6 replies to this topic

[Scott E.]

I read an older topic discussing whether to use the ideal or real gas heat capacity ratio for sizing RVs.  I understand that using the ideal gas k is almost always more conservative (since it's usually lower).  However, one topic that forum did not discuss is the effect on inlet pressure drop and thus dangerous RV chattering.

 

For example:

 

I have a PSV in ethylene service.  For my conditions, the ideal k = 1.066 and the real k = 3.778.   Using the ideal k to size my PSV, I come up with a required area of 0.347 and select an "F" orifice RV.   Using the real k, I come up with a required area of 0.206 which would result in selecting an "E" orifice RV.

 

For the "F" orifice RV and using the ideal k, my sizing program calculates a rated flow rate of 18,300 lb/hr which results in an acceptable pressure drop.   However, here's my concern: if the real k is more indicative of actual conditions, the flow capacity of my "F" orifice is much higher.   For that same "F" orifice RV when using the real k, my sizing program calculates the capacity as 26,900 lb/hr.  This results in an inlet pressure drop over 5%.

 

Because there's so much conservatism already built into RV capacity determination, I see the high inlet pressure drop as a larger safety concern than the possibility of an "E" orifice RV being slightly undersized.

 

Thoughts?

Edited by Scott E., 24 October 2013 - 11:08 AM.

[Bobby Strain]

Well, first, you should not size a PSV. The valve supplier should size it. Second, you should take a look at the sizing formulas and their derivation to understand why the ideal value is used for orifice sizing.

 

Bobby

[Scott E.]

Bobby,

 

The PSV supplier is not going to size my piping for me, so the question of what the flow through the PSV is actually going to be and what that means for inlet pressure drop and potential chattering is still valid.

[ryn376]

Basically the equations developed are for ideal gases. If your gas does not follow the ideal gas law use equations in Appendix B.

 

According to API STANDARD 520 8th ed., PART 1 pg 56:

The ideal gas specific heat ratio is independent of pressure. Most process simulators will provide real gas specific heats which should not be used in the above equation; otherwise the pressure relief device may be undersized. The value of C can be obtained from Figure 32 or Table 8. For ideal gases, where k cannot be established, it is suggested that a conservative value of C equal to 315 (0.0239) be used.

 

and pg 63:

k is the ratio of the specific heats (Cp/Cv) for an ideal gas at relieving temperature; the ideal gas specific heat

ratio is independent of pressure. Most process simulators can provide real gas-specific heats, which should

not be used in Equation (18) because the real gas specific heat ratio does not provide a good

representation of the isentropic expansion coefficient (see Annex B ).

 

Some guidelines as to when the general sizing equations might not apply:

According to Section 5.6.1 (pg 53):

[The] validity of the assumption may diminish at very high pressures or as the vapor or gas

approaches the thermodynamic critical locus. One indicator that the vapor or gas may be in one of these regions is a

compressibility factor, Z, less than approximately 0.8, or greater than approximately 1.1. To ensure the most

appropriate sizing results, users should establish the limits of applicability for their own systems. See Annex B for

guidance on sizing when this section is not applicable.

Edited by ryn376, 24 October 2013 - 12:58 PM.

[ryn376]

If you are worried about chattering, look at Relief device inlet piping: Beyond the 3 percent rule

 

http://smithburgess....3-percent-rule/

 

It has equations to calculate whether or not a valve will chatter.

Edited by ryn376, 24 October 2013 - 01:10 PM.

[marchem]

possibly you mean this thread

'http://www.cheresour...ng-temperature'

with API formulation you should use the ideal cp/cv,

in the thread different test cases are discussed
and results presented,

in addition there are Excel pages with API formulation
which allow to compare the procedures

Edited by marchem, 24 October 2013 - 01:51 PM.

[Scott E.]

Thanks for the replies.  The link to the PSV discharging temperature thread was useful in understanding that using the real Cp/Cv along with the compressibility factor accounts for the compressibility factor twice.

 

The Smith & Burgess article was also extremely interesting.  I saved that one away.  I had not previously come across a proposed systematic method for rationalizing pressure drops higher than 3%.