基于KPCA算法的环网柜故障检测方法

Fault Detection Method for Ring Main Unit Based on KPCA Algorithm

为保证供电系统的安全运行,针对智能电网中环网柜故障检测模型精度低,非线性变量泛化能力差等问题,提出了一种基于KPCA算法的环网柜故障检测建模方法。PCA算法是故障检测的常用方法,为了解决非线性问题,使用核函数建立核主元模型提取环网柜系统的非线性冗余信息,通过非线性映射将输入空间映射到特征空间,再计算特征值问题。将构建的KPCA模型应用于环网柜系统故障检测,采集环网柜内的多变量信息和环境变量信息,将变量数据空间分解为2个正交互补子空间,分别在特征空间构造T 2统计量和残差空间构造Q统计量进行监控,实现环网柜系统的故障报警。经过对正常数据和故障数据的仿真实验结果表明,该KPCA算法在准确检测故障的前提下,能够有效降低模型的故障误报率,改善了环网柜故障检测效果。

For ensuring safety operation of power supply system,a method of fault detection method for the ring main unit based on KPCA algorithm is proposed for the problem of low accuracy of fault detection model and poor generalization ability of nonlinear systems for ring main unit in smart grid.The PCA algorithm is a common method for fault detection.In order to solve the nonlinear problem,the kernel function is used to construct the kernel principal component model to extract the nonlinear redundant information of the ring main unit system,and the input space is mapped to the feature space through the nonlinear mapping,and then the eigenvalue problem is calculated.The constructed KPCA model was applied to the fault detection of the ring main unit system,multivariate information and environment variable information in the ring main unit are collected,the variable data space is decomposed into two orthogonal complementary subspaces,and T 2 statistics and Q statistics are constructed respectively in the feature space and residual space for monitoring and fault alarm of the ring main unit system is realized.The simulation results of normal data and fault data show that the KPCA algorithm can effectively reduce the fault false alarm rate of the model and improve the fault detection effect of the ring main unit under the premise of accurately detecting the fault.