Max flow calculation for choosing PSV

[GeorgeGrande]

This is the case i am dealing with:

A 2" pipe connecting a pressure vessel (200.5 psia) to a pump. The pump is designed to fill and evacuate the cold and not pressurized fluid. I must protect the pump from 145 psia. In case of leaking (or open from some reson) 2" valve there will be high pressure flow to the pump.
I took worst case scenario for gas flowing through 2" pipe in sonic velocity and calculated the volume flow from pipe flow area and the velocity.
I am getting very large PSV for 2" pipe.

P=200.5 psia
T=287 oF
Tb= 97oF @ atm.
Vsonic=164 m/s

 

[dcasto]

Your worst case scenerio would be gas flowing through the line to the pump with an inlet pressure of the maximum pressure on the vessel (whatever that vessels relief valve is set at) and the pressure the pump case is rated for. You can take credit for the pressure drop to calculate maximum flow rate.

 

[zdas04]

The key term is "Credible Scenario". If you can't come up with a credible set of circumstances that could put sonic gas flow at the pump then don't use that condition.

I couldn't tell if the pump is set up to suck on the vessel or fill the vessel. If it is to fill the vessel (for example), then putting two swing checks in the line would make sonic velocity non-credible (since both would have to fail for the scenario to play out).

Bottom line is that you don't have to use worst-case conditions, you should use worst-credible conditions which are generally more reasonably sized valves.

David

 

[GeorgeGrande]

Thank you both.

Zdas04, This is a filling evacuation system. Swing checks valve of any kind will cause a problem. Do you think double block by gate or ball is enough to make this case "non credible"?

 

[zdas04]

The problem with that is that only one thing has to fail (i.e., the operator forgot to close the valves) so putting 10 manual block valves would not be adequate. You could put a PLC controlled high pressure slam valve and a local-controlled high pressure slam valve in series, but your cost would be much higher than a big PSV.

Does the line really need to be full, 2-inch? If not you could put a smaller choke nipple in the line that would significantly reduce the flow rate required by the PSV.

David

 

[dcasto]

If the fill/evacuation is not automatic (or remotely actived), why not a three way valve with when in the pump mode, the threway valve opens in a maanner that there is a check valve (or two) protecting the pump. When opened the other way, the pump is isolated and the vessel can drain. Maybe some like this could work?