Optimization of Engine Performance Based on Process Simulation Combined with GA-BP and NSGA-II

• 1、School of Automobile Engineering, Chongqing University, Chongqing 400044

Abstract：In order to solve the multi-objective optimization problem of gasoline engines. The one-dimensional engine simulation software GT-POWER was used to generate the training data and test data needed by the GA-BP neural network model, which was designed by the Latin hypercube algorithm. And then GA-BP neural network model was compared the regression effect with BP neural network model. Finally, NSGA-II was used to multi-objective optimization based on a well-trained GA-BP neural network model. The calculation results show that the GA-BP neural network model could accurately predict the performance of the gasoline direct injection turbocharged engine, and optimize the main influence parameters of the gasoline direct injection turbocharged engine through NSGA-II. The average improvement of engine torque is 7.53%, and the average promotion of engine BSFC is 3.84%.

Keywords： gasoline direct injection turbocharged engine performance optimization GA-BP NSGA-II multi-objective optimization