A machine learning approach for hydrogen safety in confined spaces: forecasting the pressure peaking phenomenon

Hydrogen is the most abundant element in the universe and has gained significant attention as a clean and renewable energy source. However, hydrogen also poses significant safety risks, especially when it is stored, transported, or used in large quantities. One of the main concerns with hydrogen is its flammability and explosive nature. Hydrogen is highly flammable and can ignite in the presence of a spark or heat source. This flammability, combined with the fact that hydrogen is lighter than air, means that hydrogen gas can quickly spread and mix with air, creating a potentially explosive mixture. Another problem related to hydrogen safety is the potential for high-pressure releases, which can result in a sudden and significant increase in pressure that can overwhelm existing safety systems. This is particularly concerning in indoor environments, where the confinement of the space can exacerbate the potential for harm. To address these and other safety risks, stringent regulations and guidelines have been established for hydrogen storage, handling, and usage. However, these regulations are not enough to guarantee safety and ongoing research and development is needed to better understand the dangers associated with hydrogen and develop more effective safety measures. The focus of this thesis is to utilize advanced analytics techniques to predict physical phenomena in indoor hydrogen releases, such as pressure peaks, which can worsen accident scenarios or compromise existing safety measures. The ultimate goal is to enhance the safety of indoor hydrogen releases through predicting these physical phenomena and supporting operational safety measures.