Intelligent switching mechanism for power distribution in photovoltaic-fed battery electric vehicles

The paper provides a quick and robust power control mechanism for electric vehicles with integrated photovoltaic panels. Traditionally, photovoltaic power is solely used to charge the battery which feeds various power loads. However, this process is inefficient due to the incessant charging and discharging losses that occur in the battery. This paper proposes a distribution of power via an intelligent switching mechanism to various accessory loads so as to reduce these losses. Furthermore, a key component of this design is to estimate the maximum power available from the photovoltaic module in arbitrary environmental conditions. To do this, a fast and accurate polynomial regression model is presented. The performance of the model has been compared with several feed-forward neural networks with different hidden layers and nodes. The feed-forward neural network has been trained using the Levenberg-Marquardt back propagation method. The entire simulation has been carried out in MATLAB and Simulink 2018a. To validate the accuracy of this system, it has verified in real time on a hardware-in-the-loop testing platform using MicroLabBox hardware controller. It is shown that the proposed polynomial regression model provides an accurate estimate of maximum power in a much shorter duration compared with the neural networks. The formulated switching mechanism results in greater final SOC as compared to traditional power distribution schemes. This allows for longer cruising range for an electric vehicle ceteris paribus.