Abstract:The main purpose of this paper is to explore the application scenarios of text mining technology in power data. Through the analysis and application of the accumulated defects data during the operation of the power equipment, the automatic recommendation of the elimination repair processing measures is realized, so as to improve the efficiency of maintenance and elimination and reduce the time cost of the work order process. This paper first constructs a professional vocabulary of power equipment. The source of the thesaurus mainly includes the common vocabulary of the power industry, the vocabulary of the equipment category of the Southern Power Grid and the defect data feature vocabulary. Secondly, the unstructured defect text data, such as defect representation, defect cause and defect type, is combined with the constructed professional lexicon to extract the keywords and sort the key degrees. Finally, through the simhash algorithm and the Hamming distance calculation, the defect with the highest similarity that has occurred in the query layer in the defect database is recommended, and its processing measures are recommended as a reference for this defect. In this paper, the above method is applied to successfully implement the recommendation of input defect processing measures, and according to experts'' judgment, this measure can realize the shortage of such defects.