Method and device for discovering text deviation of power grid regulation and control system, and storage medium

• Main Authors: YUAN ZHONGCHEN, ZHANG JIE, KONG YANRU, PENG CHENG, SHANG JINGAN, CHEN JIAN, HUANG XINJIAN, CHEN ZIYUN, GAO SHANG
• Format: Patent
• Language: Chinese; English
• Published: 24.11.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention relates to an electric power regulation and control system text deviation discovery method and device and a storage medium, and the method comprises the following steps: 1, obtaining power grid topological data and accident plan text data in an electric power regulation and control system text, carrying out the preprocessing, constructing a power grid structure map, and obtaining an accident plan word vector; 2, extracting entities and relationships in the accident plan word vectors in the step 1 by adopting a pre-training language model; step 3, constructing a multi-domain knowledge graph based on the power grid structure graph constructed in the step 1 and the entities and relationships in the accident plan word vectors extracted in the step 2; 4, converting the multi-field knowledge graph constructed in the step 3 into vector representation by adopting a variational graph auto-encoder; and 5, taking the vector obtained in the step 4 as input, calculating an abnormal score of the vector by ado