Optimal Power Flow Problem Solution Through a Matheuristic Approach

Abstract

The optimal power flow (OPF) problem is a widely studied subject in the literature that has been solved through classical and metaheuristic optimization techniques. Nowadays, significant advances in computational resources and commercial optimization solvers allow solving complex optimization problems by combining the best of both worlds in approaches that are known as matheuristics, however, in order to solve the OPF problem, matheuristic approaches have been little explored. In this regard, this paper presents a novel Variable Neighborhood Descent (VND) matheuristic approach to solve the OPF problem for large-scale systems. The proposed algorithm combines the classic OPF model and the VND heuristic algorithm. The OPF problem is formulated as a mixed-integer nonlinear programming (MINLP) model, in which the objective function aims to minimize the fuel generation costs, subject to the physical and operational constraints of the power system. The integer variables of this MINLP model represent the control of taps positions of the on-load tap changers and the reactive shunt compensation equipment. To validate the proposed methodology, 17 power systems of specialized literature were tested with sizes from 14 to 4661 buses, and the obtained solutions are compared with the solutions provided by the commercial optimization solver Knitro. Results show the superiority of the proposed matheuristic algorithm compared with Knitro to solve the MINLP-OPF model for large-scale systems.

Keywords

• Matheuristic optimization
• mixed-integer nonlinear programming
• optimal power flow
• variable neighborhood search