CFD Modeling and AI-Based Prediction of Alkaline Water Electrolysis Cell’s Performance for Hydrogen Production

Abstract

This work presents a comprehensive model of an alkaline water electrolysis cell for hydrogen production to evaluate its electrochemical and fluid dynamic phenomena simultaneously. For this, a 2D alkaline water electrolysis model was simulated on COMSOL MULTIPHYSICS considering both gas and liquid phases. A Multiphysics approach is used in simulations and the results are validated with the experimental data to ensure model’s accuracy. The CFD model includes the equations for electric current conservation and the gas and liquid phase. Polarization curves are generated to evaluate the AWE cell’s performance and electrochemical response at different operating conditions. The CFD model allows to predict the distribution of the generated gases, movement of the bubbles, and turbulence within the cell as well as the impact of current density, electrolyte flow rate, electrode-diaphragm distance, and other parameters on the gas profiles. While the CFD model provides valuable insights into alkaline water electrolysis, combining it with a neural network model further enhances its potential for better cell design and performance. The trained ANN accurately predicted the complex relationships between input and output parameters with an R² value of 0.99922. The combined CFD-ANN approach provides comprehensive understanding of the AWE cell’s behavior, further optimizing its design for efficient hydrogen production.