Prediction Of The Remaining Useful Life Of Lithium-Ion Batteries Based On An Empirical Mode Approach With Artificial Neural Networks

Abstract

Forecasting future capacities and estimating the remaining useful life, while incorporating uncertainty quantification, poses a crucial yet formidable challenge in the realm of battery health diagnosis and management. In this study, a data-driven model based on artificial neural networks (ANN) and signal decomposition techniques including Empirical Mode Decomposition (EMD), Ensemble Empirical Mode Decomposition (EEMD), and Empirical Wavelet Transform (EWT) is presented to predict the capacity value of lithium-ion batteries. Signal decomposition was performed using the discharge voltage values for four different batteries. A total of 22 features were obtained. The features of the signal decomposition methods were evaluated separately as well as hybrid approaches. Mean Squared Error (MSE), Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE) performance metrics are used in the proposed method and the values obtained are 3.67×10-6, 0.001351 and 0.002311, respectively. According to the findings, the hybrid model proposed demonstrated positive results in terms of accuracy, adaptability, and robustness.