Application of Acoustic Sensors for Helicopter Health Monitoring

Although vibration analysis has been successfully used to monitor rotorcraft mechanical components, there are problems with system implementation and development which limit the method. Acoustical methods need to be evaluated for aircraft health monitoring on a realtime basis to detect defective mechanical components before failure. Physical Acoustics Corporation proposes to implement both acoustic emission non-destructive testing and neural networks to perform pattern classification on data resulting from a combination of three signal processing concepts. Neural networks are well know for their ability to solve non-linear classification problems with high execution speeds, making them well suited for real-time applications. Our technical objectives for Phase I are aimed at establishing a solid proof of concept that supports the development of a commercially viable system for monitoring rotorcraft mechanical components in-flight, as well as detecting and locating specific problems in real-time. The ultimate goal is to develop and market an advanced technology health monitoring system which will initially benefit the defense industry and eventually become available for civil applications.