Lessons learned from IFES, a full-scale diagnostic in-flight expert system demonstration

摘要

Summary form only given. The in-flight expert system (IFES) demonstration is a research and development project in the area of vehicle health monitoring (VHM). The overall objective of this continuing project is to develop a complete inflight fault detection, identification and reconfiguration expert system demonstration using real-time data and hardware. IFES was implemented, using the OPS83 production system, as a distributed system of IBM 286 and 386 PCs. A few lessons learned were that the production system's CPU performance timing and memory use are adequate for real-time VHM systems; noisy data can increase RETE algorithm processing time by 20%; system reconfiguration input-output timing dominates CPU execution timing during fault identification 90% of the time; and OPS83, although excellent for failure detection applications, was not as well suited as Ada for the procedural nature of fault identification applications.