Improved RNN (Recurrent Neural Network) model for power fault identification system

The invention relates to the technical field of power failure recognition, in particular to an improved RNN model applied to a power failure recognition system, which comprises an identity recognition security unit and a central processing unit, and is characterized in that the output end of the identity recognition security unit is connected with the input end of an electric energy meter unit; the output end of the electric energy meter unit is connected with the input end of the public network unit. The improved RNN model is used for a power fault identification system, the local control center controls the solar panel to work through the local control center and the solar panel, the utilization of a power supply is reasonably realized, the use of a monitored power supply is avoided, the safety and sensitivity of a fault recorder are improved, the service life is prolonged, and the reliability of the fault recorder is improved. Through the arrangement of the data decryptor and the data flash memory module i