Share Email Print
cover

Proceedings Paper

Open architecture framework for distributed real-time fault diagnosing
Author(s): Hongan Wang; Guozhong Dai
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Fault diagnosing has recently received considerable attention. Its applicability is severely restricted by the computational complexity involved in the diagnosing process. In this paper, it is presented an open architecture framework for distributed real-time fault diagnosing in an industrial environment. Within this framework, all work units comprising the manufacturing system are modeled using intelligent agents. Each agent has autonomous and cooperative functions, observing the local fault diagnosing process via the combination of model-based qualitative and heuristic reasoning with learning techniques; and cooperatively reporting the primary diagnosis result to the diagnosis manager, which supervises the global fault diagnosing process using the constraint propagation technology. To fast the constraint propagation, a matrix- based constraint parallel propagation method is discussed, which greatly reduces the computational overhead. Also, a formal description method for agent's structure is given, which can be used to describe agents and verify their functions.

Paper Details

Date Published: 4 February 1999
PDF: 8 pages
Proc. SPIE 3589, Process Control and Sensors for Manufacturing II, (4 February 1999); doi: 10.1117/12.339954
Show Author Affiliations
Hongan Wang, Institute of Software (China)
Guozhong Dai, Institute of Software (China)


Published in SPIE Proceedings Vol. 3589:
Process Control and Sensors for Manufacturing II
David M. Pepper, Editor(s)

© SPIE. Terms of Use
Back to Top