COMPARATIVE ANALYSIS OF MACHINE LEARNING METHODS FOR SOLVING THE PROBLEM OF PREDICTING FAILURES IN GAS TURBINE ENGINES

Abstract

Gas turbine energy technologies are one of the most important components of the modern and advanced energy industry. An important task is to ensure the uninterrupted operation of the equipment in a given period; therefore, monitoring and diagnostics of the technical condition of the equipment continue to play an important role in ensuring the quality of the gas turbine engine. The article examines the work on equipment diagnostics using machine learning. It discusses various solutions for combining machine-learning methods and dealing with unbalanced data to solve the problem of predicting the failure of gas turbine equipment on a dataset that has the above disadvantages. There is a review of the solutions and methods under consideration to deal with the problems of the dataset. At the end, the authors provide a comparative table of the results of the application of the considered solutions based on the quality metrics of the Recall, Precision, F1-score classification, and PR-AUC and ROC-AUC curves.