State-of-charge estimation of lithium iron phosphate battery using extreme learning machine

Tracking the state-of-charge (SOC) is important for battery applications. A novel SOC estimation algorithm based on extreme learning machine (ELM) is proposed. Compared with the traditional neural network method, the ELM simplifies the computation procedure and shortens the learning time. A typical model of ELM is built for an 180Ah/3.2V lithium iron phosphate battery, and data acquired from cell discharge experiments under a series of current are contributed to model training and predicting. To evaluate the effect of ELM on SOC estimation, the back propagation (BP) neural network and the support vector machine (SVM) are taken into comparison. The results show that the ELM method provides a more accurate SOC. What is more important is that the speed of network training is improved greatly.