An Attentional Interactive Deep Learning Model for Multi-source Electrical Consistency Checking in Power Grid Measurement Center

With the rapid development of the Internet and big data, artificial intelligence technology represented by deep neural network, DNN, has ushered in a golden period of development. The application of deep learning method in the field of power grid data checking has also attracted more and more scholars' attention. A deep learning model based on attentional interaction module is proposed for multi-source electrical consistency checking. Firstly, the circuit switch and the grid data are mapped into a fixed length vector, and the corresponding embedding matrix of the switch and data is obtained. Then, the embedded matrix is sent into the attentional interaction module to obtain an embedded matrix that comprehensively considers all elements of the switch and data, and is sent to the deep neural network after adding the previous embedded matrix to extract the intrinsic characteristics of the switch and data. Finally, the vector representation of switch and data is obtained and the matching score is calculated. Experimental results show that the proposed model is effective for multi-source electrical consistency checking.