Quantitative Risk Assessment of Proton Exchange Membrane Electrolyzers for Hydrogen Production at Nuclear Power Plants

Abstract

Hydrogen electrolyzers coupled to nuclear power plants (NPPs) present an attractive opportunity for cost-efficient and clean hydrogen production, enabling the transition to hydrogen fuel and energy storage in line with national and international energy policies. NPP-hydrogen pilot demonstration projects began in 2022, with further developments anticipated over the next several years, spanning multiple design variants under consideration. However, nuclear-integrated hydrogen plants are complex systems with multiple interacting components and unique operational conditions that have not been fully analyzed from a risk assessment perspective. Ensuring the safe and reliable operation of the NPP requires the safe and reliable operation of the proton exchange membrane (PEM) electrolysis facility and vice-versa. Thus, it is imperative to develop the appropriate quantitative risk assessment (QRA) algorithms, data, and modeling capabilities for hydrogen electrolyzers and use them to provide insights into design choices and regulations. In this work, we develop risk assessment methods and data for PEM electrolyzers coupled to NPPs by performing a QRA on a 1 MW PEM electrolyzer. We evaluated the risks (scenario, probability, and consequence) associated with integrated nuclear-hydrogen system operations. To identify failure events involving the components within the PEM electrolyzer system and its dynamic operation, we conducted a failure modes and effects analysis (FMEA). This FMEA produced over 850 failure scenarios on a modal level and served as the starting point to create fault tree models. We then quantified these models using existing reliability data banks and new industry data to calculate the frequencies of hydrogen and oxygen crossover and hydrogen, oxygen, and nitrogen releases. To facilitate component-level design recommendations, we conducted an importance measure analysis on the fault tree models. We developed novel Bayesian networks (BNs) to model hydrogen ignition probability. These networks consider a multitude of influencing factors relating to system design, maintenance, and external conditions to make informed hydrogen ignition probability estimations. To characterize the thermal and mechanical damage caused by hydrogen jet fires and explosions, we used the HyRAM+ toolkit and analytical physics models. Using these new models, we developed risk-informed facility layout options through location-specific individual and site risk. The key contributions of this work are: (1) the establishment of technical foundations for QRA of electrolyzers and nuclear-integrated systems, (2) identification of risk-critical components and system functional groups, (3) development of novel BNs for hydrogen ignition estimation, and (4) risk-informed design, siting, and layout recommendations for nuclear-hydrogen production facilities. By advancing the understanding of the possible hazards associated with the operation of a hydrogen facility at an NPP, we provide robust scientific information that appropriately mitigates the risks of these new facilities. Through this rigorous investigation and quantification of risks associated with a PEM electrolysis facility, we establish the framework for additional risk assessments to enable the future deployment of other hydrogen technologies and nuclear integrated systems.