基于反卷积波束形成算法的干式变压器异响故障识别技术

Deconvolution Beamforming Algorithm Based Abnormal Noise Fault Identification of Dry-Type Transformer

针对常规波束形成算法定位精度不高的问题,将基于反卷积变换的波束形成算法应用于干式变压器异响故障识别,分析了反卷积变换的波束形成算法基本原理及其用于干式变压器异响故障识别的可行性;研究了一种采用异响精准定位联合声纹图谱特征识别的干式变压器异响故障识别方法;提出了“高频特征峰能量比”的概念,用于量化机械异响严重程度;最后通过实验测试和现场验证的方法,证明方法的有效性和准确性。

To improve the accuracy of the conventional beamforming location algorithm,the deconvolution beamforming algorithm is proposed for the abnormal noise fault identification of dry-type transformer.The basic principle of deconvolution beamforming algorithm is analyzed,and its applicability to the dry-type transformer abnormal-noise fault identification is verified.A dry-type transformer fault identification method based on the accurate location of abnormal-noise is studied,where the feature recognition of voice print is considered.The concept of"the energy ratio of high-frequency characteristic peak"is firstly proposed to quantify the severity of mechanical abnormal noise.Finally,experimental test and field verification validate the effectiveness and accuracy of the proposed method.