Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Civil-Comp Proceedings
ISSN 1759-3433
CCP: 12
ARTIFICIAL INTELLIGENCE AND CIVIL ENGINEERING
Edited by: B.H.V. Topping
Paper VIII.1

Multi Expert Knowledge Transfer for Metal Forging Fault Diagnosis

N. Ravaille and J.L. Wybo

Ecole des Mines de Paris, CEMEF/EII, Sophia Antiplois, Valbonne, France

Full Bibliographic Reference for this paper
N. Ravaille, J.L. Wybo, "Multi Expert Knowledge Transfer for Metal Forging Fault Diagnosis", in B.H.V. Topping, (Editor), "Artificial Intelligence and Civil Engineering", Civil-Comp Press, Edinburgh, UK, pp 219-227, 1991. doi:10.4203/ccp.12.8.1
Abstract
As part of a long term project, the subject of which is design optimization, our goal is the elaboration of an expert system for defects diagnosis assistance in forging. The main difficulty is the association of scientific and technical knowledge because of involved multi expertises which come from: thermo-mechanical modelling, tribology, metallurgy, design, manufacturing. We have defined an expertise transfer policy in order to gather these knowledge of different origins in a coherent structure. The first part consists in making ourselves used with the application domain. Forging process is described with theoretical and technological parameters. Diagnosis process can be modelled with a two knowledge levels reasoning: an elementary one (context verification) and a methodological one (order in which the concerned contexts are to be verified). The problem to solve must be sum up by the three following difficulties: combination of expertises of different kinds; abstraction level (neither industrial one, nor scientific one) and the bibliography. Following on these specifications, the second part consists in identifying knowledge elements; organizing them in groups which constituants are normalized and in researching the completness by a groups opposition policy. We have implemented this approach in order to generalize the expertises and be able to model them together. The technical aim of 'this software is to gather knowledge about defects diagnosis and to extend it to production setting assistance. Moreover, we try to embed a didactic aspect by providing all the required information for a best communication among scientific and technical experts.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £60 +P&P)