Multi-Objective Lightweight Design Method for Lower Control Arms Based on Multi-Axial Random Vibration Fatigue Life

To reduce the computational cost of the multi-objective optimization process and improve the accuracy of fatigue life prediction for the lower control arm (LCA) of a vehicle under multiaxial random vibration, this paper focuses on the LCA of a MacPherson suspension in a specific vehicle model and proposes a multi-objective lightweight design method based on multiaxial random vibration fatigue. This method combines the Latin hypercube sampling (LHS) technique, Kriging surrogate modeling, and the second-generation non-dominated sorting genetic algorithm (NSGA-II). First, static and dynamic analyses are conducted to extract the design parameters required to meet the design specifications of the reference LCA. Subsequently, the LHS technique is employed to obtain 50 sample points, which are used to construct the sample space. The Kriging method is then applied to build surrogate models that capture the relationship between design variables and various responses. Finally, the NSGA-II multi-objective genetic algorithm is utilized to obtain the optimized solution. Considering 1.2 times the safe driving distance, the optimal solution with the minimum LCA mass is selected from the Pareto frontier. The optimization results show that compared to the initial LCA model, the mass of the optimized model is reduced by 13.06%, the fatigue life is increased by 47.48%, and the maximum displacement and maximum stress are reduced by 1.78% and 4.31%, respectively. Additionally, the first-order modal frequency decreases by 4.60%.