电力学报2025，Vol.40Issue(1)：25-32,8.DOI:10.13357/j.dlxb.2025.003
采用多通道最小二乘法的变压器虚拟声源提取及其远场噪声预测
Extraction of Virtual Acoustic Sources and Prediction of Far-field Noise for Transformers by a Multichannel Least-mean-square Method
摘要
Abstract
In order to simplify the modeling,improve the accuracy and achieve the real-time prediction for the noise prediction of the transformers inside a substation,we propose a method to extract virtual acoustic sources and predict the far-field noise of the transformers via a multichannel least-mean-square algorithm.The trans-formers are treated as equivalent virtual acoustic sources distributed around the radiating surfaces.The matrix of frequency response function from the sources to the surrounding monitoring points and the vector of the sound pressures at the monitoring points are then measured for extracting the source signals via the multichannel least-mean-square algorithm.Meanwhile,the frequency response functions at some far-field monitoring points are measured and combined with the extracted sources to predict the far-field noise radiating from the transformers.The errors of the predicted noise with respect to the measured results are all less than 2 dB,indicating that the proposed source extracting and noise predicting method is valid.The proposed method can be used to achieve real-time prediction of the far-field noise radiated from transformers with microphones located nearby.The pro-posed method can also be used to troubleshoot the sources of transformer noise issues.关键词
变压器/噪声/传递路径分析/多通道最小二乘法/虚拟声源Key words
transformer/noise/transfer path analysis/multichannel least-mean-square method/virtual source分类
信息技术与安全科学引用本文复制引用
刘联鋆,王玉兴,曾海燕,李豪,陈俊文..采用多通道最小二乘法的变压器虚拟声源提取及其远场噪声预测[J].电力学报,2025,40(1):25-32,8.基金项目
国网湖北省电力有限公司武汉供电公司科技项目(15CA02-9012008-0507). （15CA02-9012008-0507）
