振動のオクターブバンド分析を用いた駆動用電動機軸受の異常検知

抄録

We are developing condition monitoring methods aiming at preventing malfunctions of traction motor bearings during operation and extending the maintenance cycle. In order to verify an abnormality detection method combining vibration octave band analysis and machine learning, a rotation test was carried out by incorporating an artificial defect bearing in a traction motor. As a result, it was shown that abnormality was detected in high frequency vibration of 1 kHz or more. It was also shown that an abnormality of the output side bearing was detected by a vibration sensor attached to the counter output side.