Novel PI Controller and ANN Controllers-Based Passive Cell Balancing for Battery Management System

The cycle life and efficiency of a battery pack get enhanced by employing an intelligent supporting system with it called the Battery Management System (BMS). A novel Proportional Integral (PI) controller and an Artificial Neural Network (ANN)-based controller for controlling the Passive Cell Balancing (PCB) technology have been implemented for BMS. The Scaled Conjugate Gradient, Bayesian Regularization, and Levenberg Marquardt algorithms of ANN are employed individually for the control operation. Each of the techniques is executed and analyzed in a MATLAB Simulink environment. With PI controllers, improved performance of cell balancing is achieved as compared with the conventional PCB method without employing controllers. Whereas, on implementing the ANN-based controllers, more improvement in the results occurs in terms of SoC balancing, voltage balancing, power dissipation, heat dissipation, and temperature rise across the bleeding resistors connected to each cell. The average current flowing across the bleeding resistors decreases from 1.5620 A to 0.8756 A, and to 0.2032 A on utilizing the conventional PCB, the novel PI-controller, and ANN-controller-based PCB techniques respectively, indicating better SoC balancing. Consequently, the average power dissipated decreases from 2.9807 W to 1.1275 W and 0.0838 W, while the average heat dissipated decreases from 2.0224 KJ to 0.1921 KJ and 0.0052 KJ. Thus, the average temperature rise also reduces from 2.3541 °C to 1.0331 °C and 0.1091 °C. Hence, the efficiency further gets enhanced by employing the ANN-based controllers for their satisfactory use in BMS. So, these technologies ensure improving the performance and driving range of electric vehicles effectively.

I. Introduction

Electric Vehicles (EVs) are eco-friendly compared to fossil fuel-based vehicles, despite accounting for the manufacturing and charging processes of a battery. Researchers have established that despite considering the carbon emissions of generating electricity from the conventional resources for charging the battery of an EV, the carbon release on driving an EV is much lesser throughout its life span compared to a fossil fuel-based vehicle [1], [2], [3]. The battery pack of an EV usually consists of Lithium-ion cells. These cells are sensitive to overcharging or over-discharging which may hasten the cell degradation process and in extreme cases, the consequence can be of catching fire and exploding [4], [5]. Mostly there are small variations in the capacity, internal impedance, State of Charge (SoC), self-discharge rate, and temperature characteristics among the cells in a battery pack, although they are of a similar model, manufactured in the same batch and purchased from the same maker [6]. These variations are due to their different internal chemistry kinetics, which can result in a variance in the cell voltages eventually [7], [8].