To ensure the secure and stable operation of lithium-ion batteries, the state of health (SOH) and the remaining useful life (RUL) are the critical state parameters which must be estimated precisely. Here, a joint SOH and RUL estimation approach based on an improved particle swarm optimization extreme learning machine (PSO-ELM) is proposed. The approach adopts Pearson coefficients to screen multivariate information of the discharge process as health indicators and uses them as inputs to enable accurate estimation of SOH and RUL prediction of lithium-ion batteries on the basis of the PSO-ELM model. The validity of the model is demonstrated by the NASA lithium-ion battery data set: the maximum root mean square error (RMSE) of SOH estimation of the tested battery is 0.0033, the maximum RMSE of its RUL prediction is 0.0082, and the maximum absolute error of RUL prediction is one cycle number. In comparison with the prediction results of the traditional extreme learning machine, the proposed optimized model estimates the SOH of lithium-ion batteries and RUL with relatively high accuracy.

Incisively estimating the state of charge (SOC) of lithium-ion batteries is essential to ensure the safe and stable operation of a battery management system. Neural network methods do not depend on a specific lithium-ion battery model and are able to mirror the lithium-ion battery’s nonlinear relationships, thus receiving widespread attention; however, traditional neural network methods exhibit a long training time and low accuracy in estimating SOC. This paper presents an original algorithm of an improved particle swarm optimization (IPSO) extreme learning machine (ELM) neural network, improving the particle swarm algorithm using nonlinear inertia weights to enhance the global optimization seeking capability of ELM for solving the problem of poor precision of previous battery SOC estimation. The lithium-ion battery voltage and current are the input variables of the model, while SOC is used as the output variable. The results of the experiments revealed that the root-mean-square estimation errors of the proposed IPSO-ELM algorithm for SOC are within 0.31, 0.32, and 0.14% of the root mean square under the hybrid pulse power characteristic (HPPC), the Beijing bus dynamic stress test (BBDST), and the dynamic stress test (DST) operating conditions. Compared with the prediction results of the PSO-ELM and ELM neural networks, the simulation results prove that the SOC optimization method in this paper possesses superior precision and overcomes the shortcomings of traditional neural networks.

No abstract