| 引用本文: | 蔡谦,钱勇,徐治仁,王辉,盛戈皞.基于广义二次相关和改进飞蛾扑火算法的变压器局部放电定位技术[J].电力自动化设备,2025,45(7):218-224 |
| CAI Qian,QIAN Yong,XU Zhiren,WANG Hui,SHENG Gehao.Transformer partial discharge location technology based on generalized second cross correlation and improved moth flame optimization algorithm[J].Electric Power Automation Equipment,2025,45(7):218-224 |
|
| 摘要: |
| 在当前变压器局部放电定位研究中，针对存在复杂噪声环境下对局部放电信号处理不足、信号时延估计误差大、由时延误差引起的定位算法失效等问题，提出了一种基于广义二次相关和改进飞蛾扑火算法的变压器局部放电定位技术。对测得的特高频信号采用广义二次相关求得信号的时延，具有抗噪性能好的优点；对基本飞蛾扑火算法进行改进，对定位方程问题进行求解；采用改进飞蛾扑火算法和几种传统智能优化算法对基本检测函数进行求解，对比最优目标函数值、运算时间和迭代曲线，证明该改进优化算法的正确性和速度性；针对定位检测的误差，采用密度聚类算法，传感器阵列对局放多次测量并对检测到的信号进行排列组合，对得到的多个局放源定位结果基于密度进行聚类，取最大簇的几何中心位置作为最终的局放源位置。通过仿真和现场实验，验证了所提定位检测方法的有效性。 |
| 关键词: 变压器 局部放电 定位 广义二次相关 飞蛾扑火算法 密度聚类算法 |
| DOI：10.16081/j.epae.202504006 |
| 分类号:TM41;TM855 |
| 基金项目: |
|
| Transformer partial discharge location technology based on generalized second cross correlation and improved moth flame optimization algorithm |
|
CAI Qian, QIAN Yong, XU Zhiren, WANG Hui, SHENG Gehao
|
|
School of Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
|
| Abstract: |
| In the current research on partial discharge location of transformers, aiming at the problems such as insufficient processing of partial discharge signals in complex noise environments, large estimation errors in signal delay, failure of location algorithms caused by time delay differences, a transformer partial discharge location technology based on generalized second cross correlation(GSCC) and improved moth flame optimization algorithm is proposed. Using GSCC method to obtain the signal delay of the measured ultrahigh frequency signal has the advantage of good noise resistance. The basic moth flame optimization algorithm method is improved and the location equation problem is solved. The improved moth flame optimization algorithm and several traditional intelligent optimization algorithms are adapted to solve the basic detection function, and the correctness and speed of the improved optimization algorithm are proved by comparing the optimal objective function value, operation time and iteration curve. To address the error of location detection, the density based spatial clustering of applications with noise algorithm is adapted, the sensor array measures partial discharge multiple times and arranges and combines the detected signals. The location results of local discharge source are clustered based on density, the geometric center position of the largest cluster is taken as the final partial discharge position. The effectiveness of the proposed location detection method is verified through simulation and field experiments. |
| Key words: electric transformers partial discharge location generalized second cross correlation moth flame optimization algorithm density based spatial clustering of applications with noise algorithm |
