A NOVEL MODEL-BASED REASONING APPROACH TO SYSTEM-LEVEL DIAGNOSTICS OF A HELICOPTER INTERMEDIATE GEARBOX

摘要

A novel Model-Based Reasoning (MBR) methodology for fault diagnosis based on the behavior of system components is introduced in this paper. The logical representation of system models and the subsequent formulation of a fault propagation tree and a diagnostic decision tree from the structural and functional abstraction of the system are studied. Techniques for ambiguity resolution amongst likely fault candidates are examined, and an overall diagnostic reasoning architecture for system-level diagnostics is proposed. Finally, the implementation of this architecture on the intermediate gearbox (IGB) module of a helicopter power-train is presented. Available in-flight data collection technologies for the IGB record only the overall system behavior, which is a composite of the individual behaviors of the component units. For example, the IGB is monitored only through accelerometers mounted on the frame of the unit. The abstraction of the mechanical components as mass-spring-damper models allows for the frequency domain analysis of the system response as a linear aggregate of the individual responses. Frequency demodulation techniques present an effective method for resolving distinct fault modes in this case. Simulation results show the successful application of this methodology in the detection of input gear pinion damage in the IGB of an H-60 helicopter.