基于冗余卷积编解码器的变压器噪声抑制

Research on Transformer Noise Suppression Based on Redundant Convolutional Encoder Decoder

变压器运行过程中产生的声信号是评估变压器运行状态以及变电站噪声水平的重要依据。提出一种基于冗余卷积编解码器网络(redundant convolutional encoder decoder,RCED)的变压器噪声抑制方法,利用短时傅立叶变换得到干净声信号和含噪声信号的时频特征,构建了变压器声信号降噪模型。通过对某换流站换流变压器的噪声信号进行测试验证,结果表明:所提模型具有较好的降噪效果,所提方法对于变压器声纹振动在线监测系统以及变电站噪声的准确检测具有参考意义。

The acoustic signal generated during the operation of the transformer is an important basis for evaluating the operation status of the transformer and the noise level of the substation.A transformer noise suppression method based on redundant convolutional codec(RCED)network is proposed.The time-frequency characteristics of clean acoustic signal and noisy signal are obtained by using short-time Fourier transform,and the noise reduction model of transformer acoustic signal is constructed.The noise signal of converter transformer in a converter station is tested and verified.The results show that the proposed model has good noise reduction effect,and the proposed method has reference significance for the online monitoring system of transformer voiceprint vibration and the accurate detection of substation noise.