An Approach to Develop Knowledge Representation for Expert System to Diagnose Faults in Domestic Systems using Qualitative Modeling.

Show full item record

Title: An Approach to Develop Knowledge Representation for Expert System to Diagnose Faults in Domestic Systems using Qualitative Modeling.
Author: Kotcher, Pratik C
Advisors: Dr. Wesley Snyder, Committee Member
Dr. Mark White, Committee Member
Dr. Edward Grant, Committee Chair
Abstract: Faults, if remained undetected lead to huge financial losses. Troubleshooting faults requires analytical skills and knowledge of the system. Elementary skills help in solving common faults but faults that are infrequent or those which cannot be debugged by a consumer or even a field technician due to lack of in-depth knowledge of the system requires an expert's intervention. Experts are not always available in the field and lack of knowledge on the field technician's part may lead to huge financial losses and inconvenience on the part of the consumer. To avoid such situations an approach to make available expert's knowledge and capability to the field technicians is demonstrated through this research. A qualitative modeling approach to develop a knowledge representation is outlined here. A qualitative knowledge of the system from expert's point of view is modeled for a Bosch dishwasher for whom knowledge is extracted to generate a rule base, which reasons the characteristics of the system like an expert. Further it has been explored that using the rule base one can develop a questionnaire using the concept of model based diagnosis for troubleshooting. A qualitative model is useful for this kind of system because it resembles the analytical thinking capability of an expert and in general for any human being making it easier to communicate and interpret. Thus providing such a system to the field technician or even the consumer will simplify the troubleshooting process and will help in reducing downtime and financial losses.
Date: 2006-06-01
Degree: MS
Discipline: Electrical Engineering
URI: http://www.lib.ncsu.edu/resolver/1840.16/815


Files in this item

Files Size Format View
etd.pdf 3.114Mb PDF View/Open
code.zip 27.93Kb Unknown View/Open

This item appears in the following Collection(s)

Show full item record