An Expert System for Accelerator Fault Diagnosis
E Malandain, Paul Skarek · CERN Document Server (European Organization for Nuclear Research) · 1987
This paper discusses an expert system to assist operation and running of the CERN PS accelerators.The expert system aimed for in future will be an essential tool to ease maintenance, fault finding and eventually operation.It will reduce downtimes of the accelerators and should facilitate setting up beams.Two prototypes are discussed, one written with an EMYCIN-based shell, the other with a powerful hybrid frame, logic and production rule based development tool.The prototypes are based on deep-knowledge modelling of the structure and functionality of the accelerator and the computer control system.It will create its knowledge base directly from data bases describing the system.In addition it includes heuristics of experts' knowledge about the running of the accelerator and dealing with breakdowns.