The state of health (SOH) and remaining useful life (RUL) of lithium-ion batteries are important aspects of a cell chemistry which describe the current aging degree of the batteries from different perspectives. To improve the effectiveness and accuracy of the batteries’ health assessment models, I've implemented a method for SOH and RUL estimation of lithium-ion batteries using convolutional neural networks (CNNs), bi-directional long short-term memory (BiLSTM), and attention mechanism (AM) to build a hybrid network model for capacity estimation of lithium-ion batteries, and calculate the SOH and RUL.
Specifically, this project predicts the life of the Lithium Ion battery with an attention-based-CNN-BiLSTM based on the voltage, current and temperature of the charging cycles.
Lithium ion battery data has been taken from NASA Ames Prognostics Data Repository [1]. I've also included a notebook for initial data exploration of this dataset and an SVM model for SOC prediction.
The model uses a convolutional neural network (CNN) to carry out feature extraction and data dimension reduction. These features are used as the input of the bidirectional long short-term memory network (BiLSTM). The BiLSTM is used to learn the temporal correlation information in the local features of input time series bidirectionally. The attention mechanism puts more emphasis to key features in the input that have the most impact on the output result by assigning weights to important features, and multi-step prediction of the battery SOH is done. Overall, the convolutional layer extracts the intrinsic features of the time series data, the BiLSTM can predict the battery capacity based on the extracted features, and the AM assigns weights to those features. The structure diagram of the hybrid neural network model can be seen in the figure below [2].
[1] B. Saha and K. Goebel (2007). "Battery Data Set", NASA Ames Prognostics Data Repository (https://ti.arc.nasa.gov/tech/dash/groups/pcoe/prognostic-data-repository/#battery), NASA Ames Research Center, Moffett Field, CA
[2] Zhu Z, Yang Q, Liu X, Gao D. Attention-based CNN-BiLSTM for SOH and RUL estimation of lithium-ion batteries. Journal of Algorithms & Computational Technology. 2022;16. doi:10.1177/17483026221130598