Design of a Knowledge-Based Fault Detection and Diagnosis System,
Abstract
This paper presents a design of a knowledge-based fault detection and diagnosis (FDD) component intended to function as part of an intelligent processor (agent) in a distributed problem solving system. This component will permit the system to monitor its own behavior, detect an abnormal system rate, and identify the fault that caused it. Once the fault is identified, the system can repair it or reconfigure itself, thus improving its performance. The heart of the FDD component is a model of the system it is monitoring. This model represents the correct behavior of the system and the criteria for detecting deviations from this behavior. Thus it integrates information necessary for detection and diagnosis into a uniform structure. Detection is accomplished by monitoring the system state and detecting situations which violate the predefined expectations. Diagnosis is accomplished by constructing a representation of the current system state and determining the earliest points of departure from the expected behavior as represented by the correct system model. These points constitute the identified faults. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 20, 1984
- Accession Number
- ADA148859
Entities
People
- E. Hudlicka
- V. Lesser
Organizations
- University of Massachusetts Amherst