17 replies to this topic

[sharonkuo]

Hi,

Usually for Closed Outlet scenario, we would assume a normal flow rate which has sufficient pressure to cause relief as our required relief rate. However, I have a question about the following scheme. (drawing attached)

 untitled.bmp   576.05KB   63 downloads

The Inlet line has a pressure controller, set pressure at 545 psig.
The pressure into the PCV is 800 psig.
The PSV set pressure of the equipment is 600 psig.

Upon evaluation of Closed Outlet, could we make the assumption that the pressure will built up downstream of PCV until 545 psig and the PCV will then respond in close; thus, no relief.

Or otherwise as follow:

Is there anything specified in the Code(ASME or API) that we would have to assume a normal opening position of the PCV in determining the relief load regardless of favorable intrument response? if this is the case, then Closed Outlet scenario will be considered even if upstream PCV regulate at 545 psig.


Any Comment will be appreciated.

[latexman]

It really depends on the definition of a credible scenario you are using. Some companies use simplified definitions of credible scenario, and you have to look carefully at announced and unannounced events/failures. Then, you have to worry about pre-existing conditions, like a leaking control valve. Others would do a risk analysis. It depends on the consequence - is it a big deal or a small problem. Bottom line, consult your companies/clients definition of a credible scenario.

[sheiko]

Hi,

Usually for Closed Outlet scenario, we would assume a normal flow rate which has sufficient pressure to cause relief as our required relief rate. However, I have a question about the following scheme. (drawing attached)

I would consider maximum operating flowrate.

Upon evaluation of Closed Outlet, could we make the assumption that the pressure will built up downstream of PCV until 545 psig and the PCV will then respond in close; thus, no relief.

I think you should consider the PCV in its normal position. Refer to your project specifications or API STD 521.

Or otherwise as follow:

Is there anything specified in the Code(ASME or API) that we would have to assume a normal opening position of the PCV in determining the relief load regardless of favorable intrument response? if this is the case, then Closed Outlet scenario will be considered even if upstream PCV regulate at 545 psig.

Refer to API STD 521. Good luck.

[sharonkuo]

Thank you very much, Latexman & Sheiko.


The process trend actually shows the PCV will respond in closing when downstream pressure increases.

If , indeed, I use the PCV normal opening position to calculate the required relief load, does it not assume double jeopardies? (Closed outlet and instrument failure/PCV leakage)?

[latexman]

If , indeed, I use the PCV normal opening position to calculate the required relief load, does it not assume double jeopardies? (Closed outlet and instrument failure/PCV leakage)?

Is the closed outlet announced/alarmed?
Is the instrument failure/PCV leakage announced/alarmed?

[sharonkuo]

Do you mean a physical alarm going off when the events occur?

I would say no alarms, there is only the signal of the instrument sending into the DCS. Only if the operators sees the outputs on the screen.

If it is announced, then we could assume Closed Outlet and PCV failure will not happen at the same time??

[fallah]

Thank you very much, Latexman & Sheiko.


The process trend actually shows the PCV will respond in closing when downstream pressure increases.

If , indeed, I use the PCV normal opening position to calculate the required relief load, does it not assume double jeopardies? (Closed outlet and instrument failure/PCV leakage)?

I think this would be double jeopardies and in case of BO because PCV will tend to close when downstream pressure raised,minimum flowrate through PCV for calculating relief load of PSV on vessel in downstream would be adequate.

Anyway,the equipment upstream of PCV should be evaluated for overpressure protection due to PCV closing.

[latexman]

Do you mean a physical alarm going off when the events occur?

Yes, a physical alarm.

If it is announced, then we could assume Closed Outlet and PCV failure will not happen at the same time??

As I said before, it really depends on the definition of a credible scenario you (your company/client) are using. My company defines that one unannounced and one announced independent events/failures IS a credible scenario, and two announced independent events/failures IS NOT a credible scenario, but read your own policy.

[sheiko]

Do you mean a physical alarm going off when the events occur?

I would say no alarms, there is only the signal of the instrument sending into the DCS. Only if the operators sees the outputs on the screen.

If it is announced, then we could assume Closed Outlet and PCV failure will not happen at the same time??

I think you should clarify this with your supervisor or the client but, to me, it is not double jeopardy to consider the PCV in its normal operating position (if normally open).

[sharonkuo]

Okay. Thank you very much Latexman, Sheiko, and fallah.

The policy of our company is not very specific of a credible scenario(alarmed/announced), so I wanted to see what is usually assumed for this scheme. Thank you!!

[Lowflo]

Sharonkuo,

Here's a suggestion for clearing away the fog. Add more detail to you scenario description. Right now you have an ambiguously defined scenario, and that ambiguity is the cause of this confusion. If you flesh out the scenario, I think the answer will be apparent to you.

