振動波形の等価実効値によるベアリングの簡易診断(機械力学,計測,自動制御)

概要

論文の詳細を見る
We have diagnosed the failures of bearings by using parameters such as vibration amplitude, crest factor and kurtosis. However, there are some problems in this method. The alarm levels are necessary for diagnosis, and we cannot apply this method to high speed rotations. Then, the highly accurate diagnosis technology which has generality is requested. This paper proposes the method of solving these problems. First, we investigated the possibility of diagnoses of bearings by kurtosis on the assumption that the amplitude probability density of the vibration wave form caused by a normal bearing is normal distribution. Next, it is described that the 68.3 percentile value obtained from the cumulative frequency distribution of the envelope processing of the vibration wave form is equivalent to the RMS value when it is normal condition. In addition, it is described that the ratio of the equivalent RMS value and an actual RMS value and the fourth power mean value normalized by the equivalent RMS value are both high accuracy and high sensitivity compared with conventional parameters, and they are useful for diagnosis of bearings. Finally, a palm size bearing fault detector in which this diagnosis algorithm is installed is described.
社団法人日本機械学会の論文
2001-12-25