Pattern Recognition Analysis for Ball Bearing Diagnosis

スポンサーリンク

概要

• 論文の詳細を見る

• This paper describes the fault detection of ball bearings based on sound information by using the pattern recognition technique. In experiments, normal bearings and bearings with three types of faults are used : a fault on the outer race, the inner race, or the ball. First, the sound generated by each bearing is measured with a microphone and the sound spectrum is obtained with an FFT analyzer. Second, the reduced and reference patterns for each fault are calculated by using the sound spectrum. Third, the diagnosing patterns are compared with the reference pattern of each fault to detect the fault of the bearing and to determine the type of the fault. We also apply the statistical method to the pattern recognition. The effectiveness of using sound information and applying the pattern recognition technique in diagnosing ball bearings is discussed in the paper.

• 一般社団法人日本機械学会の論文

著者

• Nakagawa Noritoshi

  Cluster I Faculty Of Engineering Hiroshima University

• Sekiguchi Yasuhisa

  Cluster I Faculty Of Engineering Hiroshima University

関連論文

• Identification of Sound Sources based on Sound Pressure Measurement Considering Observation Point Importance
• Pattern Recognition Analysis for Ball Bearing Diagnosis

スポンサーリンク