Kellogg Pressure equivalent Pe vs Pr

[tr6]

Following the Kellogg Pressure Equivalent Method, a Pe is determined, added to the max operating pressure for P and then compared to the flange rated pressure at the max operating temperature. This produces a simple, but conservative, result that will probably disprove a majority of flange connections (Peng).

I am reviewing a piping spec where the company compares the total (Pe + P) to 1.5*Pr (rated pressure).

My question is: "Does increasing the rated pressure by a factor of 1.5 represent a more "realistic" comparison or an approach that results in an under-designed piping system?".

 

[BigInch]

To quote Peng, "This produces a simple, but conservative, result". Would that not be the opposite of "under-designed".

 

[tr6]

I guess my thought was: if you take the "conservative" result and instead of comparing to the flange rated pressure, you increase the rated pressure by 1.5. Does that not raise the acceptance level by 50% and could result in under-designed?

 

[BigInch]

You'll want to see this and other bits

mgp (Mechanical) 7 Mar 02 16:52

For anyone interested I have a copy of an interesting article on the subject. It is from an ASME conference in 1981 and co-written by L-C Peng from M.W. Kellogg.

The background for the "equivalent pressure method" (i.e. the Kellogg formula) is discussed, so it may help in the absence of the Kellogg Book.

The conclusion is as follows:

quote:
"The equivalent pressure approach have become a standard method in evaluating pipe load acting on flange connections. There are two acceptable ways of checking the pipe load using the equivalent pressure. They are rating table method and stresss calculation method. The rating table method is simpler but is much more conservative by ignoring the reserve strength. It is so conservative that it would probably disprove most of the installations which are operating satisfactorily. The stress calculation method is a more realistic approach which evaluate the actual reserve strength available in a flange.
When a standard flange is rated for a certain pressure, it normally posses sufficient strength to resist the rated pressure load plus sustantial reserve strength to resist the pipe load. This reserve strength varies from flange to flange and is not known until a stress analysis is performed. It is also highly dependent on the thickness of the connecting pipe when a bore of the flange is specified to be the same as the inside diameter of the pipe.
In designing a special flange, it is necessary to provide some allowance for the pipe load. The allowance can be either based on the actual expected pipe load or based on the load that will produce a bending stress equivalent to one half of the basic allowable stress at the connecting pipe. In any case care should be exercised in the final piping system design to keep the pipe load from exceeding the allowance provided."
End quote.

This is interesting because in the stress analysis you may get an equivalent pressure (via the Kellogg equation), but you need a max allowable pressure to compare with because stress programmes normally don't produce a proper flange stress calculation. By using the flange rating many flanges will fail the analysis. Consequently one can save many redesigns by doing an ASMEVIII calculation of the actual flange and connecting pipe.

from this thread here,

http://www.eng-tips.com/viewthread.cfm?qid=18541

 

[StevenHPerry]

tr6,

From personal experience, a majority of flanges that come in with a (Pe+P)/Pr <= 1.5 will pass a Section VIII Div 1 Appendix 2 analysis.

It is possible that the engineer has looked at what could reasonably occur in the applications this spec is intended for and is okay with the 50% extra margin.

If I were in your shoes I'd like to know the reasoning, but the owner has final say.

FWIW, some process engineers love to set line conditions to the flange limits. (No matter the open ended drain line can't reach 285psig or that the water won't be water anymore when you go to 800°F and 80psig.) With ANY external loads on the flange, it would "fail" the Kellogg method without a multiplier on Pr. The spec engineer may be trying to avoid wasting your time and his $$$ qualifying flanges on such lines.

- Steve Perry

http://www.linkedin.com/in/stevenhperry

This post is designed to provide accurate and authoritative information in regard to the subject matter covered. It is offered with the understanding that the author is not engaged in rendering engineering or other professional service. If you need help, get help, and PAY FOR IT.