Improved stochastic methods for modelling imperfections for buckling analysis of composite cylindrical shells

Abstract

Imperfection sensitive CFRP cylindrical shells are used in a variety of civil and aerospace applications and feature a large scatter in buckling load levels induced from imperfections introduced from their manufacture. Currently, there are no complete methods that can realistically simulate cylinders with a full spectrum of imperfection types for a complete diagnosis of possible buckling loads. This forces shell designers to utilise an outdated, inefficient and conservative design philosophy that is unsuitable for modern manufacturing methods and materials. Stochastic analyses can optimise and improve the robust design and reliability of such cylinders through accurate prediction of the range of conceivable buckling loads by realistic simulation of structural imperfections. Such imperfections include initial shell-wall geometric, thickness and material imperfections and non-uniform applied end-loads. A procedure which aims to improve the stochastic modelling of thickness and material imperfections in imperfection sensitive composite cylindrical shells is proposed. Monte-Carlo simulations of axially compressed cylinders are performed to show that the stochastic methods described here are able to capture the scatter introduced from the imperfections. The results show that the axial buckling load of the specific cylinder analysed here can be reduced to 29.5kN and increased to over 40kN from a perfect load of 38.2kN from material and thickness imperfections alone.