I am analysising a system where the working pressure is some 92% of the flange rating at op temperature, and the system has some high moments acting on some of the flanges.
Using the Kellogg formula for Pe + Pd < Pa, is clear enough, but my question is what is Pa. I read it as Pa is the flange rating at pressure, with my moments and forces [Pe] coming from the OPE load cases rather than SUS and EXP or OCC. If this is the case I need only when the system drops to some 70% of operating pressure have I got any leeway.
I have seen a report from a well respected analyst who, with out stating it seems to be using 1.33 * Pa or more [surely not the displacment stress range v allowable stress ratio?]
Can anybody shed some light on this or set me right

The Kellog formula ( equivalent pressure method) which was later on used in ASME SEC III, really makes life difficult for stress engineers particularly if the system pressure is very high.However I don't think that the formula considers all the aspects that are involved in the Flange leakage analysis.This formula can be simply derived by equating a value of the longitudinal pressure stress with that due to the applied bending moment and axial force. So in a nutshell this only converts the forces and moments into an excess pressure which is applied on the Flange. It does not address the relative deformation of bolts, Flanges and gaskets ( particularly the differential expansion of these components ,creep relaxation behavior of the gasket and the change in gasket modulus with temp. and other parameters all of which significantly contribute to the leakage problem ).

Coming to your question on Pa, I don't think there is much room for ambiguity.It has to be the pressure at the rated temperature at which you are computing the forces and moments.It may be your SUS case or OPE case , but I don't find any rationale for using the range of reaction ( EXP case) as the essence of this method is to convert the moment and axial force into a pressure stress which by no means is a secondary stress.

The use of 1.33 factor does not have any code or standard basis but probably a tacit way of incorporating the 1.33 OCC factor , considering the high forces and moments are ( even though taken from OPE case) are of the occasional nature.Say if these loads are due to an OPE case which is design or upset condition, based on Engineering judgement they can be considered as OCC case. I can't make out any other explanation for this 1.33 factor.

I don't know about your project requirement, however I feel that instead of this equivalent pressure method, why don't you use the EN1591 method which I think addresses the leakage question more explicitely. Even the ASME J factor check ( of course my personal opinion ) is a better option for leakage check than this Kellog Method.


A.Bhattacharya

Stress Analyst

Bechtel Corporation