A theory of diagnosis from first principles
Artificial Intelligence
Artificial Intelligence
Model-based reasoning: troubleshooting
Exploring artificial intelligence
Hi-index | 0.00 |
This paper presents a novel model abstraction and representation approach for the application of model-based reasoning (MBR) to helicopter powertrain diagnostics. Although the basic principles of MBR are well understood, its application to specific domains, especially in mechanical systems, has been restricted due to insufficient model fidelity. The traditional vibration analysis approach has been the main stay of powertrain diagnosis since accelerometers have proven to be the primary viable choice for gearbox instrumentation and monitoring. However, a suitable logical representation of mechanical components so as to facilitate diagnostic reasoning about their vibration characteristics is difficult to achieve. This paper exploits the physics behind the vibration signatures of gearbox components both in nominal and faulty operational modes, and attempts to extract and represent that knowledge in an MBR-friendly formulation. Based on these model constructs a complete, integrated, MBR methodology for fault diagnosis based on behavioral characteristics of system components is introduced in this paper. Finally, the implementation of this architecture on the intermediate gearbox (IGB) module of a helicopter powertrain is presented. Simulation results show the successful application of this methodology in the detection of input gear pinion damage in the intermediate gearbox of an H-60 helicopter.