A component sizing prediction study for a series hybrid electric vehicle based on artificial neural network

In the present study, the predictive tool based on an artificial neural network is developed by means of the experimental data of two series hybrid electric vehicles. The experiments have been conducted on different driving conditions, including highways, traffic, and combined driving conditions. Then, the artificial neural network is developed to predict an arbitrary series hybrid electric vehicle's required power. The instantaneous required power is divided into dynamic and steady power to size the combustion engine, electric motor, and high voltage battery of the series hybrid electric vehicle. The effects of different ambient conditions (including ambient temperature and altitude), the inverter and high voltage battery efficiencies, and the coast-down coefficients on the components sizing of the series hybrid electric vehicle are then investigated in different driving conditions. The results revealed that the maximum instantaneous power of the electric motor is associated with rapid acceleration in low-speed conditions, and the suburban driving route determines the combustion engine's maximum power. Notably, the Worldwide Harmonized Light-duty vehicles Test Cycle is the most comprehensive among the available driving cycles, and most of the components' sizes are determined by this cycle except the combustion engine's maximum power. It is also realized that the cycle-wise investigation can be summarized into the Isfahan-Tehran route and Worldwide harmonized Light-duty vehicles Test Cycle calculations.