State of Charge Estimation Using Multi-layer Neural Networks Based On Temperature

Lithium-ion batteries are generally used in electric vehicles, mobile phones, and lap-tops. Such batteries demonstrate advantages such as environmental-friendliness, high energy density, and long life. However, if not continuously monitored, battery overcharging and over-discharging may occur. Overcharging causes fire and explosion casualties, and overdischarging causes a reduction in the battery capacity and life. In addition, the internal resistance of such batteries varies depending on the external temperature of the batteries, and as the temperature decreases, the capacity of the batteries decreases as well. In this paper, we propose a method for estimating the state of charge (SOC) using a neural network model best suited for the external temperature of such batteries. Experimental data to verify the proposed method were obtained through a discharge experiment conducted using a vehicle-driving simulator. The experimental data were provided as inputs to multi-layer neural network (MNN). The MNN models were trained and optimized for specific temperatures measured during the experiment, and the SOC was estimated by selecting the most suitable model for a temperature. The experimental results revealed that such an estimation of the SOC was better than that using conventional methods.