Abstract
The authors were tasked with designing and fabricating a thick walled (t > 4.5″) ASME Division 2 – Class 2 separator vessel. Due to its service requirements, the vessel is to be regularly hydrotested at 18.87 MPa (2,737 psig). Linear-elastic finite element (FE) evaluations of the vessel indicated that it passed all required Code checks, including the hydrotest check specified in Section VIII, Division 2, Paragraph 4.1.6.2.
To develop a greater understanding of the advantages and disadvantages of each method, the FE analyst on the project routinely reanalyzes vessels that have been evaluated per the linear-elastic procedures of Part 5 of the ASME Section VIII, Division 2 Code with the nonlinear procedures also specified in Part 5. This practice allows for direct comparisons of the linear and nonlinear results and for identification of situations where nonlinear analyses could provide benefit. Such an analysis was performed on this vessel under the hydro-static test condition. However, this analysis failed due to solver failure / gross instability (plastic collapse) before the full hydro-static load was applied. The solver failure was confirmed and repeated in multiple FE packages. This presented a conundrum for the authors: should the linear-elastic results be accepted since the vessel passed the linear evaluations, or should they be invalidated since the nonlinear evaluations indicated that failure could occur during a hydrotest, which given the vessel’s operations, will occur frequently?
This paper discusses the additional evaluations that were required to establish confidence that the vessel could be successfully hydrotested when fabricated. These included both the Code specified evaluations, and evaluations that allowed engineering judgement to be applied to the design.
