Oversizing a PSV 3

[danielwbr]

Hi everybody,

I'm sizing a relief valve for a gas-liquid separator, which belongs to a petroleum production field. I was initially considering the maximum estimated flow for the field (something around 200000 Sm3/d of gas) as the valve's relief capacity, but one of my superiors said that it should be designed to relieve the maximum capacity of the vessel (~ 500000 Sm3/d) instead. Is this correct? In my readings through API RP 520/521 and some manufacturers' sizing guidelines I saw that PSVs should not be oversized, but it's not clear when an oversized valve would be inadequate (i.e. when the normal operating capacity of the system is less than 50% of the valve capacity).

Any help would be very appreciated.

 

[liberoSimulation]

If I recall well, over sizing of PSVs might not be allowed for high back pressure cases that need proper piping design and PSV type selection.

Hope that would help.

Regards

 

[jsummerfield]

Oversized relief valves may not provide adequate protection.
The PSV POPs open. If the volume of gas is not adequate to keep the valve open then it closes. When this happens the valve chatters, rapidly, with the potential to damage the valve, AND the capacity is reduced. This can be reduced with pilot operated relief valves, where suitable such as clean gas service.


John

 

[MortenA]

When sizing a PSV you will allways identify a number of relief case:

-Fire
-Blocked outlet
etc

You must the identify the case that requires the largest relief area and choose this.

The problems mentined in this tread are important but the valve should be able to handle the largest possible flow.

Best regards

Morten

 

[danielwbr]

First of all, thanks everyone.

MortenA, the design criteria choosen was blocked outlet. In this case, IMHO, the relief capacity would be the maximum predicted gas flow for the field (the vessel is the primary production separator).

Following this idea, I just don't know why the PSV should be designed for the maximum vessel capacity if the produced gas on the field will never reach that quantity. The produced gas' current flow is something around 150000 Sm3 per day.

 

[JimCasey]

Instead of ONE oversized valve, install multiple valves with the setpoint staggered. Size the first valve for the anticipated conditions and size the combination of the first and second to handle the maximum anticipated flow. So if the first valve has adequate capacity, the second valve never pops. ASME Sec VIII details the allowable parameters.

 

[EGT01]

Danielwbr,

As a general rule for valves in vapor service, as long as the required relief rate is greater than 25% of a relief valve's capacity, then the valve would be acceptably oversized. You can find supporting information in API RP-521 (4th edition, 1997), 'Section 4.3.2.2 Vapor Emissions' that suggests that spring loaded relief valves will stay open until the flow is reduced to about 25% of the valves capacity. I have also seen this comment in an older Farris relief valve catalog.

Recently, I found this website that has related information....

http://www.sache.org/workshop/2003Faculty/presentations.asp

In particular, look at this link....

http://www.sache.org/workshop/2003Faculty/PressureReliefDevicesScottOstrowski.ppt

Just based on your numbers of 150000 and 500000 you are close to the limit of being unacceptably oversized. Considering that the valve you get to handle the 500000 will have some greater capacity then this will make the situation even worse relative to your current needs.

If you have multiple relief cases to consider with widely varying rates, then multiple relief valve installation with staggered set points should be considered.

If you are having to make modifications to your equipment to install your relief system and you install relief capacity only for your current needs, you may want to consider making your equipment modifications so that you can more easily increase you relief capacity later if that is a concern for your superiors.

 

[MortenA]

Daniel

If you question your design basis then you must bring this forward to the right persons (not this forum ).

If you dont: Then you must select the design flow rate regardsless of current flow!

I dont like the "staggered" that mush - as it may give you other operational problems (the first valve must have a lower SP then the other - wich may be too low (at leat when considering the high flow case)?

A pilot valve will usually not have the same problems with regards to oversizing - however many companies will not accept pilot operated PSV - especially not in "dirty" service.

I think that the major problem with regards to oversizing is that API520 requires the up and downstream piping to be designed for the rated capacity - not you actual capacity. This could lead to a very large discharge system.

A normal spring operated valve has a "blowdown pressure". This means that the risk of reseating is not that big. But if your upsteam piping is undersized it may cause a high volume flow initially that may cause a high pressure drop in this pipeing - that will lead to re-seating of the valve.

Best regards

Morten

 

[danielwbr]

Thanks again for the answers...

EGT01, I've found the section of the API RP 521 that you mentioned. That is the kind of information that I need. As I have little or no field experience at all, I think that is important to have some background information (although it's only a RP) before starting to say things...

In the end, my superior was right (as if I didn't know it... hehehe). He said that that the percentage we're talking about would be 30%. That's + or - the value that can be found on some sources, including the one that you stated.

"Just based on your numbers of 150000 and 500000 you are close to the limit of being unacceptably oversized. Considering that the valve you get to handle the 500000 will have some greater capacity then this will make the situation even worse relative to your current needs."

That's exactly my point - and it's even worse considering that 150000 is the CURRENT production, and this quantity only tends to decrease with time. Anyway, I'll be writing a report evaluating the field's current relief system, so in this special case i'll put a note warning about the oversizing problem. It's up to them to decide what to do, I'll just give the choices.

MortenA, the design basis for the every PSV in this field is the blocked outlet case. It's a common practice in this company (I'm from a contracted one) to do it this way. My contractor manager doesn't show too much confidence on my decisions so i try to get as much as information before making an statement (and he's right, i don't have field experience). This question is an example of what i'm talking about. I was designing the PSV for the maximum field production, and they said no, it should be designed for the vessel's maximum capacity. Initially i've thought that i was right, so before saying something, I've tried to collect the necessary information to cover my opinion, reading the API RP 520/521 from the top to the end (minus the section 4.3.2.2

), manufacturers handbooks, googleing and asking other people, including you here.

I also know that this should be done by someone experienced, but as we don't have such one... i'm making my decisions based mostly on theorethical data and recommended practices, that's why i'm being extremely careful on doing it.

Daniel

 

[MortenA]

Re-read your original post:

I think your own thoughts are correct. Size the valve for the design flow rate and choose the actual size accordingly (closes larger std. orifice size).

The you must check upstream and downstream piping according the the actual max. capacity of this valve.

Best Regards

Morten

 

[TD2K]

One option is to install a pilot operated modulating PSV. It opens only as much to maintain the inlet pressure at set-point (actually, a slight amount over set-point but less than the 10% allowed by code). Chatter then isn't really an issue though your inlet piping and outlet piping still needs to be checked for capacity at the PSV's rated capacity.

 

[MortenA]

I agree with TD2K - many compaies will however not use pilot operated valves in "firty gas service" and some will not use them at all.

Best regards

Morten

 

[danielwbr]

Thank you all, people... your answers were very helpful.

danielwbr