7 replies to this topic

I have a question, I have to calculate a relief valve size for my compressor suction drum, which has the fire case. The relief pressure is higher than the critical pressure of the fluid, is there any procedure to calculate the relief temperature?

[Art Montemayor]

What is it that you exactly have?

You infer you are protecting a compressor suction drum from a pool fire. Compressors compress gas, not liquids. By the title of the thread, are you stating that you are compressing a supercritical fluid? Which one, and what are its conditions?

What kind of compressor is this? Is its MAWP higher than the critical pressure of the fluid (gas?)? Kindly identify the fluid and all the rest of the basic data and don’t leave us in suspense.

[umeshr]

Please find the attached link,
http://people.clarks...gn/reliefv2.pdf

This gives a procedure to do the RV sizing for super critical condition. You will require hysys or other simulation software to do the property calculations. At relieving condition the fluid will be in dense phase & the orifice size calculated will be lower (may be D or E orifice).

We have came across similar situation many times, at normal condition the gas in KOD will be below critical condition but at relieving condition (ie. 21% over pressure {for ASME section 8 vessel} & corresponding relief temperautre ) the fluid will be above critical condition.

regards
Umesh.R

Thanks Umesh,
It was useful and exactly what I was looking for.

[cmp74]

Please find the attached link,
http://people.clarks...gn/reliefv2.pdf

This gives a procedure to do the RV sizing for super critical condition. You will require hysys or other simulation software to do the property calculations. At relieving condition the fluid will be in dense phase & the orifice size calculated will be lower (may be D or E orifice).

We have came across similar situation many times, at normal condition the gas in KOD will be below critical condition but at relieving condition (ie. 21% over pressure {for ASME section 8 vessel} & corresponding relief temperautre ) the fluid will be above critical condition.

regards
Umesh.R

Hello Umesh,

i am trying to build a spreadsheet using the article you mentioned above.

In step 8 , the author find the maximum(critical) flux by doing an iteration.What he does is keep lowering backpressure until the flux calculated "G" becomes maximum.
To do this, i am using P-H diagram by staying on the same entropy line and lowering pressure by a certain value.This will get Sp volume and enthalpy. However, it is very tedious unless you have a computer model(I dont know how to do this in hysys or aspen).

My question is , to find the critical flux, can we use the formula for Critical pressure (using k value)?? in this way, we dont have to do iterations.

what is the difference(thermodynamically) between these two methods?

regards,

c

[SSWBoy]

In step 8 , the author find the maximum(critical) flux by doing an iteration.What he does is keep lowering backpressure until the flux calculated "G" becomes maximum.
To do this, i am using P-H diagram by staying on the same entropy line and lowering pressure by a certain value.This will get Sp volume and enthalpy. However, it is very tedious unless you have a computer model(I dont know how to do this in hysys or aspen).


In Pro-II I achieved this by entering the formulae into a Calculator (I believe the equivalent in HYSYS is using a Spreadsheet) then using an Optimize function to maximise the target variable, from my vague recolection of HYSYS I believe you should be able to do the same.


My question is , to find the critical flux, can we use the formula for Critical pressure (using k value)?? in this way, we dont have to do iterations.

The properties of supercritical fluids vary dramatically around the critical pressure, I suspect that the method of finding the critical pressure ratio is also based on an ideal gas and hence is not applicable for this situation. Typically Cp/Cv is fairly constant with temperature change but I doubt this is the case for fluids in the supercritical region.

In any case depending on the initial conditions it is possible to relieve a 2-Phase fluid, resulting in a larger (i.e. instead of being ~0.55 you will have a ratio of 0.8) critical pressure ratio. Your formula would not predict this.

what is the difference(thermodynamically) between these two methods?

Maximising the isentropic flux by varying backpressure is reality, using the k value is only applicable ro ideal gas's.

As you note, trying to maximise the isentropic flux is cumbersome, I recommend you look into using the DIERS method for sizing this type of orifice (2-point method). Also note that the algorithim considers heat input based on C.F.A*0.82, although this is the eqn for liquid vessels bear in mind that a supercritical fluid should be considered as a gas-filled vessel, therefore the possibility of vessel rupture is a real one. Feel free to ask any questions if needed.

[cmp74]

In step 8 , the author find the maximum(critical) flux by doing an iteration.What he does is keep lowering backpressure until the flux calculated "G" becomes maximum.
To do this, i am using P-H diagram by staying on the same entropy line and lowering pressure by a certain value.This will get Sp volume and enthalpy. However, it is very tedious unless you have a computer model(I dont know how to do this in hysys or aspen).


In Pro-II I achieved this by entering the formulae into a Calculator (I believe the equivalent in HYSYS is using a Spreadsheet) then using an Optimize function to maximise the target variable, from my vague recolection of HYSYS I believe you should be able to do the same.


My question is , to find the critical flux, can we use the formula for Critical pressure (using k value)?? in this way, we dont have to do iterations.

The properties of supercritical fluids vary dramatically around the critical pressure, I suspect that the method of finding the critical pressure ratio is also based on an ideal gas and hence is not applicable for this situation. Typically Cp/Cv is fairly constant with temperature change but I doubt this is the case for fluids in the supercritical region.

In any case depending on the initial conditions it is possible to relieve a 2-Phase fluid, resulting in a larger (i.e. instead of being ~0.55 you will have a ratio of 0.8) critical pressure ratio. Your formula would not predict this.

what is the difference(thermodynamically) between these two methods?

Maximising the isentropic flux by varying backpressure is reality, using the k value is only applicable ro ideal gas's.

As you note, trying to maximise the isentropic flux is cumbersome, I recommend you look into using the DIERS method for sizing this type of orifice (2-point method). Also note that the algorithim considers heat input based on C.F.A*0.82, although this is the eqn for liquid vessels bear in mind that a supercritical fluid should be considered as a gas-filled vessel, therefore the possibility of vessel rupture is a real one. Feel free to ask any questions if needed.

SSWboy,

thnks . You cleared a lot of my questions.

I will try to find a way to do it in HYSYS as you did in PRO-2.

My case is a light hydrocarbon which is plant feed and the equipment is the feed filter. The operating pressure is > Critical pressure but operating temp is < critical temp, so i think i can still use the equation for liquid filled vessel as given in API 521.

C.

[SSWBoy]

SSWboy,

thnks . You cleared a lot of my questions.

I will try to find a way to do it in HYSYS as you did in PRO-2.

My case is a light hydrocarbon which is plant feed and the equipment is the feed filter. The operating pressure is > Critical pressure but operating temp is < critical temp, so i think i can still use the equation for liquid filled vessel as given in API 521.

C.

No problem, happy to be of help.

Bear in mind though that as the fire goes on in time the temperature will increase, pushing you into the supercritical region. I only mentioned the fact about the liquid filled vessel as a note that if it is a critical calculation then you may want to look into carrying out a more rigorous calculation as per API where you consider heat transfer from flame to shell and shell to gas... but as you may guess this gets messy very quickly

Please don't neglect supercritical relief, since as you are above the critical pressure heat input can never lead to the generation of vapour. The only relief you would have below Tcrit is due to liquid expansion, given that the rate of change of density with temperature is low would lead to a small orifice. BUT as you increase the tempearure above Tcrit you get a sudden change of density (almost instantaneous at Pcrit and more gradual as you increase the reduced pressure) which would lead to a much larger orifice size, worsened by the fact hat you may get two-phse choking.

Now as for which property package to use in your simulator... well thats a whole other question!

How deep does the rabbit hole go...