Overview
FrançaisABSTRACT
Alternative methods exist for the estimation of thermophysical properties. These methods are based on artificial intelligence tools. These tools are artificial neuronal networks and neuro-fuzzy systems. They allow for estimating the thermal diffusivity of a homogeneous material. Indeed, a structural defect locally modifies this property and this is why knowing this property can be useful in non-destructive testing (NDT). The approaches described can easily be applied to other issues concerning the estimation of parameters or properties .
Read this article from a comprehensive knowledge base, updated and supplemented with articles reviewed by scientific committees.
Read the articleAUTHORS
-
Olivier FAUGEROUX: Doctorate from the University of Perpignan Via Domitia - Senior Lecturer, University of Perpignan Via Domitia, PROMES-CNRS UPR 8521 Laboratory
-
Stéphane GRIEU: Doctorate from the University of Perpignan Via Domitia - University Professor, University of Perpignan Via Domitia, PROMES-CNRS UPR 8521 laboratory
-
Adama TRAORE: Doctorate from the University of Montpellier II - Senior Lecturer, University of Perpignan Via Domitia, PROMES-CNRS UPR 8521 Laboratory
-
Jean-Luc BODNAR: Doctorate from the University of Reims Champagne-Ardenne - Senior Lecturer, University of Reims Champagne-Ardenne, GRESPI laboratory
-
Bernard CLAUDET: Doctorate from the University of Perpignan Via Domitia - University Professor, University of Perpignan Via Domitia, PROMES-CNRS UPR 8521 laboratory
INTRODUCTION
The application of a non-destructive testing method requires the interpretation and/or exploitation of the various results it provides. It is therefore essential to have a good command of the tools used for this purpose, which can be carried out simply by using the manipulator's/experimenter's know-how. Different mathematical tools can be used, however, which are much less subject to the manipulator's subjectivity.
A classic approach is to use inverse methods. This involves comparing a behavioral model with the measurements made. In this way, it is possible to trace back the properties or parameters under consideration, simply by ensuring that the difference between the measurements and the model is negligible. This approach requires perfect knowledge of the system under study, and the ability to use this knowledge to model it in detail, otherwise the various properties sought may not be correctly identified. Parametric models can be considered. However, their generalizability is often very limited.
One solution is to use tools that enable models to be developed from examples (or case studies), and which are then able to generalize by exploiting the information learned. These tools, artificial neural networks and neuro-fuzzy systems (for those we tested), belong to the field of artificial intelligence (AI). These systems also take expert knowledge into account. Finally, the possibilities offered by genetic algorithms are currently being explored.
The aim of this document is, after presenting the AI tools used and through a simple experimental example, to show their possibilities. The key points for implementing the proposed approaches are detailed so that an experimenter can adapt them to a specific estimation problem. However, a basic knowledge of systems modeling is strongly recommended to obtain satisfactory results.
Exclusive to subscribers. 97% yet to be discovered!
You do not have access to this resource.
Click here to request your free trial access!
Already subscribed? Log in!
The Ultimate Scientific and Technical Reference
KEYWORDS
building | materials | Non-destructive testing | Materials characterization | Properties estimation
This article is included in
Software technologies and System architectures
This offer includes:
Knowledge Base
Updated and enriched with articles validated by our scientific committees
Services
A set of exclusive tools to complement the resources
Practical Path
Operational and didactic, to guarantee the acquisition of transversal skills
Doc & Quiz
Interactive articles with quizzes, for constructive reading
Artificial intelligence tools applied to NDT
Bibliography
Patents
BODNAR (J.L.), CANDORE (J.C.), CARON (D.), NICOLAS (J.L.) . – Système de contrôle non destructif étendu (SAMMTHIR), French patent filed with INPI on March 26, 2008, number 08/01646.
Exclusive to subscribers. 97% yet to be discovered!
You do not have access to this resource.
Click here to request your free trial access!
Already subscribed? Log in!
The Ultimate Scientific and Technical Reference