14 replies to this topic

[tkeung]

Hello Everyone,

I am a junior level process engineer hoping for some advice on the following design scenario that I am facing.
For a gas plant, I am sizing the PSV for the inlet gas filter, governing case to be fire gas expansion.

My question is what is the proper way of calculating the relieving temperature?

I have read similar topics on this forum, but not able to completely understand the idea.
Should I used ideal gas law (Tr=T1*(Pr/P1)^(k-1/k))?
or should I perform a flash calc in Hysys by setting the relief pressure, and vapor fraction to 1?
Does this have anything to do with vessel wetted area?

Thank you.

[paulhorth]

tkeung,

If you require the temperature at relief pressure, and if there is no liquid in the filter, the ideal gas law will be fine, but it is not as you have written it. You only need the exponent k if there is an expansion or compression. If the expansion is isothermal then k =1. if it is adiabatic then k = Cp/Cv. But this is an increase in pressure at constant volume. No expansion or compression. So the law is:
p1.V/(z1.T1) = p2.V/(z2.T2) and the volume V cancels out.

To find Z, you will need to iterate. First assume z1 = z2 and they will cancel. When you have a first value for T2, check what z is at p2 and T2, then re-iterate using z1 and z2. This will probably give a slightly different T2.

To define the relief load and size the PSV you would need the rate of heat input. There is no wetted area. Refer to API RP 521 which gives an equation for fire relief from a dry vessel, which is based on the surface area.

Paul

Edited by paulhorth, 25 November 2011 - 08:21 AM.

[fallah]

'tkeung', on 24 Nov 2011 - 9:44 PM, said:

Hello Everyone,

I am a junior level process engineer hoping for some advice on the following design scenario that I am facing.
For a gas plant, I am sizing the PSV for the inlet gas filter, governing case to be fire gas expansion.

My question is what is the proper way of calculating the relieving temperature?

I have read similar topics on this forum, but not able to completely understand the idea.
Should I used ideal gas law (Tr=T1*(Pr/P1)^(k-1/k))?
or should I perform a flash calc in Hysys by setting the relief pressure, and vapor fraction to 1?
Does this have anything to do with vessel wetted area?

Thank you.

tkeung,

As per API 521 gas absolute temperature (T1), at the upstream relieving pressure (P1) in gas expansion case would be:

T1=(P1/Pn)*Tn

Where:

Pn is the normal operating gas absolute pressure, expressed in SI units (psi);

Tn is the normal operating gas absolute temperature, expressed in SI units (°R).

Fallah

Edited by fallah, 26 November 2011 - 02:04 AM.

[Lowflo]

tkeung - when you have a gas-filled vessel exposed to fire, the relieving temperature is a transient value rather than a fixed value. The gas temperature inside the vessel will continue to rise until the vessel is overstressed and fails. So, from a safety perspective you can pick any value you want for relieving temperature - regardless of the value that you choose, the PSV isn't going to protect the vessel from fire exposure. Go ahead an install a PSV for code compliance, but if you need effective protection look at additional layers of protection (depressuring, fire-resistant insulation, water spray).

Edited by Lowflo, 30 November 2011 - 02:16 PM.

[fallah]

'Lowflo', on 30 Nov 2011 - 3:35 PM, said:

So, from a safety perspective you can pick any value you want for relieving temperature - regardless of the value that you choose, the PSV isn't going to protect the vessel from fire exposure. Go ahead an install a PSV for code compliance

tkeung,
Above statements may be applicable for PSV fire case (no other applicabe scenarios available), but you mentioned that fire gas expansion is governing case in your system, means you would have the scenarios other than fire gas expansion and you should use API 521 (as i mentioned in my previous post) in order to minimum relieving temperature calculation which would be used in calculation of effective orifice area of the required PSV. This calculated area enable you to compare it with other orifice area values calculated for other scenarios and determine governing case based on most area value.

Now you mentioned you are in process of PSV sizing (preliminary) stage for fire gas expansion case (and you need relieving temperature for your calculation) and you previously know the governing case is this case without having relevant effective orifice area and comparing!

Please clarify this point.

