Vibration analysis of motors using Gabor expansion

• Main Author: Ronquillo, Tirso A.
• Format: text
• Language: English
• Published: Animo Repository 2003
• Subjects: Vibration > Measurements; Machinery > Vibration; Algorithms
• Online Access: https://animorepository.dlsu.edu.ph/etd_masteral/3042
• Institution: De La Salle University

Gabor expansion is applied to determine the severity of vibration of motors. An algorithm was developed to make classification of vibration severity possible. Codes were written in Matlab and Visual C++ to obtain the Gabor spectogram of different vibration signatures and to determine the energy of each. Data from National Power Corporation were used and analyzed to determine the severity of machine vibration using the signal energy as parameter. Results showed that as the energy of vibration signal increases, machine vibration severity also increases. The effect of varying the size of Gaussian window in the Gabor spectrogram was examined by simulating the vibration data of different values of Gauss variance. Within certain limits, varying the size of the Gaussian window did not have any marked effect on the value of energy. The effect of the sample size used to represent vibration signature for a particular time of measurement was examined and results showed that decreasing the sample size just results in the corresponding decrease in energy value and does not affect vibration severity classification. From the vibration data, severity mapping table was established based on the quadratic polynomial model determined from the energy and vibration severity relation. The effect of noise introduction into the vibration data was considered and results revealed that vibration energy is generally increased for an added random noise. However this change in energy did not shift the corresponding vibration severity. Specifications of hardware components for data acquisition system are recommended for accurate vibration data measurements.