Operation ticket is the basic certificate of operation and maintenance of equipment in intelligent substations, and its main content is the operation steps. Before executing the operation in intelligent substations, the primary work is to verify the operation ticket. At present, it is mainly carried out manually. However, the ratio of operations people to substations decreases year by year. Automatic checking helps reduce the work pressure of operations people and improve efficiency. This paper constructed an automatic check model for operation tickets, and it is mainly based on CNN-BiGRU-Attention. Firstly, the operation ticket texts are vectorized and CNN-BiGRU-Attention is used to judge whether text content was missing and grammar errors in operation tickets. Then, the operation matrix reflecting the change of substation operation state was integrated into the text vector of the operation ticket, and the second CNN-BiGRU-Attention is used to judge whether the operation steps in operation tickets are incorrect or missing. This model can realize the judgment of the correctness and error types of operation tickets. The F1 of the first model is 97.86%, and the F1 of the second model is 93.81%.

operation ticket,intelligent substation,text classification,CNN-BiGRU,attention mechanism