| Summary: | 碩士 === 國立高雄海洋科技大學 === 輪機工程研究所 === 106 === For a long time running machine, the clearance or loosening caused by vibration and noise can affect the efficiency of the machine and life expectancy. This study using the time-frequency analysis method to investigate the instantaneous changes of vibration signals. The variable short-time Fourier transforms which variable window size determined by the peak regional index. Peak index measured in areas of relatively high, giving a larger window application to improve time-frequency resolution, whereas a relatively low peak index measurement giving a small window application. The simulation signals are added with 10dB Gaussian noise and the results shown the time-frequency resolution of the stationary or non-stationary signals with transients can be improved by the proposed algorithm. Then, the mechanical vibration signals are applied using data acquisition card and acceleration sensor. Through the measuring motor vibration signals and time-frequency image processing, the motor state analysis is achieved. From the signal frequency and amplitude strength of size to diagnose the extent of possible damage to the machine. The results shown the time-frequency analysis of vibration signals can provide the criteria to predict motor life.
|
|---|
