Code equivalent stress

[prabhuu]

hai all,

In the code stress calculation as per ASME B31.3 for sustained case the equivalent stress Se is combination of axial stress, bending stress and torsional stress alone. i.e., Se= √(│Sa│+ Sb)2 + (2St)2 ). Why Circumferential stress is not considered for finding the overall stress?

If my wall thickness chosen is adequate only to withstand the circumferential stress due to pressure then how to ensure my wall can withstand bending and torsional stress by using axial stress instead of circumferential stress in the code formula.

This contrary can be clearly understood when we have a look on kellogs method for trunion calculation, in which the bending stresses are directly combined with circumferential stress not with axial stress as like ASME B 31.3. Because of this nature even if we pass a system in stress analysis most of the time trunion support get fails in kellogs method.

My opinion is that instead of using axial stress in code calculation Circumferential stress need to be used. Can any one guide me properly..

Thanks in Advance..

 

[BenStewart]

If pressure was the only load to consider on your pipe then the circumferential stress would be the greatest (first principal stress) and would then govern the thickness requirement.

If however your piping system has additional loads - self weight, thermal, wind, displacements etc. - then these would need to be considered when calculating the stresses. Note below:

- an axial force results in longitudinal membrane stress;
- shear forces result in shear stresses;
- in and out of plane bending moment results in longitudinal bending stress;
- torsion results in a shear stress.

These would need to be summed as appropriate to the calculated pressure load stresses.

If the longitudinal stress calculated from the loads (axial force and bending moments) when summed to the longitudinal pressure stress is greater than the circumferential stress (...calculate principals, then say stress intensities...) then it is clear why the longitudinal stress would govern. If the longitudinal stress doesn't govern, you'll have already designed your pipe to be sufficiently thick for pressure, including mill tolerance, corrosion allowance etc.

I know I haven't written this response in terms of the process piping code or Kellog's method but I hope this helps with your understanding of stresses.

 

[prabhuu]

hai Benstewart..

I agree with the concept that you had explained. Still I have doubt on combining the stresses to obtain the maximum/actual stress. My question is that why bending stresses are always combined with the longitudinal/axial stresses why can't it is combined with circumferential stresses( bending moment can cause both longitudinal bending stress and circumferential bending stress). AS per ASME B31.3 bending stresses are combined with Longitudinal stresses only and As per Kellogs method Bending stresses caused by longitudinal bending moment are combined with both the longitudinal stress and Bending stresses caused by circumferential bending moment are combined with circumferential stresses separately. Here Kellogs method seems to be more conservative and realistic approach.

 

[BenStewart]

If the bending moments are global in-plane and out-of-plane moments it would be incorrect to sum the resulting bending stresses to the circumferential membrane pressure stress, as these result in longitudinal bending stress.

I am not intimately familiar with the Kellogg method but believe it's used for assessing stresses in the pipe local to a support? I also understand it has been used for many years without any reported failures (for static loading) but despite this is not necessarily an accurate method for evaluating stresses local to a pipe support; when compared to FEA. Perhaps the method attempts to calculate a local circumferential bending stress that would be additive to the circumferential membrane stress which would then need to be compared to the relevant allowable stress.

If you are really concerned about the stresses local to the support have you considered completing FEA? If so I'd recommend reviewing Part 5 of ASME Section VIII Division 2.

Perhaps there are others that are more informed on Kellogg's method that can assist, although i'd recommend a search through the eng-tips archives as the Kellogg method has been discussed before...