Research On Robust Multi-objective Optimization Of Vehicle Frontal Collision Based On Combined Agent Model

China’s car ownership has shown an explosive trend with the development of economic level and the improvement of road traffic conditions,and it has also brought a series of security issues.The problem of vehicle crashworthiness and safety is very complicated.The safety of the vehicle structure in the event of a traffic accident has a very important relationship with the personal safety of the driver and passenger.Uncertainty caused by processing technology,material parameters and fixture position during actual manufacturing has a direct impact on the performance of automotive panels,which affects the collision safety of automobiles.Therefore,the multi-objective robustness optimization design of the key energy absorbing panels for frontal collision of vehicles is of great significance for improving the collision safety of automobiles.In this paper,based on the safety of frontal collision safety,the requirements of robustness and light weight are considered comprehensively.Based on the hybrid agent model,the multi-objective optimization method is studied to improve the structural crashworthiness of the frontal collision of the car.The main research contents of the thesis are as follows:Firstly,the vehicle frontal collision model is simulated by finite element method.The simulation results of vehicle displacement,vehicle beam velocity and acceleration,engine lower acceleration and rigid wall reaction force are compared with experimental results to verify its effectiveness.Sensitivity analysis of key energy-absorbing panels on the front end of the vehicle selects design variables,selects the peak acceleration of the B-pillar,the amount of pedal intrusion and the mass of the vehicle as the optimization targets,and strives to achieve the goal of minimum peak acceleration of the E-pillar,minimum pedal intrusion and minimum vehicle mass.In order to improve the optimization efficiency and reduce the cost,the agent model is used instead of the finite element model for analysis and calculation.In order to ensure the accuracy of the vehicle multi-objective optimization model,different sampling strategies are adopted to study the optimization objectives under different agent model methods.Accuracy,after analyzing the error,established a hybrid agent model with higher precision.Based on the above hybrid agent model,multi-objective optimization of vehicle frontal collision is carried out.The multi-objective particle swarm optimization algorithm(MOPSO)is used for deterministic optimization,which greatly reduces the peak acceleration of the B-pillar,the amount of pedal intrusion and the quality of the vehicle,and improves the safety of the vehicle.Then the approximate model reconstruction method is used to reconstruct the test sample and the hybrid agent model to improve the accuracy of the model.Robust optimization of optimization targets is based on the agent model,and the optimization results are compared with the deterministic optimization.The results show that deterministic optimization and robustness optimization have a good effect compared to the original model.Due to the influence of uncertainty,the robustness optimization effect is not deterministically optimized,but the quality performance of the product is reduced.Improve the robustness of the product.