Pengembangan Metode Deteksi Cacat Bantalan Berbasis Analisis Envelope pada Prototipe Fan Industri

Abstract

 This paper examines the spectrum and envelope spectrum for detection of ball bearing fault using vibration signals taken from a prototype of industrial fan. It presents a procedure to detect fault on the outer and inner race of roller bearing. Three conditions of ball bearing i.e., normal, outer race fault and inner race fault is tested. The artificial fault on the bearing is 2 mm depth and 0,7 mm width. The vibration signals are taken by using an accelerometer mounted on the vertical-radial direction of bearing pillow block. The waveforms are then transformed into spectrum and envelope spectrum. The differences between spectrum and envelope spectrum for both types of faults are analyzed and discussed. The paper explains the effect of amplitude modulation (AM) on the inner race fault to the spectrum and envelope spectrum. The discussion about the ability of spectrum and envelope spectrum for bearing fault detection is presented and compared in detail. The results show that spectrum has good ability for fault detection since it gives clear and high amplitude of bearing fault frequencies. However the spectrum often fails when the size of fault is relatively small, in this case the envelope spectrum gives better results.