On the Necessity of Using Supervised Machine Learning for Risk-based Screening of Distribution Transformers: An Industrial Case Study

Distribution transformers (DTs) deliver a core role in electrical power distribution systems. It is mandatory to carry out timely inspection and maintenance of DTs, to achieve the anticipated reliability in power distribution systems. Risk-based inspection and maintenance analysis (RBI&MA) enable the optimum usage of available resources, to ensure the reliability and availability of power distribution systems. It is necessary to carry out risk-based screening of DTs to classify them, in order to perform detailed RBI&MA. This manuscript demonstrates the use of a supervised machine learning approach for the risk screening of DTs to classify them into low, medium, and high-risk groups. The suggested approach enables inspection and maintenance engineers and asset managers to recognize those DTs that qualify for detailed risk analysis. To illustrate the suggested approach, a risk screening is performed, using 40 units of DTs in a power distribution network located in a highly dense area. The approach developed in this study enables asset management capabilities in electric power distribution systems to be enhanced, by replicating a similar approach in other key components in power distribution systems.