Application of Acoustic Sensors for Rotorcraft Gearbox Health Monitoring and Diagnostics

The early detection of incipient failure in a Mechanical System is of practical importance as it permits scheduled inspections without costly shutdowns and indicates locations in need of repair before a system will incur any catastrophic failure. This project aims at the development of a comprehensive procedure to identify incipient mechanical faults and their propagations using acoustic signals. Considerable success has been already accomplished by the authors in identifying gear failures from both experimental and analytical vibration signals. Some success had also been achieved in correlating acoustic signals with system vibrations. However, extensive efforts in i) acoustic signal conditioning, ii) correlations between acoustic and vibration signals due to component damage, and, iii) development of pattern recognition and damage estimation procedures for acoustic signals, are required for the development of an accurate machine health monitoring system. In order to develop an accurate signature pattern recognition procedure, a complete database of the acoustic/vibration signals has to be generated. Experimentally validated numerical models will be developed to generate acoustic and vibration signal database for both normal and abnormal operating conditions. The main goals of this study are: a) development of a procedure to eliminate unwanted noise in the acoustic signal and to increase coherence with the vibration b) experimental examination of the applicability of acoustic sensors, c) correlation of acoustic signal with mechanical vibrations, feasibility study of using existing vibration monitoring techniques for acoustic data, and the development of a condition based prognostic procedure to predict component life.