The Study On The Magnetoelectric Vibration Velocity Sensor Low-frequency Characteristic Compensation

Vibration is a kind of motion in nature. There are abundant of useful information in the vibration signal. The failure features can be acquired by measuring, processing and analyzing the vibration signal. But the vibration frequency of the large scale equipment is lower. For example, the vibration frequency of the highway bridges or railway bridges is about 2 Hz, which have a span of 30~60 meters. The vibration frequency of the large-scale hydroelectric generator set is about 1 Hz. But now the vibration sensors can't measure the low-frequency vibration signal because of the higher Inherent frequency. The output signal of the vibration sensor is almost submerged in the noise. To solve the above problem, the paper introduced the low-frequency characteristic compensation method, which could be satisfied with the low-frequency respond measurement.Through analyzing the difference among the fiber-optical sensor, piezoelectric acceleration sensor and magnetoelectric vibration velocity sensor, finally, the magnetoelectric vibration velocity sensor was selected. Through analyzing its principle, the mathematical model was built, and then proved correct by measurement. The paper analyzed the passive compensation, feedback compensation, pole-zero configuration compensation, and system identification compensation respectively, finally the pole-zero compensation was selected. The compensation network parameters were calculated by the correlation theory and then proved workable by the ORCAD simulation experiment. The circuit parameters were adjusted by consulting the experimental data and the compensation circuit characteristic. Finally, the design demand was satisfied. The cut-off frequency was extended to low-frequency region more than 10 frequency interval. Linearity was improved highly.Through analyzing the characteristics of the IIR filter and FIR filter,the paper utilized bilinear transformation to design IIR filter. And the paper utilized the MATLAB software to simulate the digital filer circuit to digitize the low-frequency response compensation of the sensor. The result proved that the digital compensation network was better than the active compensation network,which provided a better base for the ulterior exploitation of the design. Finally, the summarization and prospect of this study were proposed,and the next plan was put forward.