基于麦克风阵列和SRC-PHAT的发动机异响定位研究

Engine Abnormal Noise Positioning Based on Microphone Array and SRC-PHAT

为了实现对柴油机气门异响准确定位,采用NI数据采集系统采集柴油发动机的噪声信号,并用A计权系数滤波器对噪声信号进行滤波,获得类似人耳接收到的噪声信号,经过滤波后的信号低频干扰较少,兴趣信号的能量比例由12.62%提高到29.79%。采用随机区域收缩相位变换加权(Stochastic Region Contraction Phase Transform Weighed,SRC-PHAT)方法建立初始搜索空间区域,计算该区域随机选取点的可控响应功率(Steered Response Power,SRP),再以这些点处SRP值较大的区域建立搜索空间区域进行迭代运算,直到搜索区域范围达到目标精度即为异响源近似位置。试验结果表明,该方法对柴油发动机异响定位具有良好精度,为发动机异响噪声定位提供了参考。

In order to achieve accurate positioning of diesel engine abnormal noise,an NI data acquisition system was used to collect the noise signal of diesel engine and the noise signal was filtered with an A weight coefficient filter to obtain a noise signal similar to that received by the human ear.The signal had less low frequency interference and the proportion of meaningful signals increased from 12.62%to 29.79%.By using the Stochastic Region Contraction Phase Transform Weighed(SRC-PHAT)method,the initial search space region was established and the steered response power(SRP)at the randomly-selected points in the region was calculated.A new search space area was established with the area with a large SRP value so that the iterative operations were conducted until the target area accuracy of search area reached the approximate position of abnormal noise source.The experimental results show that this method has good accuracy for diesel engine abnormal noise positioning and provides the reference for the positioning of engine abnormal noise.