Publication Type : Conference Paper
Publisher : Institute of Electrical and Electronics Engineers (IEEE)
Source : IEEE Access
Url : https://doi.org/10.1109/access.2026.3694790
Campus : Coimbatore
School : School of Engineering
Department : Electrical and Electronics
Year : 2026
Abstract :
Directional overcurrent relay (DOCR) coordination, a nonlinear constraint optimization problem, is highly important for the reliability of the power distribution system. Five metaheuristic algorithms namely, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Adaptive Differential Evolution (JADE), Moth-Flame Optimization (MFO) and Walrus Optimization Algorithm (WaOA) are compared on three benchmark networks of increasing complexity namely, a 5-bus radial feeder, a WSCC 9-bus mesh system, and the IEEE 33-bus distribution system. A multi-faceted assessment model is used with five objective functions, three standard IEC time-current characteristic (TCC) curves and two user-defined (UD1 and UD2) formulations. UD1 deploys a single set of non-standard optimized constants across the relays, while UD2 uses independently optimized shape constants for each relay. Each algorithm is compared under an identical computational budget, based on 30 independent runs, and performance measures are reported through descriptive statistics, Friedman and Wilcoxon signed-rank tests, convergence analysis, and verification of the coordination margin. The findings imply topology-related behavior in the algorithm’s performance: JADE is observed to be statistically consistent in the radial system, whereas GA is most dominant in the mesh and large-scale systems in terms of total operating time across all topologies. It is also observed that the full range of standard IEC curves with realistic Time Multiplier Setting ( TMS ) limits are not suitable for 33-bus system. This indicates that user-specified TCC curves are not merely beneficial but a requirement for the coordination of medium-scale distribution systems. The UD2 feature yields the shortest operating time for any system topology, and the optimized coefficients give physically interpretable trends that flatten as the network depth increases. These findings offer statistically confirmed principles for the protection engineer in choosing algorithms, the types of curves they need, and numerical relay to achieve effective coordination in current distribution networks.
Cite this Research Publication : Anu Kuriakose, S. Balamurugan, R. R. Lekshmi, A Multi-Faceted Performance Evaluation of Metaheuristic Optimization Techniques in Radial and Meshed Distribution Networks, IEEE Access, Institute of Electrical and Electronics Engineers (IEEE), 2026, https://doi.org/10.1109/access.2026.3694790