Research On The Vibration Signal Processing Method For Automatic Balancing Machines

With the development of science and technology, rotating machine plays animportant role in daily life and production. Due to the production processes andenvironmental factors, the imbalance will occur when the rotating machine are in thefactory and after long-term use. The consequence of imbalance is the mechanical vibrationand deafening noise. It will damage the machine or even worse, the human bodies. Toavoid the unbalanced rotor, it needs to be detected regularly. The solution is the BalancingTechnology. And the issue of the dynamic balance of the rotor has been a research hot-spotin recent years. The development tendency of the dynamic balance test system is the hightest accuracy, test speed and the stability of testing performance. The vibration signalprocessing has always been aroused widely attentions of the researchers. This papercentered on not only the unbalance signal which has the same frequency and rotation butalso the vibration signal processing method in the condition of octave interference,non-integer harmonic interference and strong noise interference.This paper demonstrated a vibration signal processing method based on the studies onthe existing method of vibration signal amplitude and phase extraction and the theory ofwavelet. The demanded unbalanced signal can be obtained by the processing of band passfiltering with the method of Non-orthogonal Morlet Complex Continuous WaveletTransform. Due to the good symmetry of complex Morlet continuous wavelet, it has agood performance in time and frequency domain. Through the method, each component’samplitude and frequency of the vibration signals can be obtained according to thecharacteristic of complex wavelet. However, the wavelet scale factor cannot be set unlessthe components of the signal and the approximate frequency are acquired. Thus, this paperproposed a method that fitting the signal with the least square method and calculated thenumber of imbalance signals with different frequency and the approximate frequency. Withthe Morlet complex continuous wavelet transform, the measurement speed and accuracyare improved. The simulation results show that the amplitude and frequency whichobtained by the proposed algorithm has a better performance than FFT. The vibrationsignal’s amplitude and frequency can be extracted with Morlet complex continuouswavelet transform. The phase can be obtained with a simple method of the all-phase FFTspectrum analysis with a high accuracy and calculation speed. The paper combined the Morlet complex continuous wavelet transform with the all-phase FFT spectrum analysis toextract the amplitude and phase of the vibration signal. The simulation experiment showsthat in case of small unbalance, the proposed method can extract the vibration signal’samplitude and phase.