4 replies to this topic

[processrock]

Hi Engineers,

May I seek your guidance on the following?

I have a vertical vessel that is fully filled with water containing a trace amount of hydrocarbon. Due to a device located near the top dished head, the pressure relief valve (PRV) cannot be installed at the top and must instead be located on the side of the vessel, below the liquid level.

Per API 521, when a PRD is located below the liquid level of a vessel exposed to fire, the PRD should be capable of relieving a volume of fluid equivalent to the volume of vapor generated by the fire. It is also stated that two‑phase PRD sizing is generally not required for fire cases, except for unusually foamy materials or reactive chemicals.

In this case, how should the PRD be sized? If the relief device is installed on the vessel shell well below the liquid level, or on a liquid feed line, the relieving fluid under fire exposure would be liquid. If the vapor generation rate is used to drive the equivalent liquid flow rate, would this not result in an oversized PSV?

On the other hand, if only vapor generation is considered, there is a concern that the valve may be undersized.

I would appreciate your advice on the appropriate sizing approach for this scenario.

Thank you for your time and support.

[Pilesar]

Per API 521, when a PRD is located below the liquid level of a vessel exposed to fire, the PRD should be capable of relieving a volume of fluid equivalent to the volume of vapor generated by the fire. 

 

Did you not answer your own question? I do not understand why this is not satisfactory to you. Have you calculated the fire case vapor generation rate for a vertical vessel containing water?

[processrock]

My understanding is that if the inlet piping to the relief valve is submerged below the liquid level, the fluid entering the valve will not be vapor. Instead, any vapor generated will act as a driving force, pushing liquid through the relief path. As the liquid level drops and falls below the relief valve inlet nozzle, the relieving flow could transition from two‑phase to vapor‑only flow. If this is the case, why do we place emphasis on estimating the orifice size based on two‑phase or vapor flow scenarios? My assumption was that liquid flow driven by the expanding vapor would represent the worst‑case condition.

 

 

Per API 521, when a PRD is located below the liquid level of a vessel exposed to fire, the PRD should be capable of relieving a volume of fluid equivalent to the volume of vapor generated by the fire. 

 

Did you not answer your own question? I do not understand why this is not satisfactory to you. Have you calculated the fire case vapor generation rate for a vertical vessel containing water?

[Pilesar]

Your situation is curious with a completely filled water pressure vessel with no vapor space at all. If the normal inlet or exit of the vessel is not blocked in, the fire case can consider those paths as a relief exit. Completely blocking in a filled water vessel with no vapor space is an additional relief scenario you must consider anyway due to thermal expansion. Why not just put the PRD near the bottom of the liquid? Is there no bottom nozzle?
  I think there must be more to this situation that you have revealed. If the vessel has no PRD now, then it must be in the design phase where you are determining where to install the vessel nozzle for a PRD. We do not know the size, pressure, function, etc of the vessel. It sounds like an interesting problem!
   

[processrock]

Allow me to briefly provide the relevant information:

• Equipment: Hydrocyclone (vertical)
• Vessel diameter: 1.0 m
• Tangent‑to‑tangent height: 1.4 m
• Head: 2:1 ellipsoidal
• Set pressure: 30 barg

That said, may I ask how locating the PRD near the bottom of the liquid would be beneficial? Also, please excuse my naïve question: according to API 521, when a PRD is located below the liquid level in a vessel exposed to fire, the PRD should be capable of relieving a volume of fluid equivalent to the volume of vapor generated by the fire. In this context, do we assume that the required liquid volumetric flow rate is equal to the volumetric rate of vapor generation due to the fire?