9 replies to this topic

[prq123]

For relief valve sizing it is normally assumed that control valves not involved in causing the overpressure stay in their original position prior to the relief event (per API 521) .

In determine capacity credit for relief valve sizing (ie. credit for flows out of the system), API 521 states "No credit should be taken for any favorable instrument response.”

My question is if a control valve would tend to move to a more closed position during the relief event, is it acceptable to take any capacity credit for flow through this control valve? I have discussed this question with our company's safety engineers and their response is that it is company practice to take credit for the flow through this valve based on its final steady state position, even though the valve may MOMENTARILY reduce the flow out of the system to less than the final steady valve, perhaps even to zero flow. I was advised that this is normal industry practice.

My opinion is that this not in accordance with API 521. Based on the following logic:

1. For the case where the control valve used for capacity credit would tend to open during the relief event (ie. where the valve position is controlled by pressure control on the upstream portion of the flow path out of the system), we would, of course, not take credit for this valve opening more and letting more flow through the valve as this would be taking credit for a favorable (good) instrument response. We would conservatively assume that the valve was in a position which would relieve the minimum flow (ie. valve at its original position).

2. Conversely, for the case where the valve would tend to close during the relief event (ie. where the valve position is controlled by pressure control on the downstream portion of the relief path – back pressure controlled, see attached example valve V-3), we can only ensure there will be flow through this valve if there is a good (favorable) control response. A good (favorable) control response would prevent the valve from going fully (or nearly) closed. If the control response is poor, the valve may go fully or nearly closed during the relief event. However, taking credit for the good control response goes against API 521. Even with a properly tuned controller, for the case of a back pressure controller, the flow out of valve would need to be at sometime less than the initial normal flowrate (as well as final steady state flow) in order to re-establish the same initial pressure. Furthermore, the final steady position of the valve may be such that it is barely open and would be outside of its steady state control range (ie. less than 5% open), thereby increasing the likelihood the valve may fully close momentarily.

3. Hence to comply with API 521, my opinion is that credit should not be taken for valves that tend to close (ie. moved to a more closed position) during the relief event as good (favorable) control response is required to prevent the valve from going fully closed or reducing the flow to less than the final steady state flow.

I would like to get the opinion of others on this issue and with regards to what is normal industry practice.

Attached Files

•  CV_in_flow_path1.pdf   150.74KB   115 downloads

[JoeWong]

prq123,


Hi Welcome !

This will be an interesting discussion ! Before our experts come in, let me drop my 2-cents opinion.

Credit on V-3 would definitely not in picture as the PT is outside the loop and it can works normally by forcing V-3 to close. Do you agree ?

Valve-A and Valve-B have their own indepent controller feeding a common vessel (2) and relief via a common PSV. Please see the following scenario before approaching your problem.

Let say valve-A is feeding vessel-1 whilst valve-B feeding another vessel-2 and there are 2 indenpedent PSVs protecting vessel-1 & vessel-2 for control fulll open load. From flare capacity, we would consider valve-A failed open and PSV-1 relieving to flare OR valve-B failed open and PSV-2 relieving to flare. The flare capacity would be maximum of either PSV-1 or PSV-2. Do you agree ?

Now compare above example with your system. They have their own independent controller and control valve but with common receiver & PSV. I may consider when controller for valve-A failed lead to valve-A full open, the other controller controlling valve is working correctly to be inline with philosophy adopted in above example.

Personally i don't think you are taking CREDIT on second loop work in correct manner BUT consider failure of second loop together is DOUBLE JEOPARDIES.

However, i am pretty sure you have shutdown valves on these lines. Don't you ? I would continue my design with above and request the protection system to be IPF checked and ensure the available SIL is meeting SIL demand by the system.

Let see Phil and other opinions.

[prq123]

Joe,
The sketch I attached was only to show what I meant by a back pressure controller so that we are all clear on the issue. I would need to further explain my particular case if we want to talk about that.

Right now I'm more interested in the general philosophy of taking capacity credit for control valves (in particular, back pressure control valves) that tend to go "more closed" in a relief event.

I can provide further details later.

Thanks

[JoeWong]

I am a bit confused with your statement and intention. When you said "My question is if a control valve would tend to move to a more closed position during the relief event, is it acceptable to take any capacity credit for flow through this control valve?", is this referring to Valve-A and/or valve-B in your sketch ?

If valve-A / B, i would say NO credit to be taken. Bacause there is the controller failure giving "wrong" signal requesting valve-A / B FULL OPEN.

If valve-3, i would say NO credit to be taken again. Because the system downstream of V-3 valve is total independent of system upstream of V-3 valve (i guess without details).

