K-Means Clustering as a tool to supervise PD Monitoring data

This thesis provides a comprehensive guide to asset management and condition-based maintenance, with a special emphasis on partial discharges. The dissertation opens with an overview of asset management and condition-based maintenance, highlighting their significance in maintaining the effective and efficient operation of electrical assets. The definition, types, and sources of partial discharges, as well as their effects on equipment and methods for detecting them, are all covered throughout the following chapter. The dissertation enters into normative IEC 60270, which offers instructions for taking partial discharge measurements and interpreting them. This chapter discusses the fundamentals of the standard, its scope, terminologies, and definitions, as well as the advantages it offers in ensuring the safe and reliable performance of electrical equipment. The cloud computing topic is covered in detail in the next chapter, along with an overview of the technology's uses in asset management and maintenance. The chapter describes how cloud computing can be used to gather and analyse data from multiple sources, including sensors and other monitoring devices, to offer insightful information about the condition of electrical assets. The next section of the thesis examines clustering, a potent method for combining related data points in data analysis. The chapter discusses the various clustering algorithms and how they are used to partial discharge data. The thesis concludes with a research example illustrating the advantages of clustering partial discharge data sets in the cloud. The case study demonstrates how clustering may be used to find patterns and anomalies in partial discharge data, giving insight into the condition of electrical assets and enabling better maintenance decisions.