Chadwick-Helmuth Co., Inc. and the University of New Mexico propose to design and evaluate an on-board intelligent health assessment tool for rotorcraft machines, which is capable of detecting, identifying, and accomodating the slow variations and catastrophic changes of rotorcraft components in an adverse operating environment. A fuzzy-based neural network paradigm with an on-line real-time learning algorithm willbe developed to perform expert advising. A hierarchical fault diagnosis architectureis advocated to fulfill the time-critical and on-board needs in different levels of structural integrity over a global operating envelope. The research objective is to demonstrate the feasibility and flexibility of the proposed health monitoring procedure through numerical simulations and experimental verifications of bearing faultsin USAF 58th SOW/MH-53J PAVE LOW helicoptors. The proposed fault detection, identification, and accomodation architecture is applicable to various generic rotocraft machines. In a similar spirit, the proposedtechnology can be easily transferred to damage detection and identification of any general purpose dynamic systems, e.g. aeropropulsion engines, underwater vehicles, chemical processes, nuclear power plants, and manufacturing production equipments. The proposed system will greatly reduce the operational and developmental costs and services an essential component in an autonomous control system.
Keywords: Rotorcraft, Artificial Neural Network, Health Assessment, Vibration Analysis, Autonomous
