This study examines the reliability of small-scale plastic models in the determination of elastic buckling pressures of thin-shell structures.
Most problems that confront the structural engineer fall into one of three categories, namely: (1) stress-distribution problems, (2) ultimate load or strength problems, and (3) stability problems. Thin-shell roof structures, because of the great efficiency with which they transmit forces, are quite slender, and consequently subject to failure through a loss of geometric stability. A meaningful analytical prediction of such a stability loss can be obtained only for extremely simple cases. The possibility of using small-scale structural models to obtain an experimental solution has proved necessary.
This work shows how, and to what extent, the theories of probability and statistics can be applied in the experimental design method. The experimental program described—conducted on polyvinyl chloride spherical domes—was intended to deduce the magnitude of the pertinent material properties and their possible variations, the repeatability of buckling pressures from different models, the effects of shell thickness and middle surface geometry scale, the effect of boundary conditions, and the effect of the means of applying load.
The results demonstrate that reliable buckling predictions can be made, but that the means of providing boundary support is very critical.