Abstract:Reliable life prediction technology is of great significance in ensuring a safe and efficient lifetime of lithium batteries. However, the traditional health factors such as battery capacity are the effects of battery aging rather than the direct causes, which cannot directly reflect the internal degradation mechanism information of the battery, and the prediction accuracy is easily affected by the working environment of the battery. Electrochemical impedance spectroscopy (EIS) data can more directly reflect the internal mechanism information of the battery, which includes a wealth of battery aging information. In order to deeply investigate the mapping relationship between impedance spectrum and remaining useful life (RUL) of lithium batteries, EIS method is employed to obtain the impedance and phase of lithium batteries under different health states and temperatures, as well as explore the visualization and quantification of impedance frequency response of lithium battery. Furthermore, the mapping relationship between RUL and the lithium battery impedance is investigated in a full impedance spectrum at different temperatures. It is found that, as lithium battery aging, its negative imaginary parts impedance increases significantly, especially in the middle of the frequency band, and has no significant dependence on temperature. While the real part impedance shows an obvious dependence on temperature. Therefore, it is found that the negative imaginary parts impedance of impedance spectrum has a well-fit characterization ability for RUL of a lithium battery. In this paper, a fusion neural network model of Conv1d-SAM (one-dimensional convolutional neural network-self-attention mechanism) was established with negative imaginary part impedance as input factor to predict battery RUL. The predicted results show that Conv1d-SAM has improved accuracy and stability in RUL prediction, and the mean absolute error function of the proposed model is increased by 72% compared with the latest published method.

Enhanced Temporal Prediction of Electrochemical Impedance Spectroscopy Using Long Short-Term Memory Neural Networks

Prediction of Fuel Cell Performance Degradation Using a Combined Approach of Machine Learning and Impedance Spectroscopy

Simultaneous Prediction of Impedance Spectra and State for Lithium-Ion Batteries from Short-Term Pulses

Application of Electrochemical Impedance Spectroscopy for prediction of Fuel Cell degradation by LSTM neural networks

Lithium-ion Battery Remaining Useful Life Prediction with Long Short-term Memory Recurrent Neural Network

Battery Degradation Prediction Against Uncertain Future Conditions with Recurrent Neural Network Enabled Deep Learning

Remaining useful life prediction of lithium-ion batteries via an EIS based deep learning approach

State of health and remaining useful life prediction for lithium-ion batteries based on differential thermal voltammetry and a long and short memory neural network

Prediction of impedance responses of protonic ceramic cells using artificial neural network tuned with the distribution of relaxation times

Analysis of Electrochemical Impedance Data: Use of Deep Neural Networks

Exploiting the Electrochemical Impedance Spectroscopy Frequency Profiles for State-of-Health Predication of Lithium-Ion Battery

Deep neural network battery impedance spectra prediction by only using constant-current curve

Long Short-Term Memory Recurrent Neural Network for Remaining Useful Life Prediction of Lithium-Ion Batteries

State‐of‐health prediction for lithium‐ion batteries via electrochemical impedance spectroscopy and artificial neural networks

Remaining useful life prediction of lithium-ion batteries based on attention mechanism and bidirectional long short-term memory network

Multi-Timescale Lifespan Prediction for PEMFC Systems Under Dynamic Operating Conditions

A Hybrid Drive Method for Capacity Prediction of Lithium-Ion Batteries

Prognostics of lithium-ion batteries health state based on adaptive mode decomposition and long short-term memory neural network

Combining empirical mode decomposition and deep recurrent neural networks for predictive maintenance of lithium-ion battery

Bayesian impedance deconvolution using timescale distribution for lithium-ion battery state estimation