中文摘要：

为了增加配电网不对称接地故障的测距精度，减少配电网故障导致的损失，提出一种基于HHT配电网不对称接地故障测距新方法。该方法应用Hilbert—huang变换提取故障零序电流在不同频率段内不同时刻的能量分布特征。应用遗传算法以故障分支、故障类型、故障距离、接地电阻和故障时刻为变量，以不同频率段内能量分布匹配程度为适应度函数，对Matlab／Simulink仿真所得各种不同故障条件下故障零序电流特征与原始故障信号进行特征匹配，利用遗传算法对故障条件进行自动搜索，以能量分布最为匹配的故障点作为输出，故障特征最为匹配的故障条件下的故障点位置即为输出的故障距离。该方法较大程度上提高了配电网不对称接地故障测距的准确性。

英文摘要：

In order to identify the type of fault accurately and reliably in the distribution network, the paper divides faults into two types, namely the asymmetric ground fault and others by using the characteristics of the zero-sequence current, and then extracts the energy distribution at different moments and different frequency bands by the Hilbert-huang transform. Finally, it trains the two neural networks whose samples are the modulus of the current at different frequency bands and the code of the type of the fault with the modulus of the current in different fault conditions imported into the corresponding BP network to get the code of the type of the fault, and then identifies the specific type of fault. A number of simulations prove that this identification method improves the accuracy of various types of fault identification and the fault distance, fault time and the grounding resistance, and the system operation mode has no impacts on the identification.