2020/12/17 by Kyri Baker, Baker, Kyri · 1 citation
Engineering · #FOS: Mathematics #Optimal Power Flow Distribution #Optimization and Control (math.OC) #Power Quality and Harmonics #Power System Optimization and Stability #Power System Reliability and Maintenance #Railway Systems and Energy Efficiency
paper · pdf · doi:10.48550/arxiv.2012.10031
openalex publication_date 2020/12/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Using machine learning to obtain solutions to AC optimal power flow has\nrecently been a very active area of research due to the astounding speedups\nthat result from bypassing traditional optimization techniques. However,\ngenerally ensuring feasibility of the resulting predictions while maintaining\nthese speedups is a challenging, unsolved problem. In this paper, we train a\nneural network to emulate an iterative solver in order to cheaply and\napproximately iterate towards the optimum. Once we are close to convergence, we\nthen solve a power flow to obtain an overall AC-feasible solution. Results\nshown for networks up to 1,354 buses indicate the proposed method is capable of\nfinding feasible, near-optimal solutions to AC OPF in milliseconds on a laptop\ncomputer. In addition, it is shown that the proposed method can find\n"difficult" AC OPF solutions that cause flat-start or DC-warm started\nalgorithms to diverge.\n