Critical Buckling Strain Equations for Energy Pipelines—A Parametric Study

A recent review of available predictive critical buckling strain equations for segments of line pipe has shown that the equations give poor test-to-predicted ratios when validated using the more than 50 full-scale experimental pipeline test results available in the University of Alberta (U of A) database (Dorey, A. B., Murray, D. W., and Cheng, J. J. R., 2000, “An Experimental Comparison of Critical Buckling Strain Criteria,” Proceeding of the International Pipeline Conference, Calgary, Alberta, Oct. 1–5, ASME, New York, pp. 71–77, Paper No. IPC00-0157.). The pipeline specimens in the experimental database were subjected to a combination of axial load, internal pressure, and monotonically increasing curvature with magnitudes representative of those that might be experienced under field operating conditions. Research has been undertaken at the U of A to develop more reliable equations and a database of over 200 experimental and numerical results now exists. The numerical results were generated using a nonlinear finite element analysis (FEA) model that was validated using the experimental database. The FEA model provided a mean test-to-predicted ratio for the peak moment capacity of 1.025 with a coefficient of variation of 0.040 and a mean test-to-predicted ratio for the local critical buckling strain of 0.997 with a coefficient of variation of 0.067 (Dorey, A. B., Murray, D. W., and Cheng, J. J. R., 2005b, “A Comparison of Experimental and FEA Results for Segments of Line Pipe Under Combined Loads,” ASME J. Offshore Mech. Arct. Eng., in press.) for the 162 load cases analyzed. This paper presents the new predictive critical buckling strain equations developed from the U of A database.