Screening of Materials for Hydrogen adsorption using Machine Learning

The transition towards renewable energy underscores the importance of robust energy storage solutions to mitigate the intermittent nature of solar and wind power. Hydrogen, with its high energy density, emerges as a viable energy storage medium, provided optimal materials are discovered for its effective storage. This thesis aims to identify such materials through a data-driven approach. A screening tool is developed employing Machine Learning (ML) and Neural Network (NN) algorithms to sift through extensive material databases, identifying candidates with high hydrogen adsorption capacities. The subsequent phases involve constructing a matrix of use-cases under varying operating conditions and developing a comprehensive database assessing the environmental impacts and sustainability of selected materials. The goal is to significantly reduce the Levelized Cost of Storage (LCOS) by pinpointing materials that showcase high hydrogen adsorption capacities while meeting sustainability criteria. By intertwining advanced ML and NN algorithms with domain-specific knowledge, this thesis endeavors to address the complex challenge of hydrogen storage material identification, contributing constructively towards the global clean energy transition.