Classification of distribution transformer using cluster analysis

This paper presents a methodology for classification transformer using Euclidean distance and cluster analysis. A number of transformers and are stored together with the corresponding load profiles in a database of patterns. The classification of a transformer not included in the database is obtained by comparing the transformer's actual load profile to the ones stored in the database. The purpose of 6 other transformers that were not included in the database is to evaluate the efficiency of each technique employed for classifying the group of transformers. This paper describes how this can be done when using the Euclidean distance. A supervised classifier using cluster and Euclidean distance in MATLAB provides an efficient and accurate classification method to separate load data into clusters base on their properties. Therefore, features are extracted from the data set, and these features are formed into feature clusters in order to identify patterns in signals as they are related to various physical behaviors of the system. The classifier curves are classes of data being separated into groups based on their characteristics and behaviors.