On the random field of the buckling load of cylinder shells under combined load
Please login to view abstract download link
This contribution presents the results of an experimental and numerical study of the load carrying capability of 22 circular cylindric shell under combined axial load and bending moment. While the stochastic distribution of axially loaded cylindrical shells is an intensively studied subject (see, e.g., [1]), the distribution under combined loading has not been considered so far. The multiaxial loading is introduced by applying the axial load with an eccentricity which leads to both a bending moment and an axial load. The measured buckling load and buckling moment are combined into a normalized load factor. This load factor is considered as a random field. The load eccentricity is the parameter which represents the coordinate of this random field. For the numerical probabilistic analyses via Monte Carlo simulations, geometric imperfection, the spatially varying shell thickness, and the Young’s modulus are considered as random parameters. The random fields of the geometric imperfections and the thickness are parametrized by Fourier series. The stochastic distributions of the highly correlated Fourier coefficients are determined from measurements. The comparison of the output random fields obtained from numerical analyses and experimental tests is not straightforward, since only discrete values along the field variable area available. The results reveals that the scatter of the studied cylinders is mainly driven by the thickness variations, due to the limited slenderness of the cylinders. Therefore, the stochastic distribution changes only slightly for increasing eccentricity, i.e. the random field of the buckling load is almost stationary.