Battery Management System for Electric Vehicles

A research on battery management system (BMS) for electric vehicles (EVs) was carried out in this thesis. First, a simplified electrical model- based was employed to estimate the battery's state of charge (SoC) due to its capability to adapt to the extensive dynamic operating conditions in online EV application and its accuracy in reflecting the electrochemical properties and charging/discharging response as well. The parameters of the equivalent circuit are identified in real time via a recursive least squares (RLS) algorithm and then they are used in reconstructing the open circuit voltage (OCV) of the LiFePO4 battery by the proposed fading Kalman filter (FKF). The main purpose of the model is to estimate the OCV of the battery from its measurements (i.e., voltage, current, and temperature) and then the SoC is determined via an OCV-SoC curve or lookup table (LUT). The main reason for using cascaded linear filtering stages for both circuit parameters and OCV estimation is to avoid nonlinear modeling to reduce the computational effort required for real time application. Standard Urban Dynamometer Driving Schedule (UDDS) and real vehicle driving cycles are implemented representing the battery load for the validation of SoC estimation by the proposed technique. Second, SoC-based self-equalization algorithm was designed for a battery pack consisting of 24 cells connected in series. The equalizer circuit topology was also developed concurrently with the rest of the hardware modules constitute the BMS hardware platform. Finally, the BMS software platform was developed, incorporating the BMS firmware with SoC and self-equalization algorithms, data management system, and web application.