基于高阶累积量的变压器多局放源定位算法研究

Research on Multi-partial Discharge Source Location Algorithm of Transformer based on High-order Cumulant

变压器内部局部放电源的定位是电气设备绝缘领域所面临的技术难题。本文中作者研究了一种用于检测电力变压器局部放电的L字型超声阵列,传感器由7阵元L字形超声波传感器构成。通过高阶累积量的孔径扩展特性对L字型阵列进行虚拟扩展,使扩展后的阵列具有16阵元平面阵列的估计性能,提升了阵列的估计范围,降低了电路的复杂度,减少了硬件成本。通过试验验证了该算法具有与经典MUSIC算法4×4平面阵列DOA效果相当的估计效果,并在高斯色噪声的背景下对比了MUSIC-LIKE算法与MUSIC算法的估计效果,结果显示MUSIC-LIKE算法在相同条件下具有更好的估计精度,能够较好的抑制高斯色噪声干扰,更符合实际工程情况的应用。

The positioning of partial discharge power supply in transformer has always been a technical problem in the field of electrical equipment insulation.In this paper,an L-shaped ultrasonic array for detecting partial discharge of power transformers is studied.The sensor is composed of 7-element L-shaped ultrasonic sensors.The L-shaped array is virtually extended by the aperture expansion characteristics of high-order cumulants,so that the extended array has the estimation performance of a 16-element planar array,which improves the estimation range of the array,reduces the complexity of the circuit,and reduces the hardware cost.Through experiments,it is verified that the proposed algorithm has an estimation effect comparable to the classical MUSIC algorithm's DOA effect of 4×4 planar array.In addition,the estimation effect of Music-like algorithm and MUSIC algorithm is compared under the background of Gaussian color noise.The results show that Music-like algorithm has better estimation accuracy under the same conditions.It can restrain the interference of Gaussian color noise better,and is more in line with the actual engineering application.