Correlation analysis of hotspot temperatures in power transformer – a case study

Abstract

This paper presents a correlation analysis of hotspot (HS) temperatures in power transformers on the example of two representative units coming from over 1000-unit population of high voltage (HV) to medium voltage (MV) transformers. Analyzed population is characterized by relatively low load and high age. Measuring data form one year period with 1h steps are gathered for analysis by on-line temperature monitoring systems mounted in the transformers. Analyzed data consist of ambient temperature, winding HS temperatures and load. Conducted correlation analysis shows that regarding that fleet there is relatively low correlation between load and HS temperature, and HS temperature is mainly depended on the ambient temperature. Observed dependencies may be used for optimization of the present thermal prediction models of the HS temperature in terms of the application for some similar, low loaded transformer populations.