Application on Virtual Instrument and Neural Networks in the Fault Diagnosis

• Main Authors: Zhang Minghu, Wang Dehu, Lv Shijun, Song Yuxi, Liu Hong, Chen Shaojie
• Format: Conference Proceeding
• Language: English
• Subjects: Analytical models; Data processing; design of hardware and software; Fault diagnosis; fault model identification; Feature extraction; Instruments; Joining processes; Neural network hardware; Neural networks; Signal analysis; virtual instrument technology; Wavelet analysis
• DOI: 10.1109/ICICTA.2009.312

The main point of intelligent fault diagnosis theory is fault mode distinguishing principle based on data processing methods. Pointing to the problems of the traditional fault diagnosis mode, a fault diagnosis method based on the virtual instrument and neural networks is proposed. The signals collection and management based on virtual instrument is introduced, the basic method of the neural networks for distinguishing the faults is analyzed. For fastness and accuracy, connecting the wavelet analysis with the neural networks organically, and based on the wavelet transfer and the neural networks, the system of the speedy features extraction and identification for the faults is founded. The method of the feature extraction for the faults based on the wavelet analysis are established, the realization idea of the fault diagnosis based on the neural networks is put forward, and the hardware and software structure of the fault diagnosis based on the neural networks are discussed. The experimental and simulated results show: it is feasible that analyses for the faults with the neural networks and the wavelet analysis. The method can remarkably heighten the accuracy and credibility of the fault diagnosis results, and the results are of repeatability.