If you talk about general philosophy, my experience is it can be anything favour to your position in a project at any time in any organisation. A typical example is a PSV is considered ultimate device and 100% reliable. 1 x 100% ONLINE PSV will be provided. However, in one of the project, there was a question raised on reliability on the PSV and demand for ONLINE spare PSV !!!

Don't you mind to share you details for further discussion ?

[prq123]

Joe, I am referring only to the action of the V-3 valve should overpressure occur in the vessel (ie. ovepressure due to upstream control valve failure). This V-3 valve would tend to close to re-establish the initial pressure downstream. For simplicity you can assume that there are no other control valves/shutdown valves/trips in the downstream system after the V-3 valve.

My question is whether it is in accordance with API 521 to take capacity credit for flow through valve V-3 during a relief event and also what is the general industry practice (if there is one). By capacity credit I mean taking credit for flow out of the system that would reduce the calculated PRV relief load .

As I mentioned previously, my opinion is that taking credit for any flow through a backpressure control valve (ie. valve V-3) during a relief event would not bein accordance with API 521. I assume you are also in agreement?

[JoeWong]

Joe, I am referring only to the action of the V-3 valve should overpressure occur in the vessel (ie. ovepressure due to upstream control valve failure). This V-3 valve would tend to close to re-establish the initial pressure downstream. For simplicity you can assume that there are no other control valves/shutdown valves/trips in the downstream system after the V-3 valve.

My question is whether it is in accordance with API 521 to take capacity credit for flow through valve V-3 during a relief event and also what is the general industry practice (if there is one). By capacity credit I mean taking credit for flow out of the system that would reduce the calculated PRV relief load .

As I mentioned previously, my opinion is that taking credit for any flow through a backpressure control valve (ie. valve V-3) during a relief event would not bein accordance with API 521. I assume you are also in agreement?

I tend to take conservative approach (if i can)...so i don't take any credit on opening of V-3. It is difficult to ensure the V-3 and it downstream system 100% reliable (without knowing the details)...

[pleckner]

For this particular valve, your V-3, you must not take credit for it, period. I don't know what "industry" your safety engineers are considering when saying it is normal industry practice to take credit for this type of scenario? Even a momentary closing of this valve during relief can send this vessel in orbit as the pressure will build very rapidly if the PSV cannot handle the flow it needs to.

The bigger problem I have is that you say this is company standard?!? Scary!!!

[JoeWong]

For this particular valve, your V-3, you must not take credit for it, period. I don't know what "industry" your safety engineers are considering when saying it is normal industry practice to take credit for this type of scenario? Even a momentary closing of this valve during relief can send this vessel in orbit as the pressure will build very rapidly if the PSV cannot handle the flow it needs to.

The bigger problem I have is that you say this is company standard?!? Scary!!!

Phil,
Again, i am happy my opinion is inline with yours.

prq123,
Your safety engineer said "I have discussed this question with our company's safety engineers and their response is that it is company practice to take credit for the flow through this valve based on its final steady state position, even though the valve may MOMENTARILY reduce the flow out of the system to less than the final steady valve, perhaps even to zero flow. I was advised that this is normal industry practice."

...company practice...
Please sincerely request some "references" from your safety engineer in order to make yourself confidence with it. There might be some exclusion or limitation in it where we may not aware of.

Personally i have strong feeling, the statement might has been written in the way leads to mis-interpretation. Please check it out.

...normal industry practice...
I would say NO. This is not industry practice, at least in Oil & Gas. Conservative approach is always the first approach. Credit will only be taken when there are very great incentive, technically justify, additional protection in place, availability & reliability of protection system is sufficient and the situation & consequences are properly analyzed, documented, deviation note approved by many levels, etc.

Last but not least, i would like to take this opportunity to praise you that you dare to analyze, investigate and challenge. You open yourself to more opinions...All engineer should have similar kind of attitude.

Good luck.

[prq123]

Phil / Joe, thanks for your responses. I assume that both of you are in agreement that the taking capacity credit for control valves that tend to go to a more closed position during a relief event is not in accordance with API 521?

I have asked our safety engineer to provide a reference from API 521 (or its interpetations) which supports the company practice of taking capacity credit based on the final steady state position of valves that tend to close on relief. No reference has so far been produced.

[djack77494]

I am certainly no expert in relief valves, but I have done a few calcs over the years. For a typical control valve that has NO role in a failure and relief scenario that you are considering, I would normally assume that that control valve might be in any position between fully open and fully closed. Taking the conservative approach (which you should always take without strong reasons for not doing so), I would assume unfavorable positions for any and all such valves.