Abstract A powerful instrument for condition monitoring for diagnosis of machines is given by the vibration analysis. The envelope analysis as a part of it turned out to be a very effective method for diagnosis of anti friction bearings. Different bearing failures lead to different patterns of the envelope spectrum. Checking the amplitudes of characteristic bearing frequencies is mostly not sufficient for a reliable diagnosis. Therefore human experts compare the pattern of the spectrum with typical patterns of bearing defects. In this paper a fuzzy system is presented which imitates this action of human experts. The knowledge which is necessary for the diagnosis will be stored in the fuzzy rules. In this manner, the system is able to automatically recognize typical bearing defects as well as other kinds of machine failures. The input parameters of the fuzzy system are extracted by an extensive data preprocessing. The presented fuzzy-system is used in different applications in mining and steel industry.
[1]
P. D. McFadden,et al.
Model for the vibration produced by a single point defect in a rolling element bearing
,
1984
.
[2]
Hong Wang,et al.
A Neural Network Approach to the Detection and Diagnosis of Faults in a Hydraulic Test RIG
,
1994
.
[3]
Andreas Seeliger,et al.
Schwingungsdiagnose an Wälzlagern mit Hilfe der Hüllkurvenanalyse
,
1995
.
[4]
Richard P. Lippmann,et al.
An introduction to computing with neural nets
,
1987
.
[5]
J. P. Bonfert.
Using Neural Networks in Nuclear Power Plants for State Diagnosis
,
1994
.
[6]
Hans-Jürgen Zimmermann,et al.
Fuzzy data analysis: methods and industrial applications
,
1994
.
[7]
Andreas Seeliger,et al.
A Low Cost Tool for Diagnosis of Antifriction Bearings
,
1994
.