Study on Influencing Factors of Consistency in Manufacturing Process of Vehicle Lithium‐Ion Battery Based on Correlation Coefficient and Multivariate Linear Regression Model

Abstract

Lithium-ion battery manufacturing is a multiprocess serial system and the consistency of each single cell will affect the performance and safety of battery system after grouping. Therefore, optimizing the battery preparation process and improving the battery consistency have become the key technical issues in battery preparation process. In this study, the inconsistency of finished batteries caused by manufacturing process is analyzed from three processes: electrode preparation, battery assembly, and liquid injection and formation and the structure-activity relationship between manufacturing process data and battery consistency is studied. Taking the actual capacity, self-discharge rate, and internal resistance of single cell as target variables, the correlation coefficient matrix between original process variables and target variables is constructed by using correlation analysis method and the correlation index between process variables and target variables is obtained. The multiple linear regression model is established by stepwise regression method and the weight of original process variables affecting battery consistency is deduced. The results show that the data of multiple process links have a high influence weight on the actual capacity and self-discharge rate of batteries, which provide a theoretical model for the key strategies of improving quality consistency in lithium-ion battery industry.

Capacity Consistency Prediction and Process Parameter Optimization of Lithium‐Ion Battery based on Neural Network and Particle Swarm Optimization Algorithm

“…The sample data of this study come from the previous publication, [ 6,7 ] in which the sample number is 49 565 soft‐pack batteries. Based on the process data, a BP prediction model is established to predict the consistency of capacity distribution, which can intervene and control the consistency problems before large‐scale manufacturing in time.…”

“…[ 6 ] In addition, based on the collected 49 565 soft‐pack battery process data samples, Pearson correlation coefficient and multiple regression equation are carried out to quantify the correlation and contribution of process parameters to the consistency, which can be concluded that the influence weight of multiple process link data on battery grading capacity is high. [ 7 ] Therefore, the main purpose of this study is to enhance the capacity consistency through the prediction of battery capacity consistency and optimization of process parameters achieved by theoretical models to quantify the manufacturing process parameter formulation under the premise of optimal capacity consistency.…”

A method for battery fault diagnosis and early warning combining isolated forest algorithm and sliding window

“…At present, according to the different research methods of battery cell inconsistency fault diagnosis, it can be roughly divided into statistical analysis methods, machine learning methods based on outlier detection, neural network algorithm, signal processing method based on information entropy analysis, and so forth. [20][21][22][23][24][25][26][27][28] Gasper et al 23 used machine learning-assisted model recognition methods to predict battery life, with uncertainty quantified by bootstrap resampling, and the uncertainty of life prediction is greatly reduced. Xia et al 7 established a model for short-term capacity estimation and long-term remaining useful life prediction of lithium-ion batteries based on data-driven methods.…”