Temperature for PSV orifice sizing 2

[mab3]

When sizing an orifice for a pressure relief valve set at 700 psig based on external fire for a 600# ANSI pressure vessel containing LPG liquids (either propane or butane), what temperature should be used? Would the fluid act as a liquid in the super critical region at the inlet to the orifice, and then immediately flash at the outlet?

 

[TD2K]

I've been thinking about this and haven't seemed to tie it all together, perhaps if I post this someone else will be able to pull it together (Morten, Pete?).

Start out with your initial P, T and level in the drum. Calculate the mass of liquid and vapor (you'll use this later). Now, as the fire heats up the liquid, it follows the vapor liquid phase envelope until it reaches the critical point and becomes a supercritical fluid. Now, at this point, the volume for the fluid is the vessel's volume. The mass is the sum of the initial vapor and liquid masses so you know the density in the drum. You can plot this on the Mollier diagram and you know as the fire continues to input heat into the system raising the pressure and temperature, that you HAVE to follow along a constant density line (since the volume effectively isn't changing and the mass is fixed). At the relieving pressure, you can then use the Mollier chart to scale off the relieving temperature (correponding to the 'fixed' density at the relieving pressure).

Where I seem to have a disconnect is at the critical point, you should also be at the critical density (?) and how do you then 'jump' to the density you've just calculated you need to have?

Anyway, assuming we work this out AND I'm right %- Then, flash this mixture to the outlet pressure. Looking at the Mollier chart, I'm pretty sure you'll have 100% vapor and you can size the valve for a vapor relief. Watch your Z (which can be back calculated from the density, pressure and temperature), if it's below 0.5, you might want to consider a minimum 'effective' limit as Zs can be somewhat sensitive and directly affects the required area.

 

[rusman]

You need to model the simulation using isochoric flash (constant volume flash). As hilited by TD2K, since this is closed system, and your volume and mass is constant b4 your PSV is popped, then the density is constant.

You need to model using Heat Exchanger module (I bet neither HYSYS not PROVISION have isochoric flash module). In order to do this, you need to define your inlet stream vis your operating condition with predefined fluid compositional data and EOS (either PR or SRK will do). You need to heat the inlet stream to the heat exchanger and the outlet pressure of your heat exchanger shall be your relieving pressure of your psv, ie the set pressure plus accumulation (21% since this is fire case). Refer back to API RP 520 Part 1.

Having define your outlet condition (pressure), you need to do trial and error (or perharps using the adjust module in HYSYS or controller module in PROVISION) by adjusting the outlet temperature of the outlet HX and until your inlet mass density is the same with your outlet density. The calculated temperature is your relieving temperature and you can check whether it fall under supercritical region from the vapour fraction.

Hope this help.

Thanks

Rusman

 

[grt12]

If you use the depressuring module in HYSYS, it will heat the contents isochorically. What you have to do is set the initial conditions as your starting conditions, and set it to do a fire case depressuring. It will first heat the contents until it reaches the defined relief pressure - at this point you can see the relief temperature.

 

[MortenA]

I have always found this this particular question difficult. But you have to remember that the code says SP+21%

What I usually do is to calculate bubbel T at SP+21%

But if you are above PC/TC then this dosnt make much sense.

What TD2K and Rusman describes seems right (just though i would mention the 21% for sizing)

Please remember that there may be a company "policy" or whatever its called re. this subject. These might represent a special interpetation of the code/thermodynamic. Usually its pointless to argue about such policies because all old valves are designed following this rule and nobody has the stomac to re-calculate all valves and perphase having to change some of them if your "brilliant" interpetation is correct ;-)

Best Regards

Morten