Calculating SIF using FEA 1

[parijat10]

Suppose we need to calculate SIF at a tee junction and the vonmisses stress at a tee junction is calculated then to calculate the SIF how should be the other FEA model against which comparison should be made and what stress value to be obtained from the other stress model i.e. will it be vonmisses or some thing else?

Thanks and Regards
Parijat Bandyopadhyay

 

[TERIO]

Not sure what code you are using. I don't think B31.1 or B31.3 really give definitions of the SIF. Per ASME B&PV Code NC-3673.2(d) an SIF is the "ratio of the bending moment producing fatigue in a given number of cycles in a straight pipe with a girth butt weld to that producing failure in the same number of cycles in the fitting or joint under consideration."

NC suggests that an empirical relationship i=C₂K₂/2 may be used to determine the SIF, where C₂ (primary+secondary stress) and K₂ (peak stress) are the stress indices from class 1.

This EPRI Report gives some background on SIF's.

Also, the stress indices represent stress intensity (Tresca) not equivalent stress (von Mises).

 

[ColonelSanders83]

SIF's in the B31 piping code series are imperical values derived from tests. If new SIF's are required they must be developed in accordance with "B31J - 2008 Standard Test Method for Determining Stress Intensification Factors (i-Factors) for Metallic Piping Components"

http://catalog.asme.org/Codes/PrintBook/B31_J_2008_Standard_Test_Mthd.cfm

Your approach is applicable to pressure vessel type components, not piping component design.

The SIF (Stress Intensification Factor) cannot be compared to calculated FEA stress concentration stresses or even empirical stress concentration factors. The Piping SIF's listed in the code have been corrected so that comparisons can be made to the fatuige limits of a circumferential butt weld, not straight pipe (this is why pipie to pipe butt welds do not have a stress concentration factor, a stroke of genius by Mr. Markl as you don't have to evaluate every single butt joint in your system).

Your only B31 code compliant option in developing new SIF's is to test in accordance with the above standard.

A question properly stated is a problem half solved.

Always remember, free advice is worth exactly what you pay for it!

http://www.ap-dynamics.ab.ca/