| Summary: | 博士 === 國立臺灣科技大學 === 電機工程系 === 104 === Thanks to the rapid development of science and technology, artificial intelligence has been widely used in various industries, the fourth industrial revolution has quietly come, the rotating electrical machine is the mother of industry, plays a pivotal role in the power grid and plants, and predict maintenance strategy is now of concern important operation and maintenance issues, therefore, put forward a fuzzy theory-based fault diagnosis and condition monitoring systems for rotating electrical machine, which will help to enhance the reliability of the unit stable operation.
In this study, a total of ten rotating electrical machine for an experimental model to make electrical, vibration and partial discharge signal analysis, and the use of data mining technology in the mining potential of the signal fault symptoms, according to different considerations of cost and safety factors are proposed five kinds of fault diagnosis and three states monitoring system solutions. The experimental model test results, 52 % of the electrical method with the highest probability of accurately infer broken rotor bar fault, vibration method to 100 % of the highest probability of accurately infer the bearing outer ring damage and eccentric failure, partial discharge method with the highest probability of 100 % accurately infer the stator insulation failure. However, hybrid electric and vibration method can effectively improve the broken rotor bars inference probability to 83 %, and then integrate the partial discharge method, can be more accurately identification of stator insulation type of exception, and effectively avoid the possibility of misjudged, electrical, vibration and partial discharge of condition monitoring systems, taking into account relevant monitoring programs, such as the gap spacing and failure factors, and strengthening existing international standards-based condition monitoring guidelines, but in the future remains to be done to monitor the long-term operation of the rotating electrical machine, according to data analysis adjustment inference rule and the weight values to expect substantive enhance monitoring performance and operation reliability.
Concluding our proposed rotating electrical machine operation and maintenance strategy confirmed by experiments with the feasibility and effectiveness. We look forward to a positive helpful in the development process of the domestic industry, and further to avoid rotating motor abnormalities lead to serious negative impact on casualties and economic losses.
|
|---|
