A Neuro-fuzzy Network for the Condition Monitoring of Rotating Machines

R.E. Tessendorf, A.J. Hoffman, and N.T. van der Merwe (South Africa)


neurofuzzy networks, pattern recognition, conditionmonitoring


Online condition monitoring systems for assessing the health of rotating machinery have been developed as a result of the need to limit unnecessary outages in industrial applications. A condition monitoring system should use non-invasive measurement techniques and interpret the data. Classifiers making use of artificial intelligence can be used to automate the diagnostic process. Fuzzy logic basedclassifiers are useful for incorporating the experience of human operators in determining the condition of a machine. The ability of neural networks to "learn" quantitative relationships makes them very attractive. By combining neural networks and fuzzy logic it is possible to obtain a classifier, which is capable of incorporating human experience and being able to be optimised relatively easily.

Important Links:

Go Back