Selection of three RC Branches in Equivalent Circuit Model of Lithium-ion Batteries for Improved Accuracy

Battery management system (BMS) comprises of an electronic circuitry that monitors the battery operation to ensuretemperature regulation, cell balancing, protection against overcharging and deep discharging, etc. This results in increasedbattery life, prevents degradation of the health of battery, prevention of fire hazards and increased safety. BMS is critical to theperformance and wider acceptance of electric vehicle (EV) technology. Accurate estimation of the state of charge (SOC) of thebattery is necessary for the precise operation of BMS. Kalman filters can be employed for SOC estimation. This paper presentsunscented Kalman Filter (UKF) based SOC estimation for a 3100mAh, 3.7V lithium iron phosphate cell, which is employedfor developing a battery pack for EVs. Four different Thevenin models of the cell, using (i) zero RC branch, (ii) one RCbranch, (iii) two RC branches, and (iv) three RC branch, are considered in this work. With the increase in the number of RCbranches, the accuracy of the model increases. However, consequently the computational burden also increases. The UKFbased SOC estimation with the four different RC model is analysed with the help of the model developed inMATLAB/SIMULINK. The analysis reveals that the root mean square error in SOC estimation is lowest with three RC model,resulting in more accurate SOC estimation. This is achieved without significant increase in the execution time as recorded withthe Raspberry Pi based implementation of the UKF algorithm for each model.