Component sizing of a series hybrid electric vehicle through artificial neural network

Abstract

Series hybrid electric vehicles are the intermediate technology between gasoline-powered vehicles and full electric vehicles to suppress the emission and global warming issues. In the present study, the component sizing of a series hybrid electric vehicle including a high voltage battery, the combustion engine, and the electric motor is investigated employing experimental data and an artificial neural network. Despite previous studies, the experimental data is implemented to generate a robust artificial neural network model. About 3000 data series are implemented to train the artificial neural network. The results reveal that the 1.44 kWh for the high voltage battery is sufficient for all driving conditions. This battery size can overcome many challenges of this type of vehicle both in price and packaging. Three different working point of 6.7, 12.2 and 22 kW is proposed for a combustion engine in the best efficiency zone for urban, accelerating and highway driving conditions, respectively. The required power of the electric motor is 80 kW to support the drivability and vehicle acceleration in different conditions. This novel method can be extended to size the component of hybrid electric vehicles for different segments and driving conditions.