18 replies to this topic

[sandstars]

Currently, we are in need of resizing our existing PSV discharge lines since the discharge methods employed are considered "unacceptable". What I've been asked to do is to consider several scenarios that all involve verifying we have the proper sizing of discharge piping for conventional spring operated valves. So, my first initial thought after digging around in API 520 was to assume the process fluids (in most cases, steam) are incompressible since the pressures and flowrates are pretty reasonable for a medium sized plant. From there, I am planning on using Bernoulli's modified equation that reflects pressure drop from fittings, elbows, piping roughness, etc and calculate the built-up backpressure to verify it is less than 10% of the rated relieving pressure.

THis is when I ran into problems. I'm probably complicating the solution more than I need to but API mentioned that the velocity at the discharge of the valve would increase until the sonic velocity is reached. Should I be using this instead as my built up back pressure? Any information about the proper way to size discharge piping (including header sizing for several PSV discharges combining) would be appriciated. I do have and will be using the rated discharge capacity of the valves as my flowrate.

[sheiko]

First you have to determine the sizing case; that is the scenario that leads to the largest PSV nozzle size. Could it be fire? blocked outlet? Control valve failure? total or partial utility failure? etc...The relieving fluid can be liquid, vapor or both depending on the relieving T & P.
By the way, relieving pressure = set pressure + overpressure = design pressure (or MAWP) + accumulation (see the definitions on API 520 & 521) and the criteria to size the outlet line of a PSV in based on a percentage of SET PRESSURE not relieving pressure.
Once you have sized and selected a standard PSV nozzle, then you know the rated discharge flowrate and can size the inlet and outlet line. If you consider that the % blowdown is usually 6% and that the % overpressure is usually 10% (except when the sizing scenario is fire and where ASME code is applicable), then you size the inlet line for frictionnal pressure drop less than 3% (half the blowdown) of the set pressure and the outlet line for frictionnal pressure drop less than 10% (the overpressure) of the set pressure.
Usually there is also a Mach number limitation on the discharge line, maximum Mach 0.7 as far as i remember.
Again, refer to API for the correct definitons and criteria.

[fallah]

I'm probably complicating the solution more than I need to but API mentioned that the velocity at the discharge of the valve would increase until the sonic velocity is reached. Should I be using this instead as my built up back pressure? Any information about the proper way to size discharge piping (including header sizing for several PSV discharges combining) would be appriciated. I do have and will be using the rated discharge capacity of the valves as my flowrate.

What you should consider in discharge line sizing is "the pressure just downstream of the PSV nozzle shall be equal/less than 10% of the set pressure" and the way of proper sizing,usually done based on back calculation from end of discharge line,would be available in API 521 and similar sources.

Depending on piping configuration,pipe size,reducer size,relif flowrate,....you may have several choked point along the discharge line.

[sheiko]

What you should consider in discharge line sizing is "the pressure just downstream of the PSV nozzle shall be equal/less than 10% of the set pressure" and the way of proper sizing,usually done based on back calculation from end of discharge line,would be available in API 521 and similar

Dear Fallah
The pressure you define is indeed the total backpressure, while API apply the sizing criteria to built-up backpresure for conventionnal PSV.
For the OP: total backpressure = built-up backpressure (frictionnal pressure drop on the line due to PSV discharge) + superimposed backpressure (pressure before PSV dicharges)

Edited by sheiko, 16 May 2010 - 03:22 PM.

[fallah]

Dear Fallah
The pressure you define is indeed the total backpressure, while API apply the sizing criteria to built-up backpresure for conventionnal PSV.
For the OP: total backpressure = built-up backpressure (frictionnal pressure drop on the line due to PSV discharge) + superimposed backpressure (pressure before PSV dicharge)

Dear Sheiko,

The pressure i mentioned,is the build-up back pressure.Superimposed back pressure is supposed being already considered in PSV spring adjustment.

[sheiko]

Dear Sheiko,

The pressure i mentioned,is the build-up back pressure.Superimposed back pressure is supposed being already considered in PSV spring adjustment.

I regret but the pressure at the PSV nozzle outlet flange is always the TOTAL backpressure when the PSV is dicharging.

[fallah]

I regret but the pressure at the PSV nozzle outlet flange is always the TOTAL backpressure when the PSV is dicharging.

You are right and as i mentioned i meant build-up back pressure at the PSV outlet flange shouldn't be more than 10% of PSV set pressure.

[sheiko]

fallah:

Alright it is clearer then.

[bernath]

Dear Fallah or sheiko or anyone else,

what we should consider for discharge line sizing is 10% rule and mach number (<0.7). I agree on that. My problem is, when the discharge fluid (from relief valve outlet) goes to common flare header, it's easy to apply above two rules. Yet, if the discharge fluid goes directly into atmospheric venting (which is very short line), then they will have their own problem. Pressure difference between relief valve outlet and vent pipe outlet would be too huge, this will lead into very high velocity in discharge piping (300-500ft/s). The built-up back pressure would also be quite high since the discharge line is very short. I'm kind of bit missing something here. please help guys...

thank you

[fallah]

kanankiri,

In general,short tailpipes terminated to atmosphere result in lower build up back pressure than long PSV outlet line.

