摘要

This paper presents a comparison of the plastic collapse loads from experimental in-plane bending tests on three 90 degrees single un-reinforced mitred pipe bends, with the results from various 3D solid finite element models. The bending load applied reduced the bend angle and in turn, the resulting cross-sectional ovalisation led to a recognised weakening mechanism. In addition, at maximum load there was a reversal in stiffness, characteristic of buckling. This reversal in stiffness was accompanied by significant ovalisation and plasticity at the mitre intersection. Both the weakening mechanism and the post-buckling behaviour are only observable by testing or by including large displacement effects in the plastic finite element solution. A small displacement limit solution with an elastic-perfectly plastic material model overestimated the collapse load by more than 40% and could not reproduce the buckling behaviour.
The plastic collapse finite element solution, with large displacements, produced excellent agreement with the experiment. Sufficient experimental detail is presented for these results to be used as a benchmark for analysts in this area. Given the robustness of non-linear solutions in commercial finite element codes and the ready availability of computing resources, it is argued that pressure vessel code developers should now be recommending large displacement analysis as the default position for limit and plastic collapse analyses, rather than expecting engineers to anticipate weakening mechanisms and related non-linear phenomena.

  • 出版日期2010-10