Extreme learning machines improved by differential evolution algorithms to estimate the state of health of lithium-ion battery

The accurate estimation of the health state of lithium-ion batteries is of great importance for battery health management. In this paper, a model for estimating the health state of lithium-ion batteries is developed by differential evolution algorithm and extreme learning machine. Firstly, through cyclic ageing tests, the five health state related features are extracted from the battery charging curve and incremental capacity curve; secondly, the correlation between the health state of lithium-ion batteries and the features is obtained by using grey relational analysis; then, the five selected health features are used to construct an extreme learning machine based lithium-ion battery health state model; finally, a differential evolutionary algorithm is used to optimize the parameters of the extreme learning machine model parameters to improve the accuracy of Lithium-ion battery capacity estimation. By comparing several other algorithms, the results show that this model is more accurate than other methods.