GNDO-Based Optimal PV-STATCOM Relocation in Distribution Systems

Abstract

The optimal placement of PV-STATCOM devices in distribution systems poses a challenging optimization problem due to the nonlinear, non-convex, and combinatorial nature of the search space, which requires solving a multi-period Optimal Power Flow (OPF) for each candidate configuration. This paper proposes a hybrid methodology that couples the Generalized Normal Distribution Optimization (GNDO) metaheuristic with a complex-domain OPF model implemented in Julia using JuMP and Ipopt. The proposed approach is evaluated on the 33-bus test system under four operational scenarios, ranging from baseline conditions to full active-reactive injection. The GNDO algorithm efficiently explores the solution space, reducing the computational burden associated with exhaustive search, which would require evaluating 4960 possible configurations. Comparative results demonstrate that the proposed methodology not only reproduces previously reported optimal solutions but also identifies alternative configurations that achieve additional energy loss reductions of up to 1.40% compared to the literature. These findings validate the effectiveness of the GNDO-based approach for distributed energy resource planning in power distribution networks.