Fallah

[paulhorth]

I agree 100% with Lowflo's reply.
It is possible, if the PSV set pressure is high enough above the operating pressure, that, in a fire, the vessel will fail before the PSV even opens.

Paul

[tkeung]

Hello All,

Thank you very much for your constructive comments!

[fallah]

'paulhorth', on 01 Dec 2011 - 2:16 PM, said:

I agree 100% with Lowflo's reply.
It is possible, if the PSV set pressure is high enough above the operating pressure, that, in a fire, the vessel will fail before the PSV even opens.

Paul

Above statement isn't a new point in pressure relieving field and no one can disagree with it, but the point seems is being forgotten in this thread is the post originator request on "proper way of calculating Relieving temperature for Fire Gas Expansion"!

[Lowflo]

Fallah - you missed the point. There is no "proper way to calculate relieving temperature for fire gas expansion". That's like asking what is the proper hardhat to wear when you're about to be run over by a truck. The answer is....it doesn't matter.

[fallah]

'Lowflo', on 02 Dec 2011 - 2:24 PM, said:

Fallah - you missed the point. There is no "proper way to calculate relieving temperature for fire gas expansion". That's like asking what is the proper hardhat to wear when you're about to be run over by a truck. The answer is....it doesn't matter.

Lowflo,

Now is the time someone ask API Committee to delete Sec 5.15.2.2.2 of API 521 (is attached) and implement your viewpoints instead!

Fallah

Attached Files

•  API_STD_521 Fire Gas Expansion.pdf   145.7KB   623 downloads

[Lowflo]

API 521 discusses this risk, and it has several curves showing just how quickly an unwetted vessel will fail, but it's not direct enough in allerting users to this risk. The biggest problem I have with that document is that it "invites" users to gloss over this problem by applying the vapor expansion formula.

I agree that this section should be revised.

[sheiko]

Hello all,
I believe Fallah's explanation is adequate from a pedagogical point of view and given the "junior" level of the OP.

Edited by sheiko, 20 December 2011 - 06:11 PM.

[chsnoopie]

'fallah', on 26 Nov 2011 - 07:10 AM, said:

'tkeung', on 24 Nov 2011 - 9:44 PM, said:

Hello Everyone,

I am a junior level process engineer hoping for some advice on the following design scenario that I am facing.
For a gas plant, I am sizing the PSV for the inlet gas filter, governing case to be fire gas expansion.

My question is what is the proper way of calculating the relieving temperature?

I have read similar topics on this forum, but not able to completely understand the idea.
Should I used ideal gas law (Tr=T1*(Pr/P1)^(k-1/k))?
or should I perform a flash calc in Hysys by setting the relief pressure, and vapor fraction to 1?
Does this have anything to do with vessel wetted area?

Thank you.

tkeung,

As per API 521 gas absolute temperature (T1), at the upstream relieving pressure (P1) in gas expansion case would be:

T1=(P1/Pn)*Tn


Where:
Pn is the normal operating gas absolute pressure, expressed in SI units (psi);
Tn is the normal operating gas absolute temperature, expressed in SI units (°R).

Fallah

Regarding this equation, why is it not possible to calculate T1 in Celcius unit if the Tn data input is also in Celcius unit?...since (P1/Pn) is just a ratio.....

[latexman]

Try it both ways, with ^oC and ^oK, and see if you get the same answer after converting the answers to common unit. Then, recall what you learned about "ideal gas law".

[fallah]

'chsnoopie', on 07 Jan 2013 - 1:35 PM, said:

'fallah', on 26 Nov 2011 - 07:10 AM, said:

tkeung,

As per API 521 gas absolute temperature (T1), at the upstream relieving pressure (P1) in gas expansion case would be:

T1=(P1/Pn)*Tn


Where:
Pn is the normal operating gas absolute pressure, expressed in SI units (psi);
Tn is the normal operating gas absolute temperature, expressed in SI units (°R).

Fallah

Regarding this equation, why is it not possible to calculate T1 in Celcius unit if the Tn data input is also in Celcius unit?...since (P1/Pn) is just a ratio.....

chsnoopie,

As latexman said; because the base of the equation is ideal gas law you, should use absolute temperature values.