Optimal Recloser Placement for Improved Reliability in Stand-Alone Industrial Power Supply Systems

This paper proposes a cost-based decision-support framework for optimal recloser placement in stand-alone industrial distribution networks. The goal is to minimize total expected cost as the sum of capital expenditure (CAPEX) for additional reclosers and monetized interruption losses based on expected annual ENS, while discouraging violations of operating limits (voltage and branch current) via penalty terms. Reliability assessment follows a conventional analytical approach based on single-contingency (N–1) line outages and a two-stage FLISR restoration logic: (1) automatic isolation, where a subset of loads may be interrupted until switching is performed; and (2) post-switching restoration, where loads that cannot be resupplied remain disconnected until repair. For each contingency, interrupted load sets are identified through graph-based connectivity checks, and ENS is computed using failure rates, line lengths, and two restoration time components (fault localization/switching time and repair time). The planning horizon Tb is used as a practical trade-off parameter that scales the interruption-loss term linearly, allowing comparison of economically justified automation levels for different planning assumptions. Optimization is performed using a binary Genetic Algorithm. The framework is demonstrated on a real 35 kV distribution network supplying an Arctic oil and gas facility. For planning horizon values of 10, 15, and 20 years, the optimized solutions add 10, 12, and 15 reclosers, respectively, and reduce ENSyear from 12,536.71 kWh (baseline) to 3,422.73 kWh, 2,828.30 kWh, and 2,394.41 kWh. SAIDI decreases from 0.326 h to 0.115 h, 0.099 h, and 0.084 h, respectively. In all cases, total expected cost decreases compared to the baseline and operating constraints remain satisfied. A three-scenario sensitivity analysis shows that absolute ENS/SAIDI values are strongly influenced by uncertainty in failure rates and restoration times, while the main planning conclusions remain stable: longer horizons justify installing more devices, and optimized configurations remain economically beneficial across scenarios.

Overall, the proposed approach provides a technically grounded method for decision support in distribution network design and expansion planning for isolated industrial power systems, offering both economically justified device counts and reliability-cost trade-off insights.