|
With the social progress of science and technology in the field of energy, transportation, national defense involving people's livelihood, often need to manufacture large, complex, difficult to machine precision parts, so the accuracy of these precision machining parts, CNC machine tools, reliability and life put forward higher requirements. Unconventional conditions of high speed, high acceleration, heavy loads, vibration, shock, deformation and other factors often have a significant impact on CNC machine tools, CNC machine components easily lead to failure or damage. Affect the quality of the workpiece machining the one hand, on the other hand to reduce the reliability of the CNC machine tools, shorten the life span or even scrapped. effective monitoring and diagnosis of rolling bearing fault has important practical and economic significance. Traditional single sensor detection method of the rolling bearing fault detection information is less susceptible to outside interference or sensor failure and other reasons, make the diagnostic accuracy of low, or even result in a miscarriage of justice or a miscarriage of justice. Therefore, based on multi-sensor information fusion technology, techniques and methods of rolling bearing fault diagnostic CNC machine tools. The basic principle is to take full advantage of multi-sensor information fusion of multi-sensor information, and through these sensors detect the reasonable control and use of the information, redundant or complementary information of each sensor in space and time according to some criteria combination produce an explanation or description of the consistency of the object to be measured. The aim is to export multi-optimization of multi-sensor detection information more useful information, and ultimately improve the accuracy and validity of the diagnosis, and to eliminate the limitations of a single sensor. Highlight the advantages of multi-sensor information fusion is information redundancy, fault tolerance, complementary, real-time and low cost. LabVIEW software as a development platform, the design of a CNC machine tool feed system roller bearing fault diagnosis system. By analyzing the characteristics of the rolling bearing fault signal, the selection of the acceleration sensor and the current sensor detects Rolling vibration signal and motor current signal preprocessing, in the NI ELVIS into the NI PCI-6251 data acquisition card based LabVIEW software platform programming, detection signal data acquisition; combined with database technology, phase LabVIEW and Microsoft Access database, database management system, data storage, query, add, modify, and delete functions; take advantage of LabVIEW software a powerful and convenient hardware match and Matlab software signal processing functions, call the Matlab software Matlab Script Node in LabVIEW using wavelet noise reduction and wavelet analysis of the detection signal; Hilbert transform of the wavelet decomposition of the first layer detail signal demodulation , to obtain the rolling bearing fault signal, and for spectral analysis by fast Fourier transform on the rolling bearing signal; obtained by theoretical calculation Rolling fault characteristic frequency in the spectrum diagram to a rolling bearing fault frequency as the center frequency, forward 2.5Hz and rearwardly 2.5Hz 5Hz frequency bandwidth energy characteristic value of the fault signals as a rolling bearing, after normalization processing obtained characteristic value of the rolling bearing inner ring, the outer ring and the rolling element fault frequency; obtained characterized as characteristic value vector, use the momentum of the BP neural network improved algorithm, ball bearing fault detection of multi-sensor information fusion and eventually construct a neural network training and simulation, high efficiency, high reliability CNC machine tools, rolling bearing fault diagnosis system.
|