Remaining useful life prediction of lithium-ion batteries based on stacked autoencoder and gaussian mixture regression

Lithium-ion batteries have been widely applied in energy storage systems and electric vehicles (EVs), the remaining useful life (RUL) prediction is one of critical technologies for prognostics health management. However, high accuracy RUL with reliability biggest bottleneck. To improve adaptively extract indirect indicators (HIs), framework based on stacked autoencoder Gaussian mixture regression (SAE-GMR) proposed. Firstly, HIs are extracted from charging discharging data, gray relation analysis (GRA) adopted to analyze capacity. In this paper, SAE neural network proposed reduce dimensions noise battery obtain a syncretic HI. Then, GMR model estiblished not only prediction, but also describe reliability. Finally, method compared esixting methods,which shows that has superiority other methods.