A Modeling Method of Whole Vehicle Electrical Balance Simulation System Based on Neural Network Training

Abstract

With the rapid development of automotive electronic and electrical technology, the electronic configuration of automotive systems is gradually increased, resulting in a serious imbalance between the power generation and consumption in the vehicle power network. Vehicle electrical balance test is to verify whether the vehicle power system matching status can meet the design requirements. The traditional vehicle electrical balance test can only be carried out after the electronic and electrical functions of the whole vehicle are complete and the electronic and electrical system has no replacement parts. This paper proposes a modeling method of vehicle electrical balance simulation system, which can intervene in part of the test in the early stage of design, and play an aided guidance role for the vehicle power system design. The simulation system is put up through Simulink. The neural network module is used for data training, and the state-flow module is used to realize the running condition circulation which is established from the combined running condition of battery, generator and electric equipments. And a commercial vehicle of FAW is used for test verification.