配电网络电压暂降源自动定位与智能识别

Automatic location and intelligent recognition of voltage sag source in distribution network

电压暂降是配电网络中最重要的电能质量问题之一.针对短路故障、电容器投切、冲击负荷和变压器启动引起的四种常见电压暂降类型,分别搭建仿真模型分析其扰动信号特征.采用扰动功率和扰动能量法实现其扰动方向判定,并针对传统的基于矩阵算法的暂降源自动定位方案提出改进措施.分别基于小波能量熵和自组织竞争性神经网络进行暂降扰动信号的特征提取和分类训练,可实现多类型电压暂降源的智能识别,算例分析验证了所提方法的实用性和有效性.

Voltage sag is one of the most important power quality problems.The simulation models of four kinds of voltage sags are built respectively,which are caused by short-circuit fault,capacitor switching,impact load and starting transformer,and the characteristics of those disturbance signals are analyzed. The disturbance direction judgment algorithm based on disturbed power and disturbed energy is adopted to judge the disturbance direction.The existing automatic location scheme based on the matrix algorithm is improved.The features of sag disturbance signals are extracted and classified training based on the wavelet energy entropy and the self-organizing competitive neural network.By this way,multi-type voltage sag sources can be identified intelligently.The case verifies the practicability and validity of the proposed method.