Modelling the Impact of Manufacturing Uncertainties on Lithium-Ion Batteries

Abstract

The electrochemical properties of a lithium-ion battery are significantly influenced by the manufacturing process. Optimizing the manufacturing process requires to understand the relationship between the process parameters, the structural parameters and the resulting properties of the battery. Furthermore, tolerances in the single manufacturing steps lead to uncertainties which propagate along the process chain and affect the quality of cells. The uncertainties of the various structural parameters have differently strong influences on the properties. As such sensitive structural parameters need to be detected for a targeted optimization. Additionally the influence of cell-to-cell and lot-to-lot variations have a major influence on quality, and thus on the rejection rate when building battery packs. [1,2] In our prior studies, we established a coupled multi-level model approach that is able to analyze the influence of tolerances in the manufacturing process and study the process-structure-property relationship. [3] In this approach, a process chain model is coupled with a physical battery model based on Doyle et. al. [4] The coupled model approach is visualized in Figure 1. The process chain simulation consists of single process models describing the relationship between the input process parameters and the structural parameters for each manufacturing step. Tolerances can be considered in each process step by integrating stochastically distributed process and structural parameters. The process chain simulation provides a characterized battery cell, which is the input for the battery cell simulation, where the electrochemical properties are determined. In this work, the implementation of the coupled model approach and a first case study are presented. The case study is chosen in such way, that the effect of tolerances on the structural parameters, the properties, the propagation of uncertainties and interactions can be studied. Figure 1 shows the four investigated scenarios. In the first scenario, no uncertainties appear in all manufacturing steps. It is used to represent the reference point. Scenario two and three were chosen so that fluctuations occur in only one manufacturing process and propagate without interactions along the process chain. The tolerances either occur in the coating or the calendering process. The fourth scenario combines scenario two and three. This scenario is used to study interactions between the fluctuating processes. The results of the process chain simulation for the structural parameters indicate that for each process step a mainly affected structural parameter could be identified: In the coating process, the electrode thickness is mainly affected and in the calendering process the electrode porosity. Additionally, it is shown that the interactions of uncertainties in different manufacturing steps show, that the deviations are not accumulated along the process chain. Furthermore, the results of the battery cell simulation reveal distinct impact dependent on the implemented structural parameters and the applied discharge rate. An optimum in the calculation area of the considered range of the deviations leads to a skewed distribution of the electrochemical properties. Finally, the sensitivity of the structural parameters on the electrochemical properties are studied and evaluated. The results provide guidance to focus on the most relevant manufacturing tolerances in order to effectively optimize the product quality and reduce costs. Figure 1: Multi-Level Model Approach with a coupled process chain model and a battery model. Also included are the Scenarios for the Case Study. [1] S. Santhanagopalan and R. E. White, “Quantifying Cell-to-Cell Variations in Lithium Ion Batteries,” Int. J. Electrochem., vol. 2012, pp. 1–10, 2012. [2] F. An, L. Chen, J. Huang, J. Zhang, and P. Li, “Rate dependence of cell-to-cell variations of lithium-ion cells,” Sci. Rep., vol. 6, pp. 4–10, 2016. [3] M. Thomitzek, O. Schmidt, F. Röder, U. Krewer, C. Herrmann, and S. Thiede, “Simulating Process-Product Interdependencies in Battery Production Systems,” Procedia CIRP, vol. 72, pp. 346–351, 2018. [4] M. Doyle, “Modeling of Galvanostatic Charge and Discharge of the Lithium/Polymer/Insertion Cell,” J. Electrochem. Soc., vol. 140, no. 6, p. 1526, 1993. Figure 1