| Title |
AI Modeling Methods Research for Detecting Production Facility Problems, Using Power Pattern, Control Characteristics, Operation Condition Data |
| DOI |
http://doi.org/10.5207/JIEIE.2022.36.7.009 |
| Keywords |
AI modeling; Machine learning; Ranked-FBD; Visualization method |
| Abstract |
Although it is necessary to maximize efficiency through systematic maintenance of production facilities, convergence of various professional technologies and knowledge such as materials, machinery, and electrical facilities is required. Systematic management is difficult due to the diversity of expensive measuring equipment and objects. In order to overcome such problems, various experiences (implicit knowledge) of field experts are converted into a form that computers can understand. AI modeling and inference results for production facility abnormalities and quality abnormalities based on translated data (labeling data) are provided on-site Visualization technology is needed for. ① Learning data are generated from raw data of production facilities (operating conditions, operating conditions, and power patterns). ② We present a Ranked-FBD-based visualization method that can compare expert experience and machine learning results. It is expected that the visualization method based on Ranked-FBD and generation of learning data through this study can predictively diagnose unit production facilities (injection molding, press, fusion, painting, etc.) of small and medium-sized parts manufacturers and provide a customized solution. |