Mining Sensor Data for Fault Prediction in Wheel Rim Molding
Bildiriler > Özet Bildiri · 2023
Yazarlar: İnanç KABASAKAL
10th International Management Information Systems Conference (IMISC 2023), Türkiye, İstanbul, Uluslararası, Sözlü Sunum, Özet Bildiri, 2023-2023, ss 75-75
Özet
<p>Analysis of process data with data mining helps to discover useful information for root cause analysis and failure prediction. Association rule mining is widely employed for describing associations within large datasets. On the other hand, it is also possible to extract classifiers from association rules, which consists of sequential events on the left-hand side, and to predefined event on the right-hand side. In our study, we analyze out-of-control points within sensor data and discover sequential events that might signify faults in production. This approach is demonstrated on sensor data obtained from a large-scale wheel rim manufacturer in Turkey. The findings present classifier rules with high lift scores that might help to identify root causes of potential problems and failures in wheel rim molding.</p>