Data-driven climate resilience assessment for distributed energy systems using diffusion transformer and polynomial expansions

A growing consensus is reaching that climate mitigation measures could fail to reach a sustainable future without immediate counterparts on climate adaptation, especially for global power systems driven by increasing share of renewable energy. A reasonable assessment procedure on the impact of arbitrary climate conditions should be studied before necessary countermeasures could be done. This paper concentrates on distributed energy systems with distributed renewable generators and studies how to quantitatively assess their climate resilience degrees for the first time. A data-driven diffusion transformer model is suggested to generate any climate scenarios according to both prior feature variables and historical climate datasets. Then, the interactive mechanisms between climate change and distributed energy system operations are analyzed. And three quantitative capacity credit criteria are proposed to measure system climate resilience degrees. Moreover, polynomial expansions yield an alternative model to alleviate the burden of replicative computations during criteria solutions. Case study applies historical climate data from Zhangbei, China. Results indicate that 1) the proposed diffusion transformer is capable of up to 20 % fluctuations tolerance of subjective feature expectations; 2) nominal capacity of renewable energy and hydrogen storage are closely related with climate resilience degrees; 3) third order polynomial expansion model estimates climate resilience results safely with R2 over 98.4 %; 4) practical policy recommendations are suggested to combat the impact of climate changes on distributed energy systems.