Knowledge transfer-oriented deep neural network framework for estimation and forecasting the state of health of the Lithium-ion batteries

This paper proposes an efficient data-driven framework for estimating and forecasting the state of health (SOH) Lithium-ion (Li-ion) batteries. The proposed is established upon a deep neural network (DNN) model, knowledge transfer asset, autoregressive integrated moving average (ARIMA) model. property reduces required data training model hence approach becomes fast good fit SOH Li-ion Among various possibilities, most features are picked by Pearson correlation coefficient least absolute shrinkage selection operator (LASSO) regression. To suppress existing noises, Savitzky-Golay filter applied to signals. allows use limited portion dataset (about 25 %) phase guarantees high accuracy (almost 96 estimation according determination. Mean squared error (MSE) estimations 0.00075 which small enough trust on results. MSE not only during via % measured, but also after 20 30 calculated as well. Training results in great downfall performance with 26.6 rise value. Surprisingly, does add any noticeable study confirms that learning DNN combination could achieve dramatic reduction purpose.