一种基于多元线性拟合的多源噪声分离方法  

A Method of Noise Separation from Multiple Sources Based on Multiple Linear Regression

作　　者：严青[1] 张体强[1] 丁少华[1] 何清怀 YAN Qing;ZHANG Tiqiang;DING Shaohua;HE Qinghuai(Sichuan Electric Power Design and Consulting CO.,Ltd.,Chengdu 610041,China)

机构地区：[1]四川电力设计咨询有限责任公司,成都610041

出　　处：《噪声与振动控制》2020年第4期194-198,共5页Noise and Vibration Control

摘　　要：噪声户外监测往往受到多种声源的共同影响,难以对各声源产生的影响进行分离。对在不同源强条件下各声源对预测断面的噪声贡献值进行研究,结果表明各点位在两种工况下受到各声源的噪声贡献值变化趋势均一致,且两种工况下各声源贡献值增量与源强增量均一致。据此采用多元线性拟合法计算出各声源对点位的实际噪声贡献值,并反推出各种声源的实际源强和实际背景噪声,实现对多种声源的分离。利用分离出的各个源强,通过Cadna/A软件计算出各点位的拟合噪声值,其与实测数据吻合较好,表明多元线性拟合法有效可行,可实现各点位实测数据中不同声源贡献值的分离,并对各声源产生的噪声影响分别进行评价。Outdoor noise monitoring is always influenced by several sound sources,which makes it difficult to separate the effect of each sound source from the different sound sources.In this paper,the contribution of each noise source to the receivers in the prediction section from all the sources was investigated under different source intensity conditions.The result gave the similar variation curves of noise contribution to receivers between the initial and supposed source intensities.Moreover,the contribution increase was identical to the increase of source intensity.According to the conclusion above,actual noise contributions of different sources,intensity of each source and background noise were obtained with multiple linear regression And the problem of multi-source separation was solved.Then,the fitted source intensity was imported into Cadna/A to calculate the co-contributed noise value of each receiver.The calculated values show a good agreement with the monitored values,which manifests the feasibility and validity of this method.And what’s more,separation of the combined contribution and individual assessment of the sound sources are accomplished.

关 键 词：声学 变电站 交通 多元线性拟合 贡献值 

分 类 号：O422.6[理学—声学]