Metamodel-based lightweight design of an automotive front-body structure using robust optimization

摘要

Both conventional automobiles and new energy cars require urgently lightweight design to realize energy economy and environmental protection in a long run. The weight reduction of the body structure plays a rather important role in decreasing the weight of the full vehicle. In real engineering problems, the variation in sheet gauge, geometrical size, and material parameters caused by environmental factors and other uncertainties may affect the structural performances of body components. Therefore, a lightweight design without considering this kind of tolerance may result in the loss of feasibility and reliability in engineering application. From the viewpoint of crashworthiness performance, this paper presents a study on the lightweight design of the automotive front-body structure based on robust optimization, considering the variation in design variables including sheet gauge and yield limit of materials. Coupled with the design and analysis of a computer experiment, four metamodelling techniques, namely support vector regression, kriging, radial basis functions, and artificial neural networks, are employed to build the metamodels of structural crashworthiness performance indicators for comparison of approximation accuracy. An adaptive deterministic optimization process is used to upgrade further the approximation accuracy of metamodels for the following robust optimization. A double-loop strategy is chosen when solving the robust optimization formulation and the basic Monte Carlo simulation method is applied to perform a reliability analysis. A genetic algorithm solver is used to obtain both the deterministic and the robust optimum results for comparison. The reduced weight obtained by using robust optimization is 7.8003 kgf (19.45 per cent) and the result achieved from robust optimization is more conservative than that obtained through deterministic optimization as expected. However, the robust optimum design is ensured to be feasible and reliable when the variation in design variables works in a real engineering application.