Modified Neuro-Fuzzy Failure Classifier of Helicopters Turboshaft Engines

In this work, the method of implementing neuro-fuzzy algorithms for solving the problems of monitoring and diagnostics helicopters turboshaft engines on modern on-board computing devices has been further developed. The construction of a mathematical model of a helicopter turboshaft engine and a failure classifier is described, their advantages and disadvantages are reflected in comparison with the classical methods used. It is shown that the developed neuro-fuzzy classifier makes it possible to detect defects in the main components of helicopters turboshaft engines at flight modes with an error not exceeding 0.78 %.