Multi-Objective Optimization of Control Parameters for Hybrid and Electric Vehicles Using 1D CAE Model

Since the operation of the powertrain system and the engine speed and torque are determined in the ECU in hybrid vehicles, control parameters in these vehicles are more sensitive to a variety of performance factors than those employed in conventional vehicles. The three performance factors acceleration performance, NVH and fuel consumption in particular are in a tradeoff relationship, the calibration of control parameters in order to satisfy these performance targets entail considerable development costs. Given this, it is possible to increase the efficiency of hybrid vehicle development by determining Pareto design solutions for the three performance factors via multi-objective optimization using CAE, and selecting target performance and control parameters based on these Pareto design solutions. However, the prediction time necessary for CAE models of the three performance factors and the number of design variables involved in the creation of control parameters data represent issues in the ability to perform multi- objective optimization within a computation time that makes it possible to apply the technique in the product development process. In relation to the former issue, using an “Quietness map(Qmap)” formulated from actual vehicle measurements rather than 3D CAE for the prediction of NVH performance made it possible to conduct predictions for all three performance factors simultaneously in real time or less. In relation to the latter issue, the application of morphing technology reduced the number of design variables by an average of 98.3%. The combination of 1D CAE models, morphing technology, and optimization algorithms has created a system for the multi-objective optimization of control parameters, making it possible to calculate 20,000 patterns of control parameters necessary for convergence of the Pareto design solutions in 4.5 days.