Electric Bus Frame Optimization for Side-Impact Safety and Mass Reduction Based on the Surrogate Model Method

The body strength, stiffness and crashworthiness are the key aspects for the mass reduction of the commercial bus body frame. Heavy computation cost is one of the critical problems by the finite element (FE) method to accomplish a high-efficient multi-objective optimizing design. Starting from this point, in this paper, the surrogate model method is adopted to optimize the electric bus frame to reduce the mass as possible while guaranteeing the side-impact strength. The optimizing objective comprises the total mass and side-impact intrusion while the performances of static strength and stiffness in bending and torsion conditions are chosen as the constraints in optimization. First, an FE model is developed to perform the static strength analysis, modal analysis and side-impact strength analysis. Nine groups of candidate variables are determined as the optimizing design variables by sensitivity analysis. Then surrogate models have been formulated based on the methods of least squares regression (LSR) and radial basis function neural network (RBFNN). The precision of the surrogate models are evaluated and validated by comparing with the FE simulation results. Based on the surrogate models the bus body frame is finally optimized by the multi-objective genetic algorithm (MOGA) method. With the optimized parameters, the performance of the body frame is evaluated by comparing with that before optimizing. It is demonstrated that the design objective of lightweight (mass reduction) has been achieved and the side-impact crashworthiness have been improved as well while guaranteeing the basic performance including the static strength and stiffness.