The term LEARNING is understood to mean the process of assimilating experience.

Computer programs with the ability to learn are programs which process information made available (to the learning program) from outside. Such information may be in the form of measured values, images, acoustic signals etc.

A learning program 'RULEARN' has been developed and is classified into the group of Machine Learning Programs (MLP) and is therefore a part of Artificial Intelligence (AI) applications. The program module is linked to a commercially available expert system (XPS) shell. Corrosion rules which are automatically generated from a factual database containing corrosion data are verified by the program itself via consistency checks as well as by a human expert team. This insures reliability on all generated output of the program.

The rule-finding process is described below by reference to a simplified example of corrosion research in the application of silicon alloyed stainless steels e.g. NICROFER 2509 Si7 for the sulfuric acid branch of the chemical process industry (CPI).

Subject
Steel structures, Materials, Reliability, Acid concentration, Corrosion rate, Sulfuric acid, Uniform corrosion rate, Corrosion testing, Tools, X-ray spectroscopy, Silicon, Corrosion data, Stainless steel
Keywords:
Machine Learning Programs (MLP), Artificial Intelligence (AI), expert system (XPS), silicon alloyed stainless steels, sulfuric acid, database
© 1992 Association for Materials Protection and Performance (AMPP). All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise) without the prior written permission of AMPP. Positions and opinions advanced in this work are those of the author(s) and not necessarily those of AMPP. Responsibility for the content of the work lies solely with the author(s).
1992
Association for Materials Protection and Performance (AMPP)
You do not currently have access to this content.