Project abstract
In France, water electrolysis using renewable energy is nowadays considered a promising solution for producing hydrogen without dependencies on fossil fuels. Among different types, polymer electrolyte membrane water electrolysers (PEMWE) are considered the most favourable technology for hydrogen generation from renewable sources. However, the lifetime of current PEMWE technology is generally much shorter than the target value in Europe. Optimizing PEMWE operating parameters has been considered a potential approach to mitigate material degradation and extend the PEMWE lifetime, but it has been difficult so far because PEMWE degradation and its correlation to the performance and operating parameters involves complex multiscale physicochemical phenomena. Project DuraPEME contributes to improve the durability of PEMWE by developing an artificial intelligence (AI)-accelerated multiscale degradation model. In parallel and in connection with the models developed, it will also propose accelerated stress tests for PEMWE. We will achieve this goal by 1) characterizing degradations from multiple scales; 2) developing and accelerating a multiscale degradation model 3) generalizing the model in different uses and powers. The project will adopt a highly interdisciplinary approach by integrating methods in PEMWE multiphysics, material science, numerical calculation, and machine learning. As outcomes, project DuraPEME will empower PEMWE technology with an efficient multiscale degradation model. The model will enable optimizing PEMWE operating parameters to mitigate degradations.