基于多重多项式结果-二重网格搜索的变压器局部放电定位

Partial Discharge Location of Transformers Based on Multi-polynomial Resultants Approach and 2-Lattice Search

为快速筛选出变压器局放位置,该文设计了超声和特高频一体化传感器阵列排布,并基于此提出了一种多重多项式结果-二重网格局部放电搜索定位方法。该方法的基本原理为:首先,离线生成两层网格搜寻虚拟空间以对标真实空间;然后采用多重多项式结果法求解时差方程组,得出初始坐标点;之后,以初始坐标点为基础利用二重网格搜索求解目标故障坐标点。为验证方法的有效性,通过搭建变压器局部放电试验平台,内置针-板故障缺陷,对该方法下的故障点定位偏差进行了计算和分析。结果表明:该文设计的一体化传感器阵列时延差异显著;与传统多重多项式直接求解结果相比,二重网格化搜索后的定位结果偏差值分别降低40%和16.7%,2-范数值降低53.5%和6.0%,算法最终定位误差最大不超过5 cm。该方法可为实际油浸式配电变压器放电定位应用提供理论参考。

In order to quickly screen the location of transformer partial discharge,the paper presents a design for an integrated ultrasound and ultra-high frequency sensor array layout,and proposes a searching and positioning method for multiple polynomial roots-double grid local discharge based on this design.The basic principle of this method is as follows:Firstly,a two-layer grid is generated offline to search the virtual space to benchmark the real space;secondly,the multi-polynomial resultants approach is used to solve the time difference equations,and the initial coordinates are obtained;then,based on the initial coordinate points,the 2-Lattice search method is used to solve the target fault coordinate points.In order to verify the effectiveness of the method,a transformer partial discharge test platform was built with built-in needle-board defects,and the fault location deviation under the method was calculated and analyzed.The results show that the delay difference of the integrated sensor array designed in this paper is significant.Compared with the results directly obtained by the traditional multiple polynomial roots method,the deviation values of the localization results after the double-grid search are reduced by 40%and 16.7%,and the 2-norm values are reduced by 53.5%and 6.0%,with a maximum positioning error of no more than 5 cm.This method can provide theoretical reference for the practical application of discharge localization in oil-immersed distribution transformers.