Does a large PSV adjust its opening to flow only the inlet feed to the vessel?

[Iradah]

Hello,

I was wondering how the PSV behaves when it reaches the set point and opens. for a PSV sized for a blocked outlet relief load, it will be sized for the flow rate into the vessel. After some years, the production might be less and the required relief load will be less, however, the PSV still has the same size required for the initial larger relief load. if at low feed to the vessel, the outlet gets blocked and the pressure reaches the PSV set point, the PSV will open. what will be the flow rate from the PSV? is it the original designed flow rate? in this case the pressure will drop and the PSV will close and then opens when the pressure builds up again. Or will the PSV adjusts its opening to relief a flow equal to the inlet feed to the vessel?

I am carrying out a study on the flare header and it is showing that there is an issue at the original designed relief flow rates. The current production rate is less than day one, so does this mean that the relief flows for blocked outlet will be also less which will solve the issue on the flare header?

 

[Latexman]

The PSV is still the same size. It "pops" open at the same set pressure. It has the same flow area. Therefore, it will have the same "rated" flow, which is usually quoted at 10% overpressure. It closes at the same re-seating pressure. The simple model of a PSV is - it is either wide open or closed.

If the volume of the protected system is the same and the flow into the system has decreased, then the PSV opens and closes less frequently.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[Iradah]

Thanks Latexman,
This is what I think. At original inlet rates to the vessel and at blocked outlet case, the PSV will be continuously flowing the designed flow rate for its size. at lower inlet to the vessel than the original design, the PSV will be still flowing the same rate it was designed for but in cycles as it will be opening and closing. This means that the flare header will be exposed to the same flow rate but in cycles which indicates that the risk now is higher than before because of the cyclic behavior.
Can the type of PSV (conventional, balanced or pilot) have an effect, or the PSV will be always either fully open or closed?(the fact that the PSV might not seat later at pressures lower than opening setting is not part of this discussion)

 

[e43u8]

Iradah,

Only a modulating pilot operated PSV can do the job, i.e. can adjust its opening to momentarily flow; then there would be no cyclic behavior in PSV relieving toward flare network...

 

[don1980]

The standard procedure for modeling pop valves to a flare header is to use the rated flowrate for the tailpipe and the required flowrate once the stream enters the flare header (ref API 521 6th ed., 5.4). So, you can update the relief documentation to show the lower required flowrate and use that value in the flare model.

 

[Iradah]

Although the required flare volume is less, the PSV will still flow the initial flow, so how can I update the document for less flow?

 

[don1980]

Review the definitions of rated flowrate and required florate.

When evaluating a flare header, use the required flowrate. For tailpipes, one uses the rated capacity of the valve, and that's true regardless of whether the tailpipe discharges to atm or a flare header. If you're using a software program (e.g. Flarenet) for modeling the flare system, the program will automatically switch from rated flowrate to required flowrate once the stream exits the tailpipe and enters the flare header.

For this specific problem, it sounded like you were saying that the required flowrate is lower than the original valve used in the PSV design. If so, then you can update the PSV documentation (update the required flowrate) and use that lower value in the flare header rating. Alternatively, you can proceed with using the current required flowrate value in the flare rating. The flare load may turn out to be OK based on this value, even though you know that it's higher than the real value.

 

[e43u8]

Iradah,

You mentioned in your first post:"... there is an issue at the original designed relief flow rates" for which the main point is lack of adequate info and detail. Can you clarify a bit more about mentioned issue?

 

[Iradah]

OK, I will try to clarify it more.
The flare header is already sized and installed. It was sized for a certain required load. the Flare header seems to be incorrectly sized as it shows locations of choked flow.
some people mentioned that there should be no issue now because the required relief load is now less due to less plant production. However, I think that this is not accurate. the PSVs are still of the same size installed for the large required relief load and they are still set at the same pressure. so I think that although the required relief load is less, the actual load will be high because the PSV will be going on fully open position and flow its rated flow rate. but since the input to the vessel is low, the pressure will drop and the PSV will close. and the cycle will be repeated. there is someone else who argued and said that the PSV will although rated for larger flow rate it will flow only the required flow and I think that this is incorrect. Clarifying this point is very important because it changes the whole case from high risk classification into safe one.

http://www.eng-tips.com/viewthread.cfm?qid=365370

 

[e43u8]

Iradah,

The PSV with lesser incoming load than before due to less production rate might be opened momentarily at the set point, relieving the rated load and reseat back, i.e. sort of unstable operation due to lower actual generated load. But it might even, with high degree of probability, not attain a full lift to relieve the rated load because there isn't the sustained generation of previous higher load. Then i think the PSV can relieve a load mostly not higher than required flow rate leading to no excess load on flare header to be a source of concern...

 

[Iradah]

Thanks e43u8 and all for your inputs. This dicussion was useful.

 

[MortenA]

I have been following this tread with interest but im not sure that i agree with the general consensus that seem to be here: The valve opens fully and releases design capacity and then closes fully in a cycle if the require relief rate is less than the design.

