Desalter PSV calculation 1

[fdomin]

The desalter is full of liquid, thus in a fire contingency a two phase relief wil occur.

How can I estimate latent heat for crude for relief capacity calculation? What about the amount of gas being relieved along with the liquid?

 

[TD2K]

Typically, you flash about 5% of the liquid to vapor at the PSV relieving pressure in a simulator and use that duty and mass of vapor formed to determine the latent heat. The flashed vapor also gives you the density and other physical properties for the PSV sizing. This means you need a typical crude analysis, with an accurate value for the light ends.

You can also do a PSV capacity check for the volume of liquid that would be displaced due to thermal expansion until the crude starts to vaporize using the method in the GPSA manual (this method used to be in API 520 also but I couldn't seem to find it in the latest edition) but this won't be the sizing case in my experience.

Once the PSV starts to relieve, assuming it is on the high point of the desalter, it will relieve only vapor.

 

[fdomin]

Thanks TD2K.
I think that, as you stated is not the sizing case but blocked flow, nevertheless I need to check all other contingencies.
You said that it will only relieve vapor. Would not the relief be vapor-liquid?.

 

[TD2K]

It depends on the PSV installation configuration with respect to your piping and the vessel. Typically, the PSV is located off a high point so as the material boils in the vessel only the vapor is relieved through the PSV. If the PSV for example was located on the liquid piping, then it 'could' relieve only liquid, but with the volumetric flow rate equal to the amount of vapor being generated inside the vessel due to the fire.

You need to look at your installation detail and decide what is going to happen. Normally, for fire relief, even though do you a 'flash' on the liquid, I've seen vapor relief only listed but I stress you need to look at your installation.

 

[rusman]

Just wanna add to TD2K thread # 1. You can refer to the latest API RP 521 for the thermal relief of the block fluid (which include the cubical thermal expansion of the crude). Latest API RP 520 Part 1 has done major revision. API RP 521 will explain detail about the liquid thermal expansion.

TD2K is right about the relieving rate and scenario of the PSV for fire case. But you still need to check whether the crude is at critical condition at the relieving state. This can be done using process simulation software.

Think about the blocked discharge, if this is the first equipment installed in your system, with the design pressure is less than the upstream design pressure, then you should consider block discharge case.

hope this help

 

[maxxum]

rusman,

what about we fall in the case of critical conditions??

 

[grt12]

Maxxum,
If you are in the dense phase (above the critical point) then the relief has to be regarded as a thermal expansion of a vapour and the relief calculation becomes a little more difficult. Latent heat is not an issue under these conditions.

 

[fdomin]

TD2K, the PSV's are elevated about 10 ft above top of desalter.
So, vapor will be the only relief in case of fire?
Pipe linking PSV & desalter is 4". As this pipe also contains liquid, will the relief be two phased at the beggining and then only vapor phase ?

 

[TD2K]

I'd size it for the calculated vapor relief (heat input divided by your effective latent heat). I'd also check it for a thermal relief expansion case but this should be less than the vapor relief.

One could argue that as the crude light ends starts boiling and turns to vapor, a resulting equal volume of liquid initially would have to be displaced through the PSV before the gas can fill the top portion of the vessel and 'supply' 100% vapor to the PSV. If used as the approach, this would make a very large PSV load. However, I think you can argue that if this occured (exceeding the PSV's capacity), the pressure in the vessel would have to rise because of the lack of capacity. The increase in pressure in turn suppresses the vapor formation and you'd be back to the equivalent of a blocked in thermal expansion case for the PSV. So, I'd argue this is not a valid sizing approach and go back to a vapor relief load.