BEARING FAULT DETECTION AND ITS LUBRICATION CONDITION MONITROING USING JERK MEASUREMENTS

• Main Author: HERYADI (NIM : 13106022); Pembimbing : Dr. Ir. Zainal Abidin, BUDI
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/15192
• Institution: Institut Teknologi Bandung

Roller bearing has an imporant role in industry, its failure makes the machine totally stopped so that the production process has to be terminated. Vibration method is well developed to detect bearing fault, but it can not detect early symptoms of the fault. Nowadays, PeakVue, Spike Energy, and Shock Pulse method are widely used to fullfill this demand, unfortunately these methods have no standard as the vibration method has. <br /> <br /> <br /> This research is done in order to explore the possibility of applying jerk and its frequency spectrum to detect bearing fault as well as to monitor its lubrication condition. Hence, jerk – the first derivative of acceleration – is introduced to replace vibration, PeakVue, Spike Energy, and Shock Pulse method on detecting bearing fault. <br /> <br /> <br /> An experiment subjected to needle bearings is carried out to meassure acceleration and jerk with several levels of load, types of fault, and lubrication conditions. In addition, another experiment subjected on tappered bearings is accomplished to measure acceleration, jerk, and shock pulse with several levels of speed, load, and bearing fault severity. <br /> <br /> <br /> The result shows that jerk can be used to detect bearing fault and to monitor its lubrication condition. Jerk value is more sensitive with speed, lubrication condition, and bearing fault severity than acceleration and shock pulse value. But, shock pulse spectrum is better than rectified acceleration and jerk spectrum on detecting bearing fault. Noise level in acceleration&#8223;s spectrum and jerk&#8223;s spectrum are high enough so that the BPFO signal and its harmonics doesn&#8223;t seem clearly.