Pure hardware design and implementation on FPGA of an EKF based accelerated SoC estimator for a lithium‐ion battery in electric vehicles

Authors:
Open Access
Publisher website
Google Scholar
Abstract
The State of Charge monitoring is a key operation for the safety and reliability of electric vehicles. The Extended Kalman Filter (EKF) is a powerful algorithm that enables an accurate estimation of the SoC under noisy measurement conditions. However, its real‐time implementation needs a higher computation burden and lots of resources. To achieve a faster run‐time and an accelerated SoC estimation, this paper proposes a pure hardware implementation technique on FPGA of the EKF algorithm without soft‐core. Moreover, a new protocol is designed to ensure a flexible data transfer between the FPGA and the external interface. The effectiveness of the proposed hardware architecture is validated on a Hardware‐In‐the‐Loop platform that consists of a Xilinx FPGA board Zynq‐7000 and LabVIEW. The obtained SoC average estimation errors are close to the literature outcomes and are equal to 0.84% in continuous load profile and 5.34% in dynamic load profile. Moreover, the average prediction error of the battery voltage is equal to 34.9 mV. The results are obtained with the benefits of low energy consumption and high operation speed to run one iteration of the EKF core. Indeed, the FPGA consumes only 530 mW and needs a run‐time equal to 1.025 μs at a 100 MHz clock frequency.

This publication has 31 references indexed in Scilit: