Heterogenous high performance computing for accelerating power system transient simulation

Qiuhua Huang

qiuhuahuang@mines.edu

Mehmet Belviranli

belviranli@mines.edu

Large-scale renewable energy development and integration is the main trend and will shape the future of energy systems and will have profound impact on our economy and society. However, there are increasing delays in the interconnection process, with over 2000 GW (more than existing U.S. total generation capacity) renewable energy projects waiting for interconnection for years [1]. One main bottleneck is the inefficiency of the renewable interconnection studies that increasingly require detailed electromagnetic transient (EMT) simulation.   To address this issue,  this project aims to expedite  large-scale EMT simulation by leveraging graphic processing unit (GPU)-boosted parallel computing, thereby streamlining the interconnection of inverter-based resources (IBRs) into transmission systems, supporting larger-scale and faster renewable energy deployment. This project leverages NREL’s open-sourced, CPU-based, parallelizable EMT simulator called ParaEMT [2]. This will provide a good, easy-to-use platform for both EE and CS students to collaborate and extend it to solve critical energy and climate change problems. References: [1] https://emp.lbl.gov/publications/queued-characteristics-power-plants-1 [2] M. Xiong, B. Wang, D. Vaidhynathan, J. Maack, M Reynolds, A Hoke, K Sun, J. Tan, “ParaEMT: an open source, parallelizable, and HPC-compatible EMT simulator for large-scale IBR-rich power grids,” IEEE Trans. Power Del., vol. 39, no. 2, pp. 911-921, Apr. 2024.  

The team will be advised by two faculty members who are the experts in heterogeneous computing and power system simulation, respectively. The team will work together to understand existing implementation methods in and codes of ParaEMT (open-sourced on GitHub [1]) and figure out suitable strategies to convert it from CPU-based to GPU-based implementation, for example, using the Jax library [1]. References: [1] https://github.com/NREL/ParaEMT_public [2] https://github.com/google/jax

Grand Challenge: Engineer the tools of scientific discovery.

 

1. https://emp.lbl.gov/queues
2. https://www.energy.gov/sites/default/files/2024-04/i2X%20Transmission%20Interconnection%20Roadmap_1.pdf
3. https://github.com/NREL/ParaEMT_public
4. Review of Methods to Accelerate Electromagnetic Transient Simulation of Power Systems. https://ieeexplore.ieee.org/document/9459192

Qiuhua Huang, qiuhuahuang@mines.edu | Mehmet Belviranli, belviranli@mines.edu