故障极识别与故障测距一体化的智能算法  

Intelligent Algorithm Integrating Fault Pole Identification and Fault Location

作　　者：毛王清 杨琦[2] 杨明玉[1] 

机构地区：[1]华北电力大学电气与电子工程学院,河北保定071003 [2]中国电力科学研究院,北京100192

出　　处：《陕西电力》2015年第6期43-46,共4页Shanxi Electric Power

摘　　要：通过归纳现有的基于神经网络直流输电故障定位方法,分析了故障定位过程中存在的问题,提出了一种针对双极HVDC系统,可实现故障极识别和故障定位一体化的智能化算法。提出分层分布式神经网络结构,依据人工神经网络较强的模式识别功能建立特有的故障极识别模块。采用粒子群算法优化神经网络权值,避免陷入局部极小值点,加快收敛速度。Matlab和PSCAD的联合仿真验证了该算法的精确性。Through the summary of existing fault location based on neural network, the problem in the process of fault location is analyzed, a new intelligent algorithm for bipolar HVDC fault location is proposed, which integrating the identification of fault pole and fault location. A hierarchical and distributed neural network structure is put forward, the unique fault type identification module is established based on strong pattern recognition of artificial neural network. The particle swarm algorithm is adopted to optimize neural network weights, can avoid plunging into the local minimum point, and accelerate the convergence speed. The co-simulation of MATLAB and PSCAD proves the precision of the algorithm.

关 键 词：HVDC 故障极识别 故障测距 分层分布式神经网络 粒子群 

分 类 号：TM773[电气工程—电力系统及自动化]