Life prediction of lithium-ion batteries based on stacked denoising autoencoders

• Published in: Reliability engineering & system safety Vol. 208; p. 107396
• Main Authors: Xu, Fan, Yang, Fangfang, Fei, Zicheng, Huang, Zhelin, Tsui, Kwok-Leung
• Format: Journal Article
• Language: English
• Published: Barking Elsevier Ltd 01.04.2021; Elsevier BV
• Subjects: Battery cycles; Clustering; Clustering by fast search; Computer architecture; Deep learning; Feature extraction; Life cycles; Life prediction; Lithium; Lithium-ion batteries; Lithium-ion battery; Noise reduction; Predictions; Product safety; Quality assessment; Rechargeable batteries; Reliability engineering; Stacked denoising autoencoder; Life prediction; Lithium-ion battery; LSTM; Stacked denoising autoencoder; DAE; SVM; RMSE; SOH; BP; Deep learning; MAE; CFS; Clustering by fast search; SDAE; RF; SAE
• ISSN: 0951-8320; 1879-0836
• DOI: 10.1016/j.ress.2020.107396

•Introduce the CFS clustering model for feature selection.•Using the SDAE for life prediction at early life stage.•Extracting the original features using discharge capacity and temperature only. Accurate life prediction of lithium-ion batteries is important to help assess battery quality in advance, improve long-term battery planning, and subsequently guarantee the safety and reliability of battery operations. In this study, a deep learning-based stacked denoising autoencoder (SDAE) method is proposed to directly predict battery life by extracting various battery features. In general, the SDAE contains autoencoders and uses a deep network architecture to learn the complex nonlinear input-output relationship in a layer-by-layer fashion. Many features enabling the life prediction of lithium-ion batteries are extracted from discharge temperature and voltage curves. As redundancies in these features may result in poor prediction accuracy, a clustering by fast search (CFS) method is adopted to filter and select essential features. The CFS selects effective features by aggregating the types of battery features into clusters. All selected features are then fed into the SDAE to predict battery life cycle. Key hyperparameters are investigated, such as the number of iterations, the learning rate, and the denoising rate of the SDAE network. Experimental results show that the proposed selected-features-based deep learning method can provide more accurate and efficient battery life predictions with less fluctuation than the method without feature selection.