Hi,

I'd like to have some advise about how to choose a safety relief valve for both liquid and vapour relief.

Following scenario's for overpressure can occur in our batchreactor (PSV is set @ 3barg)

- Fire: 12000kg/h relief capacity (ethanol).
- Coil rupture: 385m³/h of cooling water must be relieved (Max pressure of pumps = 4.6barg).

Do I need two different PSV's? One for liquid and another for vapour relief? Or is it possible to combine those two functions in one PSV?
For both scenario's the downstream pipe of the PSV needs to be installed 4 meter above the highest floor, what implicates that the build - upbackpressure for liquid relief will be increased with ~ 1bar.

Is there anyone who has experience with this kind of cases?

Best regards.

Thanks for the many interesting answers to the question.

Jnol,

Hang on to your hat! I guess the forum guys were busy elswhere. Anyway, here is my take:

API 520 Part I does differentiate between PSVs for vapor and liquid service. Having said that:

Para 2.2.1.2.9 specifies that " spring loaded pressure relief valve designed for liquid ( or liquid and gas) applications and which are balanced to minimize the effects of back pressure are recomended for two phase applications when the fluid being relieved may be liquid, gas or a multi phase mixture"....

In other words one can get a valve suitable for both gas and liquid service. The vendor should be consulted for advice on the type of valve. It should definitely be balanced. Para 2.2.1.2.10 further specifies the type of valve based on location.

Futher thoughts:

a) I do not know the system set up, but it looks like you have cooling water flowing through the coil and ethanol is in vapor phase in the reactor. This means when you have a coil rupture, the vessel will fill up with water. If the vessel becomes water filled, then the pressure will rise. If the cooling water is coming from a centrifugal pump discharge, then the pump will go to dead head. If one ensures that design pressure of the vessel is above the dead head of the pump, then one need not have a safety valve for this case. You mention the dead head at 4.6 barg and I am sure the vessel design pressure would be above this.

The reason I am saying this is that I would typically see the ethanol vapors routed to a flare. In that case one cannot dump 380 m3/hr of water into the flare header. This would have significant consequences on the flare header support and KO drum sizing.

Hope this helps.
Regards

Thanks for the clear answer

jnol