Open AccessAutomated defect detection in oil-lubricated parts and units of D30KP/KP-2 aircraft gas turbine engines by results of microwave plasma method
Author(s) -
A Y Hodunaev,
В. Г. Дроков,
В. В. Дроков,
V. V. Murishenko
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1384/1/012017
Subject(s) - gas turbines , classifier (uml) , oil analysis , microwave , turbine , automotive engineering , random forest , engineering , process engineering , computer science , artificial intelligence , pattern recognition (psychology) , mechanical engineering , telecommunications
Classifier between states of “normal/high maintenance/defective” for oil-lubricated parts and units of D30KP/KP-2 aircraft gas turbine engines is developed. The classifier is based on “random forest” machine learning algorithm. It is trained on results of microwave plasma measurements of metallic admixture in oil filter wash samples of engines. Technical state for train set was determined earlier by expert method and was confirmed by factory disassembly study. Classifier result for states “normal/high maintenance/defective” matches expert method in 73 %, 52 %, 66 % respectively.