Buckling of near-perfect steel torispherical and hemispherical shells subjected to external pressure

The aims of the present series of tests on externally pressurized machined steel domed ends were twofold. The first was to verify the recently predicted and unusual variation of a dome's buckling resistance with the knuckle radius-to-diameter (r/D} ratio. The second was to verify that optimizing a dome's geometry, while keeping the weight constant, can bring substantial increases in the buckling strength. The agreement between the experimental results and the theoretical predictions is very good. The initial increase in the buckling strength, its reversal, the minimum buckling strength, and the rapid increase thereafter, all as predicted by theory, are confirmed by tests. The tests on near-perfect shells highlight the sensitivity of the buckling strength of torispherical shells to the r/D ratio, in the range 0.4 < r/D < 0.5. In the optimization study, in which the geometric ratios D/t, r/D, and RS/D were varied, the ratio of the buckling strengths of the best and the worst shapes is more than 5:1. The best and worst shapes have the same weight; it is clear that optimizing a dome's shape can produce substantial increases in strength. The good agreement between the test results and theoretical predictions is further confirmation of the adequacy of the BOSOR 5 shell buckling program for problems of this type.