Sensitivity Analysis of Vibration Characteristics for Localization of Imbalance Fault of the Helicopter Blade using Weight Distributions of Self-organizing Neural Network

Document Type : Research Paper

Authors

1 دانشکده مهندسی هوافضا، دانشگاه صنعتی مالک اشتر

2 Aerospace Complex, MA University of Technology, Tehran, Iran

Abstract

One of the important subjects and barriers to the development of health and usage monitoring systems for air vehicles is the large amount of required data for algorithm validation. In this research, an appropriate solution is proposed to reduce the number of sensors, the volume of data, and the calculations in the helicopter's main rotor health monitoring system. Using the self-organizing neural network, a new algorithm is developed for sensitivity analysis among vibration characteristics that are extracted from the rotor dynamic model to minimize the required data acquisition tests. Firstly,, a comprehensive nonlinear dynamic model of a helicopter is used for the simulation of the signals for 15 imbalance faults as an example of rotor faults. Then, 16 vibration characteristics are calculated and extracted from linear and rotational accelerations of the main rotor hub in steady flight conditions. Finally, different samples of 96 characteristics from this dataset are clustered using a self-organizing neural network, and the effective and key characteristics for fault detection are determined with the weight distribution function of each characteristic.,

Keywords