Optimal D-PMU placement for improved voltage-based MV earth fault localization

The increasing value of lost load necessitates efficient fault localization methods (FLMs) in MV distribution networks. Our previous work [1] introduced a novel FLM that effectively identifies fault zones using a limited number of D-PMU (Distribution-level Phasor Measurement Unit) measurements. This study investigates the use of genetic algorithms (GAs) for optimal D-PMU placement to minimize the fault location zone. The analysis compares GA with an exhaustive search (IES) approach also proposed in our previous work [1] on six realistic feeders reconstructed from Enedis Open Data. Results show that optimal D-PMU placements tend to be located near remote end nodes, aligning with theoretical expectations. IES performs well for less complex feeders but becomes computationally expensive for complex feeders with numerous remote end nodes. The GA approach, particularly when restricted to remote end nodes, demonstrates promising results in terms of both solution quality and computational efficiency.