摘要
动态电压恢复器是解决电压暂降等电能质量问题的有效方法。针对电压暂降特征量(暂降幅值、相位跳变、持续时间)实时检测的需要,在电网三相电压的对称分量法模型基础上,提出了改进的基于最小均方学习规则神经网络的特征量参数估计算法。该算法具有运算量小,检测精度高,动作快的特点。在理论分析的基础上,对所提算法和常规dq变换法进行了仿真比较,结果验证了该算法的可行性。
Dynamic voltage restorer is an effective method to solve voltage sag and other power quality problems. Due to the need for real-time detection of voltage sag characteristics （temporary decline in magnitude, the phase transition, the duration）, this paper presents characteristics estimation algorithm of an improved neural network learning rule based on least mean square, this algorithm has a small amount of operation, high precision and fast action. After the theoretical analysis, the proposed algorithm and conventional dq transformation method are compared in simulation, and the results prove its feasibility.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2012年第11期82-86,92,共6页
Power System Protection and Control
基金
甘肃省电网公司科技项目(2010406011)
甘肃省自然科学基金项目(0809RJZA006)~~
关键词
动态电压恢复器
对称分量
最小均方学习规则
参数估计
dynamic voltage restorer （DVR）
symmetrical components
least mean-square learning rule
parameter estimation
