Estimation of Lithium‐Ion Battery State of Charge for Electric Vehicles Using an Adaptive Joint Algorithm

Abstract

As a new means of transportation, electric vehicles (EVs) have a lot of potential. On the other hand, EVs that employ lithium‐ion batteries face certain difficulties in forecasting the battery's health and remaining useful life. This paper uses the adaptive joint algorithm approach to calculate the battery's online parameters and accurate state of charge (SOC). To establish the battery online parameters, the forgetting factor recursive least square (FFRLS) technique is utilized, and the extended Kalman filter (EKF), unscented Kalman filter (UKF) are employed to estimate accurate SOC. Compared to the EKF/UKF method, the joint algorithm (FFRLS‐UKF) approach produces better results. The results are validated using the urban dynamometer driving schedule cycle and the ECE extra‐urban driving cycle (low powered vehicles) to determine the performance of the proposed algorithm. The error of the estimated SOC has fallen from 3.3% to 2%. The proposed adaptive joint algorithm has substantially improved the system's accuracy and provides better results than the EKF/UKF technique. Furthermore, the random variable noise is also supplied to the test data to ensure that the proposed method is robust.

An Adaptive State of Charge Estimation Method of Lithium-ion Battery Based on Residual Constraint Fading Factor Unscented Kalman Filter

“…In contrast with the FFRLS and FMRLS methods in Ref. [32][33], and the adaptive identification capability of the model became stronger. The UKF transformation process was derived on the basis of EKF, and the introduction of the residual constraint fading factor was done for the UKF method improvement, reducing the influence of system noise and observation noise.…”

“…In Ref. [32], adopting the forgetting factor recursive least squares (FFRLS) technique, and combining extended Kalman filter (EKF) with untraced Kalman filter (UKF) for an accurate estimation of battery charge state, SOC estimation accuracy can be reached as high as 2%. In Ref.…”

State of charge estimation of lithium‐ion battery based on extended Kalman filter and unscented Kalman filter techniques

“…The hybrid technique employs two SOC estimation methods, one of which serves as a reference for error estimation and the other of which is utilized to estimate the SOC. The CCM and KF method and NNM and adaptive filter methods come under hybrid methods for SOC estimation 24,25 . Xu et al 26 propose a hybrid technique that combines a gated recurrent unit neural network with the KF method.…”

“…The CCM and KF method and NNM and adaptive filter methods come under hybrid methods for SOC estimation. 24,25 Xu et al 26 propose a hybrid technique that combines a gated recurrent unit neural network with the KF method. In other aspects, hybrid methods face high computational costs, exact battery information, and are sensitive to training data.…”

An adaptive spherical square‐root double unscented Kalman filtering algorithm for estimating state‐of‐charge of lithium‐ion batteries

“…33 However, the extended Kalman filtering algorithm will have accumulated errors, it will not only ignore the effects of temperature factors on the battery but also the convergence speed is slow. 34 Currently in the latest developments in the UKF algorithm, Wu Tiezhou and Hu Liquan have studied an improved electromotive (EMF) battery equivalent model, 35 and they have used an open circuit voltage method to obtain the SOC real value, then they combined the UKF algorithm and the EMF battery equivalent model to estimate the SOC value, and the result error can be reached 5%; Hou, J has proposed a novel variational Bayes-based unscented Kalman filter (VB-UKF) algorithm to simultaneously estimate the SOC and current of nonlinear lithium-ion battery systems, 36 the status estimate of the nonlinear system with unknown input has been solved, and the result showed the SOC root mean square errors of VB-UKF can reach 3%; Sakile, R and Sinha, UK have taked the forgetting factor recursive least square (FFRLS) to estimate the SOC with the UKF algorithm, 37 and the random variable noise is also supplied to the test data, the result showed the SOC errors of FFRLS-UKF can reach 2%; Liu Jichao aimed at the problem of multi-parameter identification of battery model, a parameter identification method based on improved particle swarm optimization (PSO)algorithm was proposed, 38 then it was used to estimate the SOC with the UKF algorithm, and its estimation results show that the relative error of the method is about 3%. Although the existing algorithms have been improved in the traditional UKF algorithm, the SOC estimation effect is improved to varying degrees, and its estimation accuracy has a large improvement space.…”