Fault Diagnosis of CNC Machine Using Hybrid Neural Network

This paper proposed a technology of fault diagnosis for CNC machine based on hybrid neural network. Before the fault diagnosis was made, the fault patterns were firstly obtained from technical manuals and field experience to build a sample set, and then they were classified and coded according to th...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Applied Mechanics and Materials 2012-01, Vol.128-129, p.865-869
Hauptverfasser: Yan, Xian Guo, Du, Juan, Wei, Na Sha
Format: Artikel
Sprache:eng
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper proposed a technology of fault diagnosis for CNC machine based on hybrid neural network. Before the fault diagnosis was made, the fault patterns were firstly obtained from technical manuals and field experience to build a sample set, and then they were classified and coded according to the rule. In order to improve diagnosis speed, the fault diagnosis system was designed as a hybrid neural network system which consists of two-grade neural networks. When the fault pattern was input the system, the fault was first classified by the first-grade BP network, and according to the fault type, the corresponding second-grade ART network was activated to perform fault diagnosis. In this paper, the train algorithms of two kinds of neural networks were programmed by MATLAB. Comparing with the traditional diagnosis method, the presented technology possesses advantages of automatic fault diagnosis and ability for self-learning and self-organization.
ISSN:1660-9336
1662-7482
1662-7482
DOI:10.4028/www.scientific.net/AMM.128-129.865