动力传动系统
自动化
人工神经网络
控制工程
汽车工程
电动汽车
计算机科学
工程类
优化设计
扭矩
功率(物理)
人工智能
机器学习
机械工程
物理
量子力学
热力学
作者
Kihan Kwon,Sang-Kil Lim,Dongwoo Kim,Kijong Park
标识
DOI:10.1016/j.etran.2023.100267
摘要
Many studies have been conducted on various powertrain systems, such as multi-motor, multi-speed, or both, to enhance the energy efficiency and dynamic performance of electric vehicles (EVs). This study developed an automated design program to obtain the optimal design of EVs for various powertrain systems. The program consists of an EV simulation and artificial neural network (ANN) modeling and optimization tools. The EV simulation tool employs an integrated EV model that can analyze the efficiency and performance of various powertrain systems in a single environment. The ANN modeling and optimization tool first constructs an ANN model, and then performs optimization using the ANN model to address excessive computational efforts arising from the multi-objective genetic algorithm. This study verified the developed program by conducting analysis and optimization of five powertrain systems with the same EV specifications. A multi-objective optimization problem was formulated by considering the design variables as the torque distribution between the motors and gear shifting patterns and ratios of the transmission, and the objectives as the energy consumption and acceleration time. A comparison of the optimization results among the five powertrain systems quantitatively showed the positive effects of the multi-motor and multi-speed powertrain systems. Furthermore, the ANN-based multi-objective optimization process allowed for the efficient determination of the optimum design solutions for the proposed EV powertrain systems. Consequently, these results demonstrated the effectiveness of the automation program in rapid decision-making on EV powertrain system configurations, satisfying each designer’s requirements.
科研通智能强力驱动
Strongly Powered by AbleSci AI