模糊权重小波去噪在变电站声压信号测量中的应用

为了对变电站周边可听噪声进行科学合理的评价,分析了不同时间点噪声分布情况,提出了一种自适应声压背景噪声滤波方法。采用小波去噪算法对声压频谱轮廓进行提取,对小波阈值进行了改进以提高信号平滑度,根据三角模糊数设计了复合评价指标权重以确定最佳小波分解尺度。利用设计的算法得到了声压频谱轮廓,并通过对比不同时间点的声压频谱轮廓确定背景噪声的频段,以此设计带阻滤波器的参数。利用设计的滤波算法对采集的声压信号进行了处理,实验结果表明,本文设计的滤波方法可以很好地滤除声压信号中的背景噪声。经设计的滤波算法处理后的等效连续A声级比未处理的要小5dB以上,可见所设计的滤波算法可有效降低背景噪声的影响,提高变电站可听噪声测量的可靠性。

Application of Fuzzy Weighted Wavelet De-noising Method in Substation Sound Pressure Measurement

In order to estimate the audible noise of the substation correctly, a self-adaptive background noise filter method was proposed according to the analysis of the noise distribution at different time points. The wavelet de-noising method was used to extract the spectrum contour of the sound pressure. The de-noising threshold of the wavelet was modified in order to improve the smoothness of the signal. To determine the best decomposition level, a composite evaluation index weight was proposed based on the triangular fuzzy number method. The spectrum contour of the sound pressure was obtained and the filtering band was figured out by comparing the day and night spectrum contour, as well as the parameters of band pass filter. After the sound pressure was processed by the proposed de-noising method, most of the background noise was filtered out. By comparing the unfiltered signal, the noise levels of the signal processed by the proposed de-noising method is less than 5dB. The results testify that the proposed method can reduce the influence of the background noise. Thus, it can be used to the substation noise level estimation.