基于长短期记忆网络的电网故障区域定位与故障传播路径推理

为了在发生故障后维持电力系统的安全稳定,有必要实现对故障区域的快速定位并确定故障冲击的传播路径,提出基于长短期记忆网络(LSTM)的故障区域定位和故障传播路径推理方法.首先,利用LSTM建立2个故障诊断模型分别实现在线检测故障时刻和确定故障区域;然后,通过计算故障点附近线路的端口供给能量确定故障冲击的传播路径;最后,以8机36节点电网为例进行算例验证,结果表明所提模型可在发生故障后短时间内检测到故障,给出故障区域和故障冲击传播的路径,且对噪声有较强的鲁棒性.

Fault section location and fault propagation path reasoning of power grid based on LSTM

To maintain the security and stability of power system after failure, it is necessary to locate the fault section quickly and determine the propagation path of the fault impact, a fault section location and fault propagation path reasoning method based on LSTM(Long Short-Term Memory network) is proposed. Firstly, two fault diagnosis models are built by LSTM to realize online fault time judgment and fault section location respectively. Then, the propagation path of the impact caused by the fault can be determined by the ESP(Energy Supply on Port) of the transmission lines near the fault. Finally, the 8-machine 36-bus power grid is taken as an example for verification, and the results show that the proposed model can detect the fault immediately after it occurs, give fault section and fault impact propagation path, and has strong robustness to noise.