Spatial-Constrained Optimization of Tie-Line Switch Placement: A Case Study of Kalimantan Using an Improved Genetic Algorithm

Abstract

Reconfiguring existing distribution networks (DN) is an important strategy within smart grid initiatives to minimize power loss, enhance reliability, and improve the system's resiliency. While the IEEE 33-bus DN is widely used as the benchmark for performing various reconfiguration methods, most studies assume ideal initial tie-switch placement and lack real-world geographical context. This paper presents a novel approach that integrates the spatial characteristics of South Kalimantan, Indonesia, into determining the optimal placement of new tie switches using an Improved Genetic Algorithm (IGA). Three case studies are detailed. In Case 1, the IGA is validated on the standard IEEE 33-bus system, achieving loss-minimization results comparable to those of existing techniques. Case 2 adapts the IEEE 33-bus model to represent South Kalimantan's distribution landscape, assigning each bus to a specific spatial location based on actual grid topology. The optimized configuration in this scenario yields a 4% reduction in power loss. Case 3 further advances the analysis by determining the optimal placement of two new tie switches in a spatially constrained setup, achieving up to a 53% reduction in loss and a 93% reduction in cable installation cost compared to Case 2. These findings underscore the importance of spatially informed planning in distribution system reconfiguration. Integrating local geographic conditions, such as those in Kalimantan, into optimization frameworks significantly enhances both technical performance and cost efficiency in the distribution system.