Characterization of microphone placement and noise sensitivity in a global active noise control system for a compact noise source

Abstract

This paper presents a sensitivity analysis to error microphone placement and extraneous noise sources for two control source configurations used for the active control of noise radiated from a compact source. The two control configurations analyzed are a linear array of four control sources with the primary source in the middle of the array and a symmetric array of four control sources surrounding the primary source. For modeling purposes, the primary source and the control sources are modeled as simple sources with a baffled free space Green's function to predict the noise measured by the four error microphones. The control source strengths which minimize the squared pressure at the error microphones are solved for based on error microphone locations. If the microphones are placed properly with no error in spatial location, this minimization process has been previously shown to result in radiated power minimization. Simulation results show the linear array is significantly more sensitive to microphone placement errors than the symmetric array. The linear array configuration is also shown to be more sensitive to added random noise in the error signals. Experimental results suggest that for the configurations used, the sensitivity to microphone placement errors was the dominant effect. The reduction in performance for the linear array configuration is significant with microphone placement errors and/or added random noise, while the reduction in performance for the symmetric array is minimal.

An IP Based Audio Test Network for Tunnel Application

“…Thus, the characterization of sound propagation in tunnel is complicated [1] .To control noise, it is important to analyze sound wave for which recording becomes necessary. The sound propagation is related to the sound velocity, distance and time, which requires all recorders to record the wave synchronically in the same phase difference and the same amplitude [2] . A test covered whole wave length needs five sensors at least.…”