Machine Learning

Transformer-Based Battery Prognostic And SoH Prediction

Machine Learning

Transformer-Based Battery Prognostic And SoH Prediction

The deep learning transformer model can provide accurate and stable battery SoH prediction SoH estimation of lithium-ion batteries is crucial for ensuring the reliability and safety of battery operation while keeping maintenance and service costs down in the long run. This research proposed a novel ML prognostic model based on a convolutional neural network (CNN)-Transformer framework. The Pearson correlation coefficient (PCC) and Principal correlation analysis (PCA) are used in pre-processing for feature selection. The NASA battery dataset is used as a training and testing dataset. The testing results show this CNN-Transformer model can predict the battery SoH with high accuracy and stability.

Accuracy:

Efficiency:

A convolutional neural network predicts battery SoH from the NASA dataset to within 1%.

Academia | P. 231 Source: Gu, Xinyu, K. W. See, Penghua Li, Kangheng Shan, Yunpeng Wang, Liang Zhao, Kai Chin Lim, and Neng Zhang

Extracted from the PDF; charts and images may read imperfectly.