SOC estimation design of lithium battery based on FPGA

Aiming at the problems of existing methods for estimating state of charge (SOC) of lithium battery, a novel SOC estimation scheme based on Field Programmable Gate Array (FPGA) with high parallelism was proposed in this paper. The second-order RC circuit is selected to be equivalent to the battery model, and the variable forgetting factor recursive least square (VFFRLS) and extended Kalman filter (EKF) algorithm is used to identify the parameters online and estimate the SOC of the lithium battery. The EKF algorithm flow is re-planned according to the high parallelism of FPGA, so that some algorithms can run in parallel, which greatly reduces the algorithm operation time and improves the SOC estimation efficiency. The modularization design idea is used to design the Verilog program structure in FPGA. The finite state machine is used to control the logical sequence between each step, so that each module runs in order and reduces the logical complexity. The pulsating integral matrix multiplier is packaged and modified, so that a single pulsating integral matrix multiplier can deal with different dimensional matrix schemes, thus reducing FPGA resource consumption. The time series simulation and experimental verification show that this scheme has high real-time performance and accuracy in estimating SOC of lithium battery.