Assessment of water electrolysis projects for green hydrogen production with a novel hybrid q-learning algorithm and molecular fuzzy-based modelling

Abstract

Determining the most important criteria for increasing the efficiency of water electrolysis investments provides businesses with a competitive advantage. Although there are many studies in the literature on this importance, there are very few studies determining the most important of these performance indicators. To satisfy this gap, the purpose of this study is to make assessment of water electrolysis projects for green hydrogen production via a novel model. First, the balanced expert evaluation matrices are obtained by q-learning algorithm. Secondly, the criteria for water electrolysis investments are prioritized using molecular fuzzy Bayesian network (BANEW). Thirdly, green hydrogen strategies for water electrolysis investments are ranked with molecular fuzzy multi-objective particle swarm optimization (MOPSO). The most important contribution of this study to the literature is the determination of the criteria that should be applied primarily for the performance increase of water electrolysis investments by creating a new model. The use of molecular fuzzy numbers is a very important contribution of the model to the literature. In this process, the use of three-dimensional geometric figures allows the reduction of uncertainties in decision-making processes. The findings indicate that lifespan of electrolysers and production capacity are the most essential criteria. Additionally, proton exchange membrane electrolysers and alkaline water electrolysis are found as the most critical green hydrogen strategies. Extending the life of electrolysers is crucial to increase sustainability in hydrogen production and reduce long-term costs. In this context, research incentives should be provided for the development of materials and technologies to increase the durability of electrolysers. Similarly, establishing quality standards to extend the life of electrolysers also contributes to achieving this goal.