Determination of viscosity correction factor for two phase PSV sizing

[rpmUK]

I have recently been preparing preliminary PSV sizing estimates for a new build oil treatment facility. Some of the PSV sizing required consideration of two phase relief scenarios. For these scenarios I employed the Omega calculation method described in API 520 Part 1 (Eighth Edition). To complete the calculations I assumed the viscosity correction factor, Kv, to be 1.0. This assumption follows all of the worked examples included in API 520 Part 1 for two phase relief.

I have since wondered how I could have demonstrated that Kv = 1.0 (or not as the case may be). API is not explicit as to how this factor is to be determined for two phase flow cases. The most obvious way would be to use the Kv graph presented in the API standard (Figure 37). However to make use of this graph I think it would be necessary to employ weighted averages for physical properties. It is not clear, however, that this is a correct methodology. In particular I am wondering as to the validity of using averaged viscosities.

I have been trying to read up on the subject, including reading other board queries on this subject. As a result of this reading I have ordered the following paper:

Darby,R. and Kamyar Molavi, "Viscosity Correction Factor for Safety Relief Valves", Process Safety Progress, Vol. 16, No. 2, Summer 1997.

Hopefully this will address my (two phase) query, but in the meantime I wondered if anyone here had any experience with estimating Kv's for two phase relief and / or guidance on the subject?

 

[Chance17]

Well I looked at my references for two phase sizing using the omega method.
Your correct. There is no liquid viscosity used nor any viscosity correction.
My references also had the following statement.

"This methodology is based on specific assumptions about two-phase fluid behavior that may not be applicable to all fluids.
The use of alternative, generally accepted physical property models and rigorous thermodynamic computation
based on the Homogeneous Equilibrium Model (HEM) may be considered (refer to API Std-520-Part-I Appendix D for more information including examples"

There are viscosity corrections for the various liquid sizing cases.
In the old days, two phase sizing used to be "area of vapor" (using vapor sizing equation)
plus "area of liquid" (using liquid sizing equation).
If you have very high viscosity (<5,000cp) you may want to consider the old method.

 

[Latexman]

Chance17,

Did you mean ([red]>[/red]5,000cP)?

Good luck,
Latexman

 

[rpmUK]

Thanks for your comments Chance17. I would agree that further investigation would be appropriate when dealing with high viscosity liquids in two phase relief scenarios. Use of the old "liquid + gas areas" technique would make a useful check.

I hope that the API procedure will at some point be updated to clarify how Kv should be estimated for two phase cases.

For information, I have now reviewed the Darby and Molavi paper I referred to in my first post. In essence it describes the basis and accuracy of the API graph for liquid Kv. It does not specifically address 2 phase cases.