Power Grid Line Breaking Identification Method Based on Parallel Power Flow Figure Using Deep Learning

The power flow may get redistributed once fault occurs and transmission line get broken. This paper proposed a power grid line breaking identification method based on parallel power flow Figure using deep learning. The method is based on the convolutional neural network (DN). Firstly, the principle and technical idea of using parallel power flow Figure (PFF) to identify line disconnection are introduced. After that, based on the static PFF concept proposed by the author in the early stage, the trend characteristics of the power flow are added, the concept of PFF is proposed, and the method of generating massive PFF sample sets is given. Finally, a DN that conforms to the PFF dimension is designed, and the method can accurately identify the line disconnection. The example shows that the designed DN structure can extract the spatiotemporal features in parallel PFF samples. Compared with the static PFF method, the accuracy of using the parallel PFF for grid line breaking is significantly improved.