A Knowledge Extraction System Based on Weight Optimization Applied and Evaluated to Distribution Network Fault Assistant Decision

At present, in the event of emergency fault, the power system is still highly dependent on manual analysis and calculation of text procedures and experience knowledge to make decisions, resulting in a large amount of waste of resources, and greatly prolonging the fault time. Therefore, a knowledge extraction system based on weight optimization is proposed in this paper. Firstly, a data preprocessing module is established to vectorize the unstructured data and form a word vector set that can retain the original semantics. Then Bi-LSTM module is used to extract the entity from the word vector set. Secondly, the weight optimization module is used to focus on the key knowledge in the text data to complete the relationship extraction. Then the error correction module trains the relationship between adjacent labels to obtain the global optimization of text labels. The simulation results show that the system can assist the decision making of distribution network fault, and evaluate the working condition of the system in real time according to the output results, which not only saves the decision time, but also greatly reduces the decision error rate.