But re-reading API 520 (8. ed) im now i doubt if my initiall position is wrong. API 520 section 4.1.1.7 reads: Flow is restricted by the opening between the nozzle and the disc until the disc has been lidted from the nozzle seat approximately one quarter of the nozzle diameter. After the disc haas attained this degree of lift flow is then controlled by the bore area rather than by the area between the seating surfaces."

So if i read this correctly flow will only be dependant on upstream pressure and orifice area for lift above 1/4 of the nozzle diameter. This together with figure 6 supports the statements that the valve will effectively relive close to design capacity (slightly lower because the pressure imo will not reach 110% - but design value may be used because its impossible to say exactly what the actual relieving pressure will be). Then inaccordance with figure 6 pressure will decrease below set-p - and the valve will reseat fully. If the conditions that lead to overpressure is still there - the pressure will once again increaes ntil the PSV opens. etc.

Best regards, Morten

 

[Iradah]

Thanks Morten for your valuable input.

 

[LittleInch]

An interesting discussion and as ever I think "it depends" on the process being relieved and particular impacts of the system. In an initial relieving event the valve will clearly "pop" open with full relief capacity as per original design. However the issue then is how does the valve respond to a slowly falling pressure if, for instance, the blocked flow is now 50% of the original design flow. Pressure will reduce, but only slowly and some flow relief is required.

It could well be that as the valve disc approaches the seat, that so long as the flow is sufficient to hold open the disc that the valve in effect becomes a throttling device on the basis that the flow is no longer enough to hold open the valve at its 100% opening.

I can only suggest you ask the vendor what the valve characteristics are in this respect or even do a few tests yourself or off line to find out how the valves respond to high pressure, but lower flow.

In some parts of the plant I imagine that some valves will now operate in a open / close arrangement, especially where there is is only a small volume to relieve, but where a larger one exists which can even out the open / close popping, then below a certain flow rate as a percent of its rated duty (maybe 25-30%), the relief valves could easily sit partly open.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[e43u8]

Considering the fact that a PSV is designed to pop up at the set pressure and to be in stable full lift position at the specified over pressure and the relevant rated flow of governing case, and reseats at the value of blowdown below the set pressure if the cause of the over pressure would be removed; it can be said that in the cases other than the sizing case with lower relief load it might not to attain in full lift position even after pop up at the set pressure due to lesser incoming flow and probably the valve would reseat before attaining in full lift position and after a while again pop up and repeat the cycle. But appears if the flow in non governing scenarios not to be so low, let say higher than 25% of the rated flow, even though the valve will not go to be in full lift position after pop up, it will reseat and again pop up in a low frequency cycle not to be called it chattering. On the other hand, if the flow in a scenario would be below, let say, 25% of the rated flow; the frequency of pop up-reseat cycle would be so high can lead the PSV to chattering situation...

 

[MortenA]

@e4u8: True, chattering suggests that a "low flow" case do in fact exists. But chattering in the dimentioning case would normally be considered a design error. In fact i believe that the case where chattering is normally considered is the case where the lead pipe is underdimentioned so that when flow builds up the pressure drop in the lead pipe becomes to large that the PSV re-seats. For this calculation the rated capacity sall be used. Yet another indication that the rated caacity is the correct way when designing - and then assuming that the valve will open and close.

 

[e43u8]

@MortenA,

As you well aware PSV chattering might be created due to several rerasons mainly included: 1)Undersized inlet line, 2)Undersized outlet line, 3)Oversized PSV itself. In current discussion the chattering due to third reason are being considered i.e. a PSV which has rightly been sized per governing case would be oversized for other cases. But the point is that the PSV is how much undersized for the cases other than sizing case; and as i mentioned if, for any case, degree of oversizing wouldn't be so high the PSV might be under cyclic open-close conditions but with a time lag between opening and closing points not being called chattering; and on the other hand if degree of oversizing would be high, say for lower than 25% of the rated flow, rapid opening-closing of the PSV will lead to chattering and subsequent PSV damage...

 

[MortenA]

@e4u8,

That fact that you will have other cases that the design case cant be avoided! So of course you can find cases where the PSV will be (much) too large - but the valve still have to be able to cope with the design case. Thats just the way it is. I think this whole tread considers the design case - and any required oversizing and impact on header and falre capacity.

Howvere, as also other pointe out much earlier there are valve type sch as modulating pilot types tha can cope with this dilemma.

 

[Duwe6]

Or go with two PRV's. One sized for the smaller, most likely case. The second sized for the 'design case' that is the biggest [minus the contribution of the smaller one]. Set the second one at least 10% higher than the first.

 

[MortenA]

There are rules for "stacked" psvs too and the max set-point of additional valves is 5% over the SP (10% for the fire scenario) for the first valve according to API 520. They also require more gear and maintenance and becmes more difficult to maintain wo plant shutdown (if you relieve to a flare header anyway)