Data-Driven Optimal Power Flow: A Physics-Informed Machine Learning Approach

This paper proposes a data-driven approach for optimal power flow (OPF) based on the stacked extreme learning machine (SELM) framework. SELM has fast training speed and does not require time-consuming parameter tuning process compared with deep algorithms. However, direct application of OPF is tractable due to complicated relationship between system operating status solutions. To this end, regression framework developed that decomposes model features into three stages. only reduces complexity but also helps correct bias. A sample pre-classification strategy active constraint identification achieve enhanced feature attractions. Numerical results carried out IEEE Polish benchmark systems demonstrate proposed method outperforms other alternatives. It shown can be easily extended address different test by adjusting few hyperparameters.