Research on Fault Location Algorithm of Distribution Network under Cloud-Edge Collaboration

With the rapid development of the Internet of Things in power distribution, a large number of accesses to power distribution equipment will generate millions of power distribution data. The collection, transmission and calculation of these data will bring great pressure to the communication channel and storage calculation system of the master station, which will greatly affect the reliability of real-time transmission of communication and fault location. In order to solve the above problems, a distribution network fault location scheme under cloud-edge collaboration is proposed. This scheme combines the advantages of centralized grounding line selection and local grounding line selection, and adopts the neural network algorithm to offline train multiple fault features in the substation master station. The generated model is dispersedly deployed at the edge calculation nodes, and the fault diagnosis of the distribution network is completed locally by the edge equipment. The pressure of the distribution master station to deal with various resource information is reduced, and the fault location efficiency of the distribution master station is improved. Finally, PSCAD / EMTDC is used to conduct simulation experiments on various faults of small current grounding system, and the results are verified by neural network algorithm, which proves the adaptability and effectiveness of this scheme.