Abstract
The aim of the thesis was to create the models which describes electrolyte infiltration and characterize its performance. The electrochemical impedance spectroscopy model should be developed based on Newman model. The Lattice Boltzmann Method must be used to describe the electrolyte infiltration process. All the models were combined with experimental results. The thesis was part of the ERC-Artistic project. The first stage of the thesis was devoted to develop the Electrochemical impedance Spectroscopy model. EIS constitutes an experimental technique used for the characterization of LIB porous electrodes tortuosities. For the first time, a 4D (3D in space + time) physical model is proposed to simulate EIS carried out on NMC porous cathodes, derived from the simulation of their manufacturing process, in symmetric cells. This methodology allows to understand the limitations of using EIS, electric circuit models and homogenized physical models for the determination of the tortuosity of NMC-based cathodes, revealing a complex interplay between the conductivity of the solid phases, the electrolyte properties and the cathode meso/microstructure. The second stage of the thesis was devoted to development of the electrolyte infiltration process. This step is crucial as it is directly linked to LIB quality and affects the subsequent time consuming electrolyte wetting process. It was reported here for the first time a 3D-resolved Lattice Boltzmann Method model able to simulate electrolyte filling upon applied pressure of LIB porous electrodes obtained both from experiments (micro X-ray tomography) and computations (stochastic generation, simulation of the manufacturing process using Coarse Grained Molecular Dynamics and Discrete Element Method). The model allows obtaining advanced insights about the impact of the electrode mesostructures on the speed of electrolyte impregnation and wetting, highlighting the important of porosity, pore size distribution and pores interconnectivity on the filling dynamics. Furthermore, we identify scenarios where volumes with trapped air (dead zones) appear and evaluate the impact of those on the electrochemical behavior of the electrodes. Further an innovative machine learning model, based on deep neural networks, to fast and accurately predict fluid flow in three dimensions, as well as wetting degree and time for LIB electrodes. The ML model is trained on a database generated using a 3D-resolved physical model based on the Lattice Boltzmann Method. We demonstrate the ML model with a NMC electrode mesostructure obtained by X-ray micro-computer tomography. The extracted pore network from tomography data was also used to train our ML neural network. The results show that the ML model is able to predict the electrode filling process, with ultralow computational cost (few seconds) and with high accuracy when compared with the original data generated with the physical model. Also, systematic sensitivity analysis was carried out to unravel the spatial relationship between electrode mesostructure parameters and predicted infiltration process characteristics, such as saturation dynamics, filling time among others. Finally, the EIS, LBM and Machine learning models will be integrated into the ARTISTIC platform. The platform can be used to simulate, understand, and optimize battery manufacturing. The platform will be free of charge and all the data and codes will be available