A Knowledge-Based Diagnostic System for Pneumatic System

This paper presents an approach to a knowledge-based diagnostic system for pneumatic system. The construction of the diagnostic system is introduced, which contains the design and engineering knowledge about the pneumatic system to be diagnosed. Intelligent diagnosis and compensation functions are incorporated in a real time expert system that diagnoses faults in a pneumatic system. This expert system for fault diagnosis bases on knowledge acquisition, knowledge base and inference explanation. In particular, the role of domain models in guiding the knowledge-acquisition process is reviewed. For considering the diagnosis of complex systems like the pneumatic system, which has the nonlinear, time-varying and ripple coupling properties, traditional expert systems has its shortages, neural network techniques that may help in the design of a diagnostic system are presented. Moreover, neural network can be used together with expert system to enhance pneumatic diagnostic reasoning capabilities.