This paper presents a spectrum expansion and bias compensation method for a magnetoelectric velocity sensor used in low-frequency magnetic levitation vibration isolation systems. To enable accurate measurement in the low-frequency range, a compensation circuit is introduced in series with the sensor, extending its cutoff frequency from 3.9 Hz to 0.3 Hz. To mitigate the zero-drift phenomenon caused by ambient temperature variations, a particle-genetic optimized artificial neural network (PG-ANN) algorithm is developed for bias compensation, effectively reducing the sensor offset to quasi-zero level. A low-frequency magnetic levitation vibration isolation system is further designed using a quasi-zero-stiffness isolator that combines positive and negative stiffness through a reversed magnetic pole structure. This design enhances the magnetic field intensity in the air gap, increases the levitation force to support heavier loads, and reduces the system’s natural frequency to 1 Hz. To suppress low-frequency vibrations, an active control strategy combining absolute velocity feedback and ground feedforward is implemented. Experimental results show that the proposed method reduces vibration transmissibility by more than 30 dB at the natural frequency and achieves less than −40 dB transmissibility at 10 Hz. Comparative testing using sensors before and after spectrum expansion verifies that the proposed method significantly improves the system’s low-frequency isolation performance.
Li et al. (Mon,) studied this question.