Abstract:Accurately predicting the lifetime of lithium-ion batteries in early cycles is crucial for ensuring the safety and reliability, and accelerating the battery development cycle. However, most of existing studies presented poor prediction results for early prediction, due to the nonlinear battery capacity fade with negligible variation in early cycles. In this paper, to achieve an accurate early-cycle prediction of battery lifetime, a comprehensive machine learning (ML) based framework containing three modules, the feature extraction, feature selection, and machine learning based prediction, is proposed. First, by analysing the evolution pattern of various informative parameters, forty-two features are manually crafted based on the first-100-cycle charge-discharge raw data. Second, to manage feature irrelevancy and redundancy, four typical feature selection methods are adopted to generate an optimal lower-dimensional feature subset. Finally, the selected features are fed into six representative ML models to effectively predict the battery lifetime. Numerical experiments and paired t-test are conducted to statistically evaluate the performance of the proposed framework. Results show that the wrapper-based feature selection method outperforms other methods, and significantly improves the prediction performance of subsequent ML models. Both before and after wrapper feature selection, the elastic net, Gaussian process regression, and support vector machine present better performance than other complex ML prediction models. The support vector machine model combined with wrapper feature selection statistically presents the best result for battery lifetime prediction, with a root-of-mean-square-error of 115 cycles, and a R2 of 0.90. Finally, when compared with an existing work, the root-of-mean-square-error is substantially decreased from 173 to 115 cycles, by using the proposed framework.

Early prediction of lithium-ion battery lifetime via a hybrid deep learning model

A Physics-Informed Hybrid Multitask Learning for Lithium-Ion Battery Full-Life Aging Estimation at Early Lifetime

A Physics-Informed Hybrid Data-Driven Approach with Generative Electrode-Level Features for Lithium-Ion Battery Health Prognostics

A Hybrid Machine Learning Model for Battery Cycle Life Prediction with Early Cycle Data

A deep learning approach incorporating attention mechanism and transfer learning for lithium-ion battery lifespan prediction

A novel hybrid model for lithium-ion batteries lifespan prediction with high accuracy and interpretability

Deep Learning Powered Lifetime Prediction for Lithium-Ion Batteries Based on Small Amounts of Charging Cycles

A machine-learning prediction method of lithium-ion battery life based on charge process for different applications

Early-stage lifetime prediction for lithium-ion batteries: A deep learning framework jointly considering machine-learned and handcrafted data features

Cerberus: A Deep Learning Hybrid Model for Lithium-Ion Battery Aging Estimation and Prediction Based on Relaxation Voltage Curves

Accurate battery lifetime prediction across diverse aging conditions with deep learning

Early prediction of battery lifetime via a machine learning based framework

Early prediction of remaining useful life for Lithium-ion batteries based on a hybrid machine learning method

A two-stage deep learning framework for early-stage lifetime prediction for lithium-ion batteries with consideration of features from multiple cycles

Hybrid Data-Driven Approach for Predicting the Remaining Useful Life of Lithium-Ion Batteries

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

Joint modeling for early predictions of Li-ion battery cycle life and degradation trajectory

An Empirical-informed Model for the Early Degradation Trajectory Prediction of Lithium-ion Battery

Lithium-ion battery degradation trajectory early prediction with synthetic dataset and deep learning

Battery Life Prediction Based on a Hybrid Support Vector Regression Model

Battery Early Prognostics Based on Pseudo Meta-Learning