FRF PATTERN OF FLOOR NOISE GENERATION IN POWER SPECTRUM OF ROLLING BEARING VIBRATION RESPONSE

• Main Author: MONHENDRI (NIM : 231 09 032); Pembimbing : Prof. Dr. Ir. Komang Bagiasna, RIKO
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/17012
• Institution: Institut Teknologi Bandung

Predictive maintenance based on vibration signal monitoring is a reliable method of maintenance for rotating machinery to assure its proper operation. There are several measurements used to determine the condition of rotating machinery, among other are FRF measurement and power spectrum measurement. Most of the vibration respon power spectrum shows that the generated floor noise has pattern which similar to the pattern of FRF. This research is intended to identify the relationship between the bearing surface roughness and the generated floor noise vibration signal. The analysis was carried out using numerical method, i.e. finite element method. The following research activities, i.e.: FRF, power spectrum, as well as surface roughness measurements were conducted to accomplish this research. The FRF measurement is evaluated in order to know dynamic characteristic of the tested bearing mounted inside bearing house. On the other hand, power spectrum evaluation is performed to understand the occurance of FRF pattern in the floor noise. The surface roughness was measured using Surftest Mitutoyo SJ 301. The results of bearing surface roughness measurement were utilized to reconstruct the Hertzian contact force by utilizing the finite element model of the bearing unit. This reasearch shows that floor noise of rolling bearing power spectrum measurement has a similar pattern compared to the FRF of the bearing unit up to 96,4% similarity level. Simulation of power spectrum shows that the Hertzian contact force originated from the bearing surface roughness can generate random vibration along the measured span frequency. The form of generated floor noise generated the floor noise measurement with 91,3% similarity level.