【摘要】 地铁直流牵引网中经常出现的低频振荡电流导致继电保护系统误动频繁,欲提高保护系统的可靠性必须采取更为有效的特征提取方法。经验模态分解(EMD)和能量权重结合的特征提取方法可有效区分牵引网振荡电流与故障电流,实现对牵引网电流动态特征信息的提取。通过对牵引网电流进行EMD分解并计算各本征模态函数(IMF)的能量权重,构建基于能量权重的多尺度特征熵,进而获得故障模式识别的特征矢量。现场实测数据证明,该方法可准确识别牵引网振荡电流与故障电流,且概念清晰、算法简单。更多还原

【Abstract】 Due to the frequent malfunctions of relay protection system caused by low-frequency oscillation current of Metro DC traction network, more effective feature extraction methods are in need to improve the reliability of protection system. The extraction method combined empirical mode decomposition(EMD) and energy weight can effectively distinguish oscillation current and fault current of traction network, and achieve the extraction of dynamic feature information in traction network current. In this method, the traction network current is decomposed by EMD and the energy weight of each intrinsic mode function(IMF) is calculated, and the feature vector of fault pattern recognition is obtained by constructing the multi-scale feature entropy based on energy weight. The actual measurement data illustrates that the new method not only can accurately identify oscillation current and fault current of traction network, but also has the advantages of clear concept and simple algorithm.更多还原