AI – machine learning algorithms applied to transformer diagnostics
The machine learning algorithms have shown accuracy when analyzing complex power transformer data, but human judgment is crucial in their training process.
by Dr. Luiz Cheim
With the arrival of the age of big data, e-commerce, and smartphones, there has been a growing interest in the application of fast and sophisticated tools, namely machine learning algorithms, to handle massive amounts of data and extract meaningful information that can boost and speed up regression and classification problems, as for example in short term load forecasting and asset condition assessment. This paper describes the use of machine learning (ML) algorithms as supporting tools for the automatic classification of power transformers operating conditions. The work consists of training of multiple ML algorithms with real-life data from one thousand transformers that were individually analyzed by human experts. Each transformer in the database was scored with a ‘green,’ ‘yellow’ or ‘red’ card depending on the data and the interpretation of human experts, thus serving as the target variable in ML supervised training mode.
The machine learning algorithms have shown an impressive accuracy when analyzing complex power transformer data, however, human expert judgment is crucial in their training process
Artificial intelligence and machine learning algorithms are novel techniques that find more applications each day. The article describes the use of machine learning algorithms for the automatic classification of power transformers operating conditions, diagnostics, and asset condition assessment.
Learn more about the main steps towards training of the multiple MP algorithms and the stunning output produced by those algorithms by reading the whole article.
Keywords: automated tool, condition assessment, machine learning algorithms, transformer diagnostics
