Abstract: In the field of vehicle lightweight design, mega-casting components have attracted significant attention due to integrated characteristics and involvement of multidisciplinary issues. Lightweight design methodology for such mega-casting components to achieve global lightweight design under multidisciplinary performance constraints is systematically established. For continuous body structures without natural boundaries, a subsystem partitioning strategy that considers the nonlinear characteristics of collisions is proposed, and model reduction techniques are employed to significantly improve computational efficiency while ensuring analysis accuracy. A compromise programming approach is used to normalize static and dynamic performance objectives and construct a comprehensive objective function. An objective decision sequence calculation method based on the performance information of the initial design state is proposed, which combines the Analytic Hierarchy Process to determine the subjective and objective combined weighting coefficients for sub-objective synthesis, and guides the design of strengthening rib effectively with topology optimization technology. A parameterized method was developed for casting components with variable thickness structures, and key thickness variables affecting multiple properties were identified through relative sensitivity analysis and the entropy-weighted TOPSIS method. The best surrogate model was selected to conduct optimization using the differential evolution algorithm, achieving a balanced approach to multi-disciplinary performance and global lightweight. The weight was reduced by a total of 5.0 kg. The lightweight method offers a robust theoretical and practical foundation, which is useful for reducing complex system models and optimizing the design of mega-casting components.

Key words: mega-casting, model reduction, topology optimization, surrogate model, parameter design, lightweight