Diagnosing multiple failures using knowledge of component states

Author

Holtzblatt, Lester J.

Author_Institution

Mitre Corp., Bedford, MA, USA

fYear

1988

fDate

14-18 Mar 1988

Firstpage

139

Lastpage

143

Abstract

A prototype, called the generic model-based diagnostic system (GMODS), has been implemented to detect and diagnose failures in launch system hardware at the Kennedy Space Center. GMODS diagnoses both single and multiple failures from a model of the structure and function of a physical system. The diagnostic algorithm extends the strategy proposed by J. de Kleer and B. Williams (1987) for efficiently diagnosing multiple failures. GMODS departs from previous work by using knowledge of component states to diagnose failures in components that operate in one of several states. The resulting approach also prunes suspects by using knowledge of the failed-state behavior of components

Keywords

aerospace computing; automatic testing; expert systems; knowledge engineering; GMODS; Kennedy Space Center; component states; failed-state behavior; generic model-based diagnostic system; knowledge; launch system hardware; multiple failures; Artificial intelligence; Collaboration; Combinatorial mathematics; Condition monitoring; Control system synthesis; Fault diagnosis; Hardware; Nose; Prototypes; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Artificial Intelligence Applications, 1988., Proceedings of the Fourth Conference on

Conference_Location

San Diego, CA

Print_ISBN

0-8186-0837-4

Type

conf

DOI

10.1109/CAIA.1988.196094

Filename

196094