Energy Optimization of Electrolyzers Using Machine Learning: Enhancing Hydrogen Production Efficiency

This doctoral project aims to improve the energy efficiency of alkaline electrolyzers used in hydrogen production by using machine learning and data analysis techniques. Hydrogen production is crucial for energy transition and reducing the carbon footprint. However, current electrolyzers have limited energy efficiency, with significant losses due to high energy demand and heat generated during the electrochemical reaction.
The main goal of this thesis is to identify and optimize critical parameters of the electrolysis process to maximize hydrogen production while minimizing energy consumption. This includes exploring optimal operating temperatures, improving electrolyte composition, designing more efficient electrodes, and managing overvoltage. To achieve this, advanced machine learning techniques, such as artificial neural networks and genetic algorithms, will be used to develop precise predictive models. These models will help determine the optimal operating parameters by considering the complex interactions between different process variables.
Optimization algorithms based on deep learning techniques will also be developed to adjust the operating conditions of the electrolyzers in real-time. The research methodology includes collecting real data from electrolyzers, the electrical grid, and weather conditions. This data will be transformed and modeled to facilitate exploration and analysis using data science techniques.
The expected results of this research will not only improve the energy efficiency of electrolyzers but also reduce production costs and environmental impact, making hydrogen production more competitive and sustainable. This thesis will also contribute to the advancement of scientific and technological knowledge in energy optimization and machine learning applied to energy transition.