Machine learning for optimal electrode wettability in lithium ion batteries

Electrode wetting is a critical step in the Lithium-Ion Battery manufacturing process. The injection of electrolyte in the electrodes' porosity requires the application of pressure-vacuum pumping strategies without warranty that the full porosity will be fully occupied with electrolyte at the end of this process step. The electrode wettability strongly depends on the contact angle between the electrolyte and the electrode, the electrode microstructure characterized by its porosity, pore network and tortuosity factor, the electrolyte viscosity and density. Compu-tational fluid dynamics approaches such as the Lattice Boltzmann Method can provide relevant information of the filling process, yet these approaches come with significant computational cost. The use of machine learning techniques can provide surrogate models for the optimization of this multi-parameter process that depends on both chemical and physical properties. Within this context, we propose a general workflow for realizing this objective and provide detailed simulation-based experiments. These physics-informed surrogate models open the path to tractable, rapid solutions of parameter identification and design optimization problems. They also pro-vide a general workflow for applications on other optimal battery material design problems.

Automated summary

Electrode wetting is a critical step in the Lithium-Ion Battery manufacturing process, which requires the injection of electrolyte into the electrodes' porosity. The wetting quality depends on various factors such as the contact angle, electrode microstructure, and electrolyte properties. Computational fluid dynamics and machine learning techniques can be used to optimize this multi-parameter process.