Intelligent maintenance frameworks of large-scale grid using genetic algorithm and K-Mediods clustering methods

Abstract

Large-scale power grids, especially smart grid systems, consist of a huge amount of complex computerized electronic devices, such as smart meters. A smart maintenance system is desired to schedule and send maintenance worker to locations where any computerized devices become faulty. A grid management system (GMS) is purposely designed in the way that the following three conditions are generally fulfilled: 1) all workers are working on full capacity everyday; 2) the highest severity level faulty devices are fixed the quickest; and 3) the overall traveling distance/time is minimized. In this study, two intelligent grid maintenance framework are proposed considering the above three conditioned based on two state-of-arts algorithms, namely, genetic algorithm and K-mediods clustering method, respectively. Five real-world datasets collected from five different local cities/counties in eastern China are adopted and applied to verify the effectiveness of the two proposed intelligent grid maintenance frameworks.