SoC Estimation of Battery in FCHEVs Using Reformulated Constrained Unscented Kalman Filter

Abstract

The most essential requirement for the battery management system (BMS) in fuel cell-based hybrid electrical vehicles (FC-HEVs) is an accurate estimation of state of charge (SoC) of battery. To estimate the SoC of the battery several methods have been suggested, each of which is a tradeoff between computational complexity and battery accuracy. For massive energy density, power density, and for the long life of the battery along with strong environmental adaptability and high cell voltage Li-ion battery became very popular. Due to the chemical mechanism of LI-ion batteries direct extraction of SoC is not possible. In this paper, Ampere-hour counting, Extended Kalman filter (EKF), and proposed Reformulated Constrained Unscented Kalman filter (RCUKF) techniques are used to estimate the battery SoC as well as the terminal voltage of Li-ion batteries. The proposed RCUKF technique is used to eliminates the errors of time-delayed measurements. MATLAB simulations are performed for various characteristics of the Li-ion battery like charging, discharging by considering different charge rates (C-Rates). Comparison of state of charge of battery is estimated by Ampere-hour [A-h] counting method, EKF method, RCUKF method for various initial SoC is presented also the functional relationship of open-circuit voltage and state of charge is observed.