Research On Dynamic Unbalance Signal Processsing And Calibration Algorithm

Using the dynamic balancing machines as the research object, the thesis will explain the basic principle of low speed rotor dynamic balance and review balancing technology research status in China and other countries. According to the character of balancing signals, the thesis will analyze the shortages of directly using the Fourier transform handling dynamic balance signal amplitudes and phases ;test and verify the collected signals with Matlab tool.In order to solve the defects of Fourier transform in dynamic balance applications, firstly using wavelet denoising processin to research particularly the principle of wavelet denoising ,analysis the advantages and the disadvantages of hard and soft thresholding ,then blend the OSTU algorithm in wavelet transform denoising of dynamic unbalance signals and obtain the good result.The method of using sliding window contro the length of acquisition datas what are calculating the average in each cycel on wavelet transform. Using wavelet transform to separate the noise and other useless information from the sampled non-stationary signal, then analysis the reconstructed signal by frequency spectrum and extract dynamic unbalance signal amplitude.When demarcate the amplitude-unbalance quality, use successive approximation algorithm to determine the wheel not balance phase & give up dematcating unbalance quality and phase at the same time through influence coefficient method ,considering the limited handling ability of MCU,then MCU can calculate dynamic unbalance quality from vector data instead of scalar data.Experimental results show that the traditional Fourier transform is good at handling dynamic balance signals to get the amplitude and phase of unbalance points in theory,but if the signal interferenced by large external noise, the calculated result will deviate , especially in the error of phase. Using wavelet analysis to handle these signals can extract unbalance signal amplitude preferably. Using successive approximation method decreases can decrease phase error in the actual applications & improve the efficiency and accuracy of dynamic balancer.