A operation error prediction method of gateway capacitive voltage transformer based on long short-term memory neural network and multi-task learning strategy

Accurate gateway capacitive voltage transformer (GCVT) operation error prediction can ensure the accuracy and fairness of power trading, which plays a vital role in ensuring the stability of the new power system and improving its operating efficiency. Aiming at the inefficiency of traditional GCVT periodic power failure detection methods, a operation error prediction method of gateway capacitive voltage transformer based on long short-term memory neural network and multi-task learning strategy is proposed, and it is applied to photovoltaic grid-connected gateways and the provincial power supply gateways. First, the multi-order Lagrangian difference method is used to fill in the missing values of the dataset, and all data are checked and corrected in standard format. Then, the ratio error and phase error in the GCVT operation process are predicted in parallel by using bidirectional long-short-term memory neural network, and the multi-task learning strategy is used to speed up the training speed of the model and improve the training effect. Finally, experimental results verify the effectiveness of the proposed method.