Building expert systems for repair domains

Abstract

Electromechanical systems such as computer peripherals, cash registers, and some office equipment have service manuals which describe repair procedures at a certain practical level of detail, eg., about board or electromechanical unit replacement. Repair of such systems is normally accomplished by direct replacement of Field Replaceable Units (FRUs). Detection and redacement of defective FRUs are the basic odjectives of the repair task in these systems. systems for Depending on the type of system serviced, service manuals vary considerably in complexity. Buildingm I For some svstems, comdex decision trees and repair clomams flowcharts are part of the manual while in other systems only rudimentary information on repairAbstract: This paper describes a conceptual framework for building expert systems in repair domains. A specific example of the use of this framework in building an expert system for a field service repair problem is given. The example system is designed to assist fieZd service technicians-in ~roub~shooting and repait.ing electronic systems at the board level. This system, a Field Service Advisor (Fieldserve), is designed ( I ) to diagnose single or multiple defects in electronic systems, (2) to guide a technician through appropriate repairprocedures, (3) to verifi diagnoses and monitor repair effectiveness and ( 4 ) tc control. zaintain records for subsequent evaluation and quality.-

Putting the expert in charge: graphical knowledge acquisition for fault diagnosis and repair

“…Hofmann, Caviedes, Bourne, Beale, & Brodersen (1986) reported that their expert had a difficult time structuring the knowledge required for their (classification-based) system but instead preferred an algorithmic, decision-tree style.…”

Intelligent troubleshooting of complex machinery

“…Nevertheless, because the symptoms of a problem are strongly dependant on one another, exploring a hypothesis, that is subsequently rejected, can simultaneously reinforce or weaken other hypotheses. Also, one of the characteristics of troubleshooting is an efficient hierarchical decomposition of problems into sub-problems [Hofman et al 1986, Pickering et al 1986. Bonissone and Johnson 1983, Fink and Lusth 1987.…”

An Expert System for Diagnosing and Proffer Solutions to Causes of Overheating of a Bulldozer Engine (Case Study Model D60s-6 Komatsu Products)

“…For several years expert system has been a reliable, dependable and convenient tool for technical diagnosing model for solving problems in different fields of engineering and technology (Lugar & Stubblefield, 2004). Expert system has a lot of advantages such as cost efficiency, convenience, positive impact on maintenance strategy, accuracy, time saving, fast and reliable when used in making decisions (Hofmann et al, 1988). Hofmann et al (1988) proposed and adopted two approaches to address the problems encountered in reason strategies, these approaches are Explainable Expert System (SME) which is enriched to include explicit strategic knowledge and Reconstructive Explainer (REX) which has a representation of domain rule knowledge.…”