If choked flow occurs at the outlet of a short tailpipe vented to atmosphere,high build up back pressure would be establshed.

Regards

[bernath]

Dear Fallah,

sorry if my question seems odd for you since I'm a newbie to relief system.

you said that, "In general,short tailpipes terminated to atmosphere result in lower build up back pressure than long PSV outlet line." Could you please explain in a more detail way? Is this something to do with built-up back pressure? When the K value and f*L/D value are high, the discharge line pressure drop will be great. We all know that the built-up back pressure = outlet vent pipe + discharge line pressure drop. Therefore, if the pressure drop along the discharge line is great, then the built-up back pressure would be low. Is my understanding correct? please advice


"If choked flow occurs at the outlet of a short tailpipe vented to atmosphere,high build up back pressure would be established."

So I guess if choked flow occurs, then atmospheric venting would not be a good choice for the discharge line? is this correct?

Is there any rule of thumb for determination of venting pipe sizing and length for atmospheric venting? The discharge line is only consist of small amount horizontal pipe, one long radius elbow as well as vertical pipe (for stack) and the optional weather cap as specified on API 520 partII.

thank you, br

[bernath]

Dear Fallah,

sorry if my question seems odd for you since I'm a newbie to relief system.

you said that, "In general,short tailpipes terminated to atmosphere result in lower build up back pressure than long PSV outlet line." Could you please explain in a more detail way? Is this something to do with built-up back pressure? When the K value and f*L/D value are high, the discharge line pressure drop will be great. We all know that the built-up back pressure = outlet vent pipe + discharge line pressure drop. Therefore, if the pressure drop along the discharge line is great, then the built-up back pressure would also be great. Is my understanding correct? please advice


"If choked flow occurs at the outlet of a short tailpipe vented to atmosphere,high build up back pressure would be established."

So I guess if choked flow occurs, then atmospheric venting would not be a good choice for the discharge line unless we choose to employ balanced bellow type or pilot operated relief valve. is this correct?

Is there any rule of thumb for determination of venting pipe sizing and (minimum) length (or height) for atmospheric venting? The discharge line is only consist of small amount horizontal pipe, one long radius elbow as well as vertical pipe (for stack) and the optional weather cap as specified on API 520 part II.

thank you, br

[fallah]

Dear Fallah,

sorry if my question seems odd for you since I'm a newbie to relief system.

you said that, "In general,short tailpipes terminated to atmosphere result in lower build up back pressure than long PSV outlet line." Could you please explain in a more detail way? Is this something to do with built-up back pressure? When the K value and f*L/D value are high, the discharge line pressure drop will be great. We all know that the built-up back pressure = outlet vent pipe + discharge line pressure drop. Therefore, if the pressure drop along the discharge line is great, then the built-up back pressure would be low. Is my understanding correct? please advice
Back pressure due to pressure drop in the PSV discharge line when PSV opens named as "Build up Back Pressure".Longer discharge line results in higher pressure drop leading to higher build up back pressure.
"If choked flow occurs at the outlet of a short tailpipe vented to atmosphere,high build up back pressure would be established."

So I guess if choked flow occurs, then atmospheric venting would not be a good choice for the discharge line? is this correct?
Selecting larger valve outlet size,or choosing bellows/pilot operated type may solve the problem
Is there any rule of thumb for determination of venting pipe sizing and length for atmospheric venting? The discharge line is only consist of small amount horizontal pipe, one long radius elbow as well as vertical pipe (for stack) and the optional weather cap as specified on API 520 partII.
It should be evaluated case by case.
thank you, br

[bernath]

thank you my friend fallah..

[bernath]

Dear All,

I still have something to discuss with you guys related to the above mentioned atmospheric discharge line. I have simulated a very simple atmospheric discharge line using Aspen Flarenet 2006 and I wanted to attach the '.fnw' file to the forum, but this forum doesn't allow me to attach '.fnw' file or '.rar' file. If anyone knows how to upload my simulation file, please kindly do let me know.

With regard to my flarenet simulation, the situation is like this,

I still have the problem with choked flow. The back pressure is okay, only the Mach number of the stack trouble me. My orifice area of 'T' letter, 8" of inlet and 10" on outlet. Both the tailpipe and stack diameter is 10". Choked Flow occurs at the end of stack. The rest is okay. I used balance bellow valve to accommodate for relative high back pressure.

This choked flow problem will be solved if I increase both the diameter of tailpipe & stack to 24", but then if I do this I will have to put the reducer to connect relief valve outlet (10") with tailpipe (24"). I believe this will lead to another choked flow at relief valve outlet.

