Construction and identification of defect knowledge map for DC protection system

Abstract With the continuous acceleration of DC engineering construction, the equipment scale of the DC protection system continues to expand. As the core part of DC transmission, the DC protection system has become an important factor affecting the safe and stable operation of the power grid. The diagnosis and handling of defects in DC protection system equipment often rely on expert experience and technical service support from manufacturers, making it difficult to detect and handle them in a timely manner. Defect analysis and handling are time-consuming and labour-intensive, and disposal experience cannot be shared, posing hidden dangers to the safe and stable operation of the power grid. The DC protection system contains a massive number of components and has extremely complex internal logic. Traditional data mining methods make it difficult to fully extract the semantic features of defect texts in DC protection devices. Therefore, this paper analyzes the types and causes of protection devices in DC transmission systems, uses the Word2Vec-Bootstrapping algorithm to extract the DC protection defect knowledge, uses the Bootstrapping algorithm to classify the word relation, and proposes a modeling method of DC protection device defect Knowledge map. By converting the extracted knowledge into triplet form and constructing a knowledge map, the causes and treatment measures of DC protection defects are diagnosed. An example shows that this method can fully extract the semantic features in the DC protection device defect record text, improve the utilization rate of the defect text record, and accurately identify the defect of the DC protection device.