A KNOWLEDGE-BASED EXPERT SYSTEM FOR SILICATE GLASS PLANT MAINTENANCE

Abstract

A Knowledge-Based Expert System (KBES) which will aid diagnosis of plant faults within Silicate Glass Plant domain was developed. Knowledge was elicited using a combination of techniques (protocol-generation, teach and teach back, protocol analysis techniques) and root cause failure analy-sis was used to identify primary cause of fault as they were presented by the domain experts. The elicited knowledge were structured and coded in form of rules using C Language Integrated Produc-tion System „CLIPS SHELL‟. The KBES developed comprised a main system and four sub-systems with utility files. Performance evaluation was carried out on the KBES for consistency and exactness. Field tests were conducted and the data acquired were analyzed using some maintenance parame-ters. Results obtained when the KBES was used in the plant‟s faults diagnosis compared with when it was not used showed that with expert system in use, an average reduction in plant downtime of 36.62% was obtained for field test carried out. Also, 24.28% and 34.03% were obtained for average increase in Plant Availability and Total Plant Maintenance Productivity respectively.