Fault detection for rotating machine using time frequency localization method
Bearing is one of the important things in machining. Bearing are considered as critical mechanical components and a defect in a bearing causes malfunction to machine. Failed machines can lead to economic loss and safety problems due to unexpected and sudden production stoppages. These machines need...
| Summary: | Bearing is one of the important things in machining. Bearing are considered as critical mechanical components and a defect in a bearing causes malfunction to machine. Failed machines can lead to economic loss and safety problems due to unexpected and sudden production stoppages. These machines need to be monitored during the production process. Because of that, on-line condition monitoring become alternatives to solve this problem compared with off-line monitoring. Objective of this project is to analyze data acquired from testing of fault detection to differentiate between the defective bearings and good bearings using an accelerometer. A set of good bearing and defective bearing with different failure was using in this experiment. Four units of bearing which is one is good bearing, one corroded bearing, one sandy bearing, and one bearing with damage at the ball was used in this experiment. The data were obtained from experiment on test rig using Bruel & Kjaer accelerometer and data acquisition system. All the bearings were run with different speed which is 4000rpm, 7000rpm, and 10000rpm. The data were analyzed using PULSE LabShop software. The data from three rotations for each bearing was analyzed using time domain, frequency domain, and time-frequency domain analysis. The time-frequency domain method used in this experiment is Short-time Fourier Transform (STFT), and S-transform. STFT and S-transform are applied to detect the location of the signal that has high vibration. The highly damaged bearing is detected based on the high magnitude distribution value in the obtained time-frequency domain. Based on the result, it has a different in vibration between all the bearings. The data for a good bearing were used as benchmark to compare with the defective bearing. For a good bearing, higher vibrations occur at low frequency which is below than 5 kHz using a STFT and below 5μHz when using S-transform. For the defective bearings, the higher vibrations occur at high frequency which is above than 5 kHz when using a STFT analysis and above 5μHz when using S-transform. From the graph, the different between good bearing and defective bearings can be made. The findings indicate that time frequency localization transform method can be used to develop an effective condition monitoring tool. The use of signal processing analysis in this study can be used in industrial applications. This signal processing analysis is recommended to use in on-line monitoring of parameters while the machine is producing. |
|---|