振動法によるボールベアリング油切れ異常の監視システムの警報の信頼性 : カルマンフィルタと指数形加重移動平均法の比較

概要

論文の詳細を見る
This paper examines alarm reliability based on the vibration monitoring of rotary machinery. Two procedures for diagnosing bearing condition are compared experimentally. One is the Kalman filter combined with Bayes' theorem and the other is the exponentially weighted moving average method. In the experiment, the rotor is operated at a constant speed and the vibration level is monitored. Increased vibration due to bearing seizure is detected by the methods mentioned above. The results show that 1) the moving average method can detect relatively faster than the method using the kalman filter but sometimes leads to a false alarm, and 2) in monitoring using the Kalman filter, no false alarm is observed, which means that higher alarm reliability can be obtained by the Kalman filter.
一般社団法人日本機械学会の論文
1993-04-25