering any possible driver limitations. Sorry but I have no templates to send you for this, but you have my good luck wishes. (We used to say "that and 25 cents will get you a cup of coffee. Today change the 25 cents to $3.00 and you got it.)

[joerd]

In relief calculations, you can't rely on blanket statements to cover all scenarios. If you have a centrifugal compressor, and need to protect the discharge system, what you want to do is to evaluate all credible scenarios. A blocked discharge is one of them.
If the compressor can overpressure the discharge system, you will need to identify what combination of factors gives you the highest (sizing) relief flow rate. This may be limited by the compressor curve, the maximum suction pressure, the maximum available motor power, variable speed drive, startup and shutdown situations, possibly operating with a different gas. You should not take credit for any automatic controls, unless they are SIL-3 style instrumented protective systems, which is a whole other ballgame.
If the blocked-in scenario causes the suction pressure to rise to a maximum, evaluate the compressor performance at that maximum pressure, and at maximum speed. Assuming that the discharge relief valve popped, you know what the discharge pressure is, so you can find the flow rate that the compressor is processing from its curve. Then see if your driver is capable of providing the required power (unless you don't want to take that into account) and you'll get the relief load.
The point on the curve that you find may be anywhere, dependent on the required compressor head, so just taking stonewall is too simplistic, in my opinion.

[aju_1807]

In relief calculations, you can't rely on blanket statements to cover all scenarios. If you have a centrifugal compressor, and need to protect the discharge system, what you want to do is to evaluate all credible scenarios. A blocked discharge is one of them.
If the compressor can overpressure the discharge system, you will need to identify what combination of factors gives you the highest (sizing) relief flow rate. This may be limited by the compressor curve, the maximum suction pressure, the maximum available motor power, variable speed drive, startup and shutdown situations, possibly operating with a different gas. You should not take credit for any automatic controls, unless they are SIL-3 style instrumented protective systems, which is a whole other ballgame.
If the blocked-in scenario causes the suction pressure to rise to a maximum, evaluate the compressor performance at that maximum pressure, and at maximum speed. Assuming that the discharge relief valve popped, you know what the discharge pressure is, so you can find the flow rate that the compressor is processing from its curve. Then see if your driver is capable of providing the required power (unless you don't want to take that into account) and you'll get the relief load.
The point on the curve that you find may be anywhere, dependent on the required compressor head, so just taking stonewall is too simplistic, in my opinion.

Hi,

Thanks for your reply. In our case, the suction pressure may go up and hence the the compressor curve with higher suction pressure will shift further right and flow rate will increase substantially. Normally, I have seen many company uses design flow rate of compressor or stonewall condition flow rate with normal suction pressure. Howeverm here the catch is when the suction pressure increases the stone wall point in curve will shift further right (because of higher suction pressure).

I hope i am in right direction.

Regards,
AJU

[rxnarang]

Centrifugal compressor

a)The suction pressure increase should be an independent event, in no way related to the compressor getting blocked out. In most cases the suction pressure is controlled, and failure of this loop simultaneously with compressor discharge contingency is adjuged a double jeopardy. In this case the capacity of the relief valve should be the rated flow rate of the compressor, or if you want to be precise, the flow rate corresponding to the relief pressure. Obviously the latter flow rate should be less than the rated flow rate. This is the reasoning most companies have this practice.




b)Sometimes it is NOT an independent event, and is a consequence of a compressor getting blocked out. e.g. compressors in parallel drawing gas from a resevoir; if one compressor trips, the others get an higher suction pressure, and if a relief event occurs for the running compressors, then they could see a different curve. This different curve for a higher flow rate should be available to you, if you have specified the compressor conditions correctly. You should not be running back to the vendor, if an alternate running case for the compressors are designed for. In case you have not designed for the alternate case, then you are in bigger trouble. Once you have a curve available, then all you have to do is to repeat point a) above.

I would also put in more thought into deciding the set pressure of the discharge valve, and I assume you have.

Do let us know if you have any further queries.



Regards

Rajiv

[JoeWong]

Hi All,

I have very interesting query here. The basis of compressor discharge relief valve sizing for blocked outlet case is really debatable and I am really not able to find any convincing method for sizing the relief valve.

1. Some companies (good companies) advise to size the relif valve for rated flow rate of compressor only (probably taking credit of certain controls).

2. Some companies (like ours) ask to size the relief valve to size for stone-wall conditions (probably the safest approach to do??)

Now, my query is when we are sizing relief valve based on stone wall conditions (assuming anti-surge won't open and it will move towards the stone-wall point and the suction pressure increases). Now, in this condition the following is a method we are using but I am not sure how correct this method is for sizing. Appreciate expert opinion on this issue.

a) we are calculating the pressure ratio from compressor performance for all operating cases and identify the one with the max pressure ratio at stone-wall condtiions.

fix the compressor discharge pressure at relieving pressure of relief valve

c) based on pressure ratio (which is fixed for machine) calculate the suction pressure considering dischare pressure as calculated in

d) Get the comp performance curve from vendor for that suction pressure got from c)

e) calculate the vol flow rate at the stone-wall condt for this curve.

This is the max flow which compressor can deliver. In case of variable speed, we have consider maximum rpm of compressor (without any limitation of compressor motor as the argument was in winter motor can deliver almost 30% more power.even SHELL DEP recommends not to take any credit of motor limit (power limit) to reduce the flow rate (i.e., below stone-wall point).

Is that approach correct or is there any other basis to calculate the same? I am really very confused as different company is following different method.

Can anybody throw more light on this subject and if possible send me one typical calcs on my email is (ajaysharma_2001"rediffmail.com)

regards,

ajay

Ajay,
First you failed to advise us if the compressor is centrifugal, axial or reciprocating. Different type of compressor, the analysis of relieving load is pretty different. From the way you described, i guessed you are referring to centrifugal compressor. Flow may be limited by stone-wall condition in centrifugal compressor, however, this may or may not occurred in reciprocating compressor. Flow passing compressor possibly subject to compressor type.

Secondly, you are mixing the driver. Motor driven, gas turbine driven, steam turbine driven, etc. The driver plays a critical rule in defining maximum power that can be delivered to compressor to compressed gas/vapor passing the PSV. If the driver is motor, then it will be limited by motor rating. If the driver is gas turbine, then maximum power can be delivered by GT subject to ambient temperature, atmospheric pressure, air inlet and flue gas exit losses, driver rating, etc. A complete studies with the assistance of GT vendor is required. Simple use 30% is just a start but not a final...

Third, "1. Some companies (good companies) advise to size the relif valve for rated flow rate of compressor only (probably taking credit of certain controls)." This approach may be adopted during conceptual stage. However, during FEED or detailed design, a proper definition of system including compressor, driver, VSD, PSV setting, suction condition, etc shall be made available in order to conduct a detailed PSV load calc.

Just an experience to share...for a fixed speed centrifugal compressor, you may encounter maximum flow at stone-wall condition. However, for VFD centrifugal compressor, the maximum flow may not occur at the stone wall condition. All subject to fluid properties, maximum power available, suction condition, etc.