Construction of Knowledge Graph of Power Defect Faults Based on Deep Neural Networks Considering Attention Mechanisms

Aiming at the current problem that part evaluation and overall substation equipment evaluation are split and cannot be organically integrated, therefore, this paper proposes a deep neural network to build a knowledge graph based on the association relationship between equipment and parts considering the attention mechanism. First, a multilayer neural network model considering textual attention mechanism and relational attention mechanism is established to build a knowledge map for each substation equipment and its corresponding parts; second, the multilevel and multidimensional association relationships between parts and parts, parts and equipment, and equipment and equipment are analyzed from the electrical space and the physical space, and the equipment association network is formed; finally, based on the association relationships, the knowledge map of parts and equipment is Finally, the knowledge map of parts and devices is interactively mapped based on the association relationships to form a spatial association knowledge map that integrates electrical and physical associations and is oriented to the spatial associations among substation equipment, components, parts and devices.