When is it important to evaluate a change in the direction of a load on a piping component? How do analysts perform such an evaluation? These questions are intended to address both changes due to operation, and due to constructed modifications to the piping.

For example: An existing A106 carbon steel B31 piping system having a hundred or so full temperature cycles (FTCs) from ambient to 750 Deg. F is modified by adding a branch connection via a hot tap.

A B16.9 welding tee in the existing system experienced in-plane bending during the FTCs that resulted in stresses greater than Sy.

The addition of the proposed branch line will result in out-of-plane bending during a projected remaining life of another hundred or so FTCs, also with stresses greater than Sy.

If the owner directs the designer to follow the requirements of the B31 Code for the design of system modifications, is a cumulative life evaluation in accordance with BPVC Sec. VIII, Div. 2, Appendix 5 necessary, or is a B31 analysis sufficient for the evaluation of the tee (and other piping components)?

I think that if the secondary stresses were below Sy, then a B31 analysis is sufficient. With secondary stresses in the plastic region (due to shakedown) I’m not understanding the effect of stress on a component when developed from loads from one direction, later receiving stresses from loads in a new direction. It appears to me that Appendix 5 addresses this situation in 5-110.3(g) where it requires the evaluation of “…two or more types of stress cycle(s)…”

The Markl tests evaluated bending in-plane in one test, and then out-of-plane in a separate test. However, bending between planes, or the cumulative effect of bending in-plane and then bending out-of-plane on the same fitting were not evaluated (that I have found).

The B31 Codes evaluate the secondary stresses as individual load cases, not the cumulative effect of various load cases on a node or component. If Appendix 5 considers it important to evaluate the cumulative effect of different types of stress cycles, why doesn’t the B31 Code require the evaluation the cumulative effect also?