| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 107 === In today's industry, high-voltage motors are indispensable sources of power. There are some characteristics about high-voltage motors, like long life cycle, high energy efficiency, low vibration noise and high stability. High-voltage motors usually need long-term operation to maintain economic efficiency. Therefore, how to maintain high-voltage motors is an important issue. Most of today's factories or electric power plants adopt a maintenance strategy for predetermined maintenance, also known as time-based maintenance (TBM). Although the probability of failure can be reduced, the potential operation status of the high-voltage motor cannot be displayed immediately. If the operation status of the high-voltage motor can be predicted early and prevented in advance, the maintenance cost can be greatly reduced, and major accidents can be avoided.
This thesis is dedicated to the establishment of a proactive high-voltage motor operation condition monitoring method based on electric signals. Firstly, the three-line voltage and current signal of the high-voltage motor running in an electric power plant are captured by the measuring platform, and the one-day data of normal operation is taken from the database. Because there is no dangerous operation data of the high-voltage motor, this study adds additive white Gaussian noise (AWGN) and linear amplification on normal state data, synthesizing warning and dangerous state data, and makes a case study. Next, calculate the relevant electrical indexes in the international standard, and then extract the least number of characteristic indexes with the most structure information through the principal component analysis (PCA). Further, we use the extracted characteristic indexes dataset as the inputs, and employ the fuzzy C-means (FCM) clustering method to cluster the data, that is, distinguish the various operation states of the motor. Finally, the data is defuzzified, and the data points are displayed in percentage for the user to refer to the high-voltage motor operating state to make the most suitable maintenance decision.
|
|---|