Does the closed outlet situation occur while the process is running normally, during S/U, S/D, clean out, or some other mode of operation? A non-credible scenario during normal operations can be very credible during some other mode of operation, and your case is a good illustration of that.

Let's start by looking at your closed outlet scenario from the perspective of normal operations. The PCV is functioning normally.....it's controlling the pressure in the drum. Then someone mistakenly closed the outlet valve (or some failure occurs downstream creating the same effect). The PCV is functioning fine during normal operations, so it responds to the sudden rise in pressure by going shut. No overpressure occurs. To say that the PCV fails at that same time, that would be double jeopardy.

Now, let's think through what might go wrong during a start-up. Let's say the outlet valve was blocked while the system was down. The board operator, thinking that everything is ready to go, puts the PCV in manual and opens the valve to some fixed position. Or, say the board operator intends to open some other valve and opens this PCV by mistake. To me, this is a very credible scenario and one that needs to be protected against.

The bottom line is that you need to add more detail to your scenarios. Think through what can go wrong, and not just during normal operations. Relief designers often make the mistake of limiting their thinking to normal operating conditions, because plants normally ARE running. But, overpressure incidents are far more common during other modes of operation, especially during plant start-ups.

[sharonkuo]

Thank you Lowflo,

I usually analyze the scenarios based on normal operations only due to many uncertainties and abnormal conditions at start-up/shut-down; seems a lot of assumptions had to be made. As you pointed out, overpressure incidents are likely to occur more frequently at start-up/shut-down mode. I should have read the start-up/shut-down procedures first before I evaluate the PSV. I will keep this in mind and describe my scenario more thoroughly.

Sorry, I am new to PSV field, so I want my thinking to be clear before evaluating more valves. Thanks a lot!!

[Lowflo]

sharonkou,

No need to read start-up and shut-down procedures. The relief designer is the devil's advocate....you want to think through all the things that can go wrong, not right. In this task you don't care how things are supposed to work. That's not relevant. Just look at the P&id's and lay out the possible scenarios in which things can get screwed up. Then evaluate whether or not those scenarios are credible. Based on what you've said, you need to get an experienced person to help you with this.

[latexman]

Use "What If". What if someone closed the manual valve? What if the PCV stuck wide open? What if someone closed the manual valve and the control room operator opened the PCV instead of closing it when pressure got too high?

[sharonkuo]

Thank you Lowflo & Latexman,

I was not sure if it is alright to assume the manual valve is closed and control valve being open at the same time regardless of a normal response. I get it now. One should assume a worst case possible and it also depends on the company's definition of a credible scenario. Engineering judgment, I think.

Thank you again for your patience and time in answering my questions.

[arvind]

Thank you very much, Latexman & Sheiko.


The process trend actually shows the PCV will respond in closing when downstream pressure increases.

If , indeed, I use the PCV normal opening position to calculate the required relief load, does it not assume double jeopardies? (Closed outlet and instrument failure/PCV leakage)?

I think this would be double jeopardies and in case of BO because PCV will tend to close when downstream pressure raised,minimum flowrate through PCV for calculating relief load of PSV on vessel in downstream would be adequate.

Anyway,the equipment upstream of PCV should be evaluated for overpressure protection due to PCV closing.

Hi Fallah,

API 521 (4th edition) section 3.10.2 Capacity credit: recommended to be consider normal flow through the instrument in evaluating the relieving requirement.

[fallah]

Hi Fallah,
API 521 (4th edition) section 3.10.2 Capacity credit: recommended to be consider normal flow through the instrument in evaluating the relieving requirement.

Please note to below fact from API 521 (5th edition) a little bit different with previous edition regarding the flow ("minimum" word to be added):

"5.10.2 Capacity credit
In evaluating relieving requirements due to any cause, any automatic control valves that are not under
consideration as causing a relieving requirement and that would tend to relieve the system should be
assumed to remain in the position required for minimum normal processing flow. In other words, no credit
should be taken for any favourable instrument response. Minimum normal valve position is the expected
position of the valve prior to the upset incident, that is, the position of the valve when at minimum design flow
rates (unit turndown conditions)"

Edited by fallah, 08 September 2009 - 02:11 AM.

[skearse]

My take, based on the information given, is that there are two potential scenarios here, with only one that I would consider a credible scenario leading to an overpressure in the vessel. I would discount the blocked outlet scenario, because if the outlet is blocked, you assume that that the PCV would continue to hold and regulate the pressure below the PSV (and presumably the vessel MAWP) of 600 psi. However, wide open failure of the PCV would be a credible scenario, with credit only taken for alternate relief paths as noted by fallah. If this is aa actuated control valve (ie, not a simple diaphragm regulator), with actuated control valves on the outlets, the case of power failure (or instrument air, etc.) should also be considered, based on the fail position of the valves.

Edited by skearse, 09 September 2009 - 08:49 AM.