Crashworthiness design for multiple loading conditions using dynamic weighting factors in HCA framework

Abstract

A lot of research work has been done in the field of crashworthiness design of structural components and systems. Most of the work in this area considers single load case resulting in designs suitable for only that one loading condition. However, in real life the structure can be impacted in many ways and hence a design should behave reasonably well under multiple loading conditions. One way of accounting this uncertainty in loading conditions is to incorporate the notion of reliability in the design process. Another, computationally less expensive, way is to design the structure for some finite number of representative loading conditions. A number of researchers have incorporated multiple loading conditions in the design method by superimposing the results for various load cases. A weighted average of certain structural properties like compliance, stress or strain energy are used for the purpose of superimposing over the multiple load cases. In most of the work constant (not changing with design cycle) weighting factors are used which are obtained either by design requirements or the probability of occurrence of corresponding load cases. In case where few of the load cases dominate over the others, it is observed that the final designs are dictated by the dominating load cases unless the weighing factors for these load cases are deliberately chosen to be comparatively smaller. However a design must behave well for all the considered loading conditions. For crash problems, force displacement behavior is one of the most widely used criteria to measure the performance of a design. In this work, three different strategies are introduced to allocate the associated weighting factors dynamically (changing with design cycle) during the design iterations such that an acceptable force displacement behavior is achieved for all the loading conditions. A well developed, Hybrid Cellular Automaton (HCA) based design methodology is used for efficient material distribution within the design domain. A crash problem with three loading conditions is used to demonstrate the implementation and effectiveness of the proposed strategies.