My conclusion is I will go with 10" diameter of tailpipe and stack and ignore the occurrence of choked flow at the end of stack. Please advise if I'm wrong.

fyi, the mass flow is 160,000kg/h with set pressure of 15.87 psig. The fluid is saturated steam at 190 C degree.

Thanks in advance my fellow engineers.

regards,
bernath

Attached Files

•  pge pic.JPG   30.85KB   53 downloads

[kkala]

Having read this interesting thread, I would like to note following (even though having no expertize on the matter).
A. As informed by a friend, a preliminary sizing (or size check) of PSVs' discharge network can be:
1. Size the network for fire case, assuming superimposed back pressure = 0.3 kgf/cm2 g. The latter is assumed as operating pressure in the flare header. During fire all the PSVs of a "fire area" are assumed open, but rest PSVs of the factory are assumed idle (factory is divided in several "fire areas" according to criteria by NFPA). This can be repeated for every "fire area", considering one strong fire at a time.
Note: Fire may cause opening of PSVs designed to cover fire, but simultaneously of other PSVs, e.g. for blocked outlet.
2. Size the network for a general power failure case, assuming superimposed back pressure = 1.0 kgf/cm2 g. This because a lot of PSVs from all factory open during a general power failure.
3. Can fire at one "fire area" cause power failure in other areas? "Yes" is a conservative answer. When we check capacity adequacy of a new unit, we also consider the case of fire in this unit and at the same time general power failure in the factory.
4. For every part of network make a conservative size selection between (1) and (3).
5. Start now checking the size of resulting network in detail, for every possible scenario of exceeding flow, to trace any necessary further size increase.
For all above, rule of PSV discharge line delta P and of max mach=0.6-0.7 is applicable. Flarenet is usually applied.
B. Concerning Bernath's recent post, it is understood that pressure upstream flare tip is at least 14.7+1.3=16.0 psia, versus a PSV set pressure of 14.7+15.87=30.57 psig. I do not feel the issue physically; but increasing line sizes (diameters) may transfer the chocked flow at the PSV itself (upstream of reducer), which is not troublesome.
Not having enough background, a way out in this issue by others would be welcomed.

Edited by kkala, 25 March 2011 - 12:22 PM.

[bernath]

B. Concerning Bernath's recent post, it is understood that pressure upstream flare tip is at least 14.7+1.3=16.0 psia, versus a PSV set pressure of 14.7+15.87=30.57 psig. I do not feel the issue physically; but increasing line sizes (diameters) may transfer the chocked flow at the PSV itself (upstream of reducer), which is not troublesome.
Not having enough background, a way out in this issue by others would be welcomed.

Yes, please anyone who has the answer for the issue.

It's like transferring the choked flow from the tip to the relief valve outlet. What is the consequences of such act?

I also know that there's 3 mechanisms of choking, i.e. end-point, expansion and restriction choking. The pressure at relief valve outlet will undergo expansion choking while end point choking will occur at flare tip. Which one is most dangerous? Should I put the choke point at flare tip or at relief valve outlet? Which one is better? What is the consequences of each act?

hope someone can help

thank you
regards,
bernath

[paulhorth]

Bernath,

The problem with having choked flow at the flare tip (which means sonic velocity) is that unless the flare tip has been specially designed for sonic flow, the flame will lift off and you will have an unignited discharge, which is dangerous. This is the reason why it is general practice to limit the velocity at a pipe flare tip to about 0.5 to 0.7 Mach.(as well as to limit noise).

It is not uncommon to have choked flow at the PSV discharge flange. Immediately downstream of the discharge flange (even before any isolation valve) should be a pipe expander to fit the tailpipe, which should be sized for 0.7 Mach approx, often much larger than the discharge flange. If you have choked flow where the tailpipe joins the header then the tailpipe is too small. High velocity at this point can give rise to vibrations which can result in fracture of the joint into the header, I have seen photos where this has happened.

You need to get an expert to check the acoustic vibration energy in this tailpipe.

Paul

[bernath]

Bernath,

The problem with having choked flow at the flare tip (which means sonic velocity) is that unless the flare tip has been specially designed for sonic flow, the flame will lift off and you will have an unignited discharge, which is dangerous. This is the reason why it is general practice to limit the velocity at a pipe flare tip to about 0.5 to 0.7 Mach.(as well as to limit noise).

It is not uncommon to have choked flow at the PSV discharge flange. Immediately downstream of the discharge flange (even before any isolation valve) should be a pipe expander to fit the tailpipe, which should be sized for 0.7 Mach approx, often much larger than the discharge flange. If you have choked flow where the tailpipe joins the header then the tailpipe is too small. High velocity at this point can give rise to vibrations which can result in fracture of the joint into the header, I have seen photos where this has happened.

You need to get an expert to check the acoustic vibration energy in this tailpipe.

Paul

Hi Paul,

Thank you for such a nice answer. The way you describe your answer is simply wonderful.

bernath