Artificial Intelligence Application to Diagnostics/Prognostics of Flight Control Systems

Increasing costs associated with aircraft maintenance snd low readiness figures require an improved, cost effective method for diagnosis snd prognosis of Flight Control System malfunctions. Diagnosis using traditional Built-In-Test (BIT) concepts have not proven fully effective. In many cases the BIT actually increased downtime as a result of false alarms. Application of' Artificial Intelligence techniques can lead to improved Flight Control System readiness and maintainability. A concept has been developed that will combine traditional BIT with expert system heuristics. The fusion of heuristics with analytical knowledge will produce a robust diagnostic/prognostic maintenance system. Specific design techniques have been identified and applied to a hydraulic subsystem. Initial results of the hybrid system indicate improvements over either a conventional stand-alone Expert System, or a conventional BIT approach. The implementation features pre/post-flight tests that require no additional sensors or test equipment. A key to the results is an improvement to the knowledge elicitation process, necessary to build the hybrid knowledge base. Practical guidelines were established snd a methodology developed to remove redundancies and ambiguities.