基于压缩感知的旋转声源多普勒效应消除方法

Doppler effect elimination method of rotating sound source based on compressed sensing

消除旋转声源的多普勒效应是监测旋转机械工作状态和定位旋转声源的前提。当前用于消除旋转声源多普勒效应的方法分为频域和时域两种,但是这两种方法都需要在周向均匀布置大量传声器,这不仅增加了测量成本,而且增加了测量工作量。针对上述问题提出了一种基于压缩感知的多普勒效应消除方法,该方法通过在周向随机布置少量传声器完成周向模态幅值的重构,并根据静止柱坐标系与随声源同步旋转的旋转柱坐标系之间的相对运动关系完成多普勒效应的消除。数值仿真和试验研究都表明采用较少传声器随机布置的方法同基于采样定理均匀布置传声器的方法在消除多普勒效应的效果上基本一致,且所提方法可有效降低测量成本和测量工作量。

Eliminating Doppler effect of rotating sound source is the premise of monitoring working state of rotating machinery and locating rotating sound source.At present,methods to eliminate Doppler effect of rotating sound source are divided into two kinds in frequency domain and in time domain,but both of them need to uniformly arrange a large number of microphones around sound source,this increases not only measurement cost but also measurement workload.Here,to solve the above problems,a Doppler effect elimination method based on compressed sensing was proposed.This method could complete reconstruction of circumferential modal amplitudes with randomly arranging a small number of microphones in circumferential directions,and eliminate Doppler effect according to relative motion relations between a static cylindrical coordinate system and a cylindrical coordinate system rotating synchronously with sound source.Numerical simulation and test study showed that the method of random arrangement of fewer microphones is basically consistent with the method of uniform arrangement of microphones based on sampling theorem in eliminating Doppler effect;the proposed method can effectively reduce measurement cost and workload.