ON THE ANALYSIS OF CYLINDRICAL SHELLS WITH RANDOM IMPERFECTIONS

In this paper a stochastic model of random imperfections of cylindrical shells is presented. For this purpose an extensive data bank has been used to estimate secondmoment characteristics of these imperfections. The imperfections are then modeled as a twodimensional, uni-variate random field. Based on the stochastic model of the initial imperfections, the Karhunen-Lo eve expansion as well as the spectral representation method respectively, are applied for simulation purposes. The simulation concept introduced in this paper is designed to improve the prediction of the large scatter in buckling loads, as generally observed in experiments