A Text-Oriented Fault Diagnosis Method for Electromechanical Device Based on Belief Rule Base

At present, quantitative data is often used for fault diagnosis of electromechanical devices, while qualitative in the form text rarely used. In order to integrate and an device, a text-oriented method based on belief rule base (BRB) proposed this paper. Specifically, key information extracted from through natural language processing (NLP) then converted into rules. Then, supplement adopted add rules BRB completion construction. This applies process construction, which new attempt at construction method. It not only solves problem that cannot use but also improves modeling accuracy comprehensive ability BRB. To verify effectiveness algorithm, we designed experiment asynchronous motor case study. The experimental result shows can construct effectively diagnose faults motors. MSE 0.0451, better than traditional (0.1461), BP (0.0613), SVR (0.0974) under same conditions.