DNkS: A distance-based neighborhood k-search algorithm for determining meter–transformer connectivity in low-voltage grids

The Distance-based Neighborhood k-Search (DN$k$S) algorithm, introduced in this article, offers a novel approach to enhancing meter–transformer connectivity modeling in low-voltage grids. By employing a local $k$-neighborhood search strategy, DN$k$S effectively subdivides the grid into manageable sections, ensuring robust connectivity assessments. Utilizing metrics such as Adjusted Mutual Information and Accuracy, DN$k$S demonstrated superior performance in trials, achieving up to 100% accuracy in certain cases, significantly outperforming existing state-of-the-art methods such as deep convolutional time-series clustering and spectral embedding-based meter–transformer mapping. Although DN$k$S is effective, its performance critically depends on the accuracy of meter and transformer coordinates. In comparative analyses across various network configurations, DN$k$S consistently outperformed other methods, affirming its utility and effectiveness. The versatile nature of the algorithm would allow its integration into existing systems in various ways, for example, through an API or a web interface. Implementing DN$k$S promises substantial improvements in the reliability and accuracy of utility network models, directly contributing to enhanced grid management practices.