Integration of neural networks and expert systems for microscopic wear particle analysis

A loosely coupled integration of neural networks and expert systems, using simple message interaction, has been developed for the analysis of microsco pic wear particles from machinery. The neural networks are used to classify particle features, whilst a knowledge based system is used for overall wea r assessment (using the feature classifications). This type of integration maintains the individual strengths of neural networks and expert systems. T he particle features are obtained from a computer vision system linked to t he neural networks, and the wear assessment relies on a knowledge base deri ved from human experts. The system is adaptable, extendable, modular and fa st. (C) 1998 Elsevier Science B.V. All rights reserved.