PEMANTAUAN KONDISI MESIN DENGAN EKSTRAKSI FITUR SINYAL GETARAN

Abstract

Machinery condition monitoring is increasingly important in industry due to the need to improve reliability and reduce the possibility of production loss due to machinery breakdown. Vibration monitoring is the most widely used method for machinery condition monitoring and is considered the most effective method to detect defect in rotating machinery. Therefore, accurate features of vibration signals are required in fault diagnosis of machinery. In this study, vibration measurement data were extracted using several statistical parameters of time domain. Vibration signal data from a variety condition of rotating machinery were obtained from machinery faults simulator. In the simulation the machinery conditions were normal, unbalance, loosenes, misalignment, and bearing defect. The result of feature extraction of vibration signal is used as input to the self organizing map (SOM). In this study, SOM is used as tool for fault diagnosis machinery. The results shows that the RMS feature has best performance and mean has poorest performance.