Research on Unbalance Signal Extraction under the Constraints of Low Operation Speed and Limited Acquisition Time

Abstract

Aiming at extracting sinusoidal signals from strong background noise and disturbance under the constraint of limited cycles of samples, the spectral feature of the picked unbalance vibration signal of a hard bearing balancing machine was analyzed and an analogue filter was designed to eliminate the main disturbance components in the frequency domain, and a signal extension method based on the AR model was introduced and investigated. Simulation and field experiments demonstrated the feasibility of the presented extension method and an improvement in accuracy was achieved by extension of the AR model.