Development of a Knowledge based System for the Diagnosis of Faults on a Plating Process

摘要

The diagnosis of intermittent faults is a common problem with otherwise reliable, highly automated plant. This paper describes the development of an on-line automatic fault diagnosis system for a plating process. The system was developed within an Expert System shell that was linked with the plant control computer a programmable logic controller (PLC) to monitor the process for faults. This was achieved by generating a 'fault vector' within the PLC. The fault vector was constructed from individual computer based logical bits collected from the PLC that represented specific errors and timers that detected when the plant stopped for longer than it should have without a specific error being generated. The fault vector was used to trigger the Expert System to follow a knowledge based diagnostic tree and interrogate the PLC to determine the most likely cause of the problem and arrive at a solution heuris-tically. This all happened transparently to the user with the system quickly and consistently diagnosing a number of faults that had previously taken plant personnel variable amounts of time to analyse.