Machine Learning-based Dynamic Climate Projections for Power System Planning Datasets
PI and co-PIs: Bri-Mathias S Hodge (University of Colorado Boulder); Aneesh Subramanian (University of California, San Diego); Claire Monteleoni (University of Colorado Boulder); Himanshu Jain (IIT Roorkee)
Funding amount: $135,000
Project overview: Climate change has the potential to disrupt power grids by changing both how much power is produced in different locations (especially by solar and wind) and also how much power is consumed (e.g. in air conditioning systems). This project uses state-of-the-art superresolution techniques to downscale global climate model scenario outputs to create publicly available climate change-informed load, wind, and solar power datasets for Western North America and India for power systems planning. Making this data publicly available can aid power system utilities in factoring changing climate into their generation and transmission build-out decisions, which is not performed today due to lack of data availability. Addressing these data limitations can improve preparedness for changing weather patterns and alleviate the impacts of climate change on energy systems.
Full abstract:
Click to expand
Electricity systems are seen as a key part of future decarbonized energy systems, but power system planning does not currently take into account the impacts of climate change on the system load and generation from renewable energy resources, chiefly due to a paucity of high resolution data. This project will utilize state-of-the-art artificial intelligence techniques, such as normalizing flows, to downscale global climate model scenario outputs to create publicly available climate change-informed load, wind, and solar power datasets for Western North America and India. The project is supported by deployment partners from the Western Electricity Coordinating Council, Jupiter Intelligence, and NVIDIA to ensure that the developed datasets have maximum real-world impact.
Power & Energy Climate Science & Modeling