直升机模型旋翼声学试验声信号分析中的抑噪方法研究

Study of an Approach of Noises Suppression in Acoustic Tests of a Helicopter Rotor Model

摘要:

将直升机模型旋翼在消声室中进行声学性能测试是研究直升机旋翼噪声特性的一种最有效最直接的方式;在直升机模型旋翼消声室试验中,发现传声器采集到的围绕170 Hz特征频率分布的频带极窄的旋翼声信号上叠加有宽频无用噪声;试验中采集到的声信号信噪比与总距角大小成正比,小总距角时声信号信噪比较小,无用信号成分会掩盖声信号真实信息,必须采用高效的方法去除无用成分并同时最大可能不损坏需要成分;模型旋翼声信号中各成分频带分布不同,信号自相关长度与其分布频带有关,文章指出可以根据信号中各成分之间自相关长度的差异性采用自适应线性预测器把所需信号成分预测分离出来；通过分别对比在总距为0°、6°、10°时声信号与各自处理结果,表明叠加在信号上的无用成分被成功滤除,信号质量得到明显改善。

Abstract:

The acoustic performance test of a helicopter rotor model is one of the most effective and direct way to study the helicopter rotor noise features. In an acoustic test of a helicopter rotor model in the anechoic chamber,it is found that unwanted wideband noises have superimposed on the rotor acoustic signals which are distributed around the 170Hz characteristic frequency in an extremely narrow band. Since the signal to noise ratio (SNR) is proportional to the collective pitch,when the collective pitch is small,a low SNR will lead to real information of the signal be covered by the unwanted signal components. It is necessary to find a way to remove the unwanted components and to remain intact wanted components. Different components of the acquired acoustic signal have different frequency band distributions,and the signal autocorrelation length relies on its frequency band distribution,so in this paper,it points out that the required signal components can be separated using the adaptive linear predictor according to the differences of the autocorrelation length of the components of the signals. The results of the processing of the acquired signals when the collective pitch are 0°、6°、10°respectively show that noises are filtered successfully and the signals are improved significantly.