Hybrid Learning Aided Inactive Constraints Filtering Algorithm to Enhance AC OPF Solution Time

The optimal power flow (OPF) problem contains many constraints. However, equality constraints and a limited set of inequality constraints encompass sufficient information to determine the problem feasible space. This article presents a hybrid supervised regression-classification learning-based algorithm to predict active and inactive inequality constraints before solving AC OPF solely based on nodal power demand information. The proposed algorithm is structured using a mixture of classifiers and regression learners. Instead of directly mapping OPF results from demand, the proposed algorithm removes inactive constraints to construct a truncated AC OPF. This truncated optimization problem can be solved faster than the original problem with less computational resources. Numerical results on several test systems show the proposed algorithm's effectiveness for predicting active and inactive constraints and constructing a truncated AC OPF.