SOC Estimation of Li-ion Batteries With Learning Rate-Optimized Deep Fully Convolutional Network

In this letter, we train deep learning (DL) models to estimate the state-of-charge (SOC) of lithium-ion (Li-ion) battery directly from voltage, current, and temperature values. The fully convolutional network model is proposed for its novel architecture with rate optimization strategies. capable estimating SOC at constant varying ambient on different drive cycles without having be retrained. also outperformed other commonly used DL such as LSTM, GRU, CNN an open source Li-ion dataset. achieves 0.85% root mean squared error (RMSE) 0.7% absolute (MAE) 25 °C 2.0% RMSE 1.55% MAE (-20-25 °C).