524
P. Yang et al.
(a)
(b)
(c)
(d)
Fig. 45.5 The fitness evolution curve of four network models (generation = 200)
45.5 Conclusion
This paper has demonstrated that CSBP neural network model as a useful technique
to recognize roller bearing fault by outer ring failure, Inner ring failure, Rolling
element failure and Cage failure with high classification accuracy for classifying.
Based on the statistics studied here, The CSBP neural network model has a very
high accuracy and stability, accurate values of close to 93% in the classification.
The effectiveness of the newly proposed algorithm is verified by examples of roller
bearing fault.