| Summary: | 碩士 === 國立中山大學 === 電機工程研究所 === 81 ===
The maintenance of electromechanical drive systems is essential to the efficient, and profitable, operation of a wide variety of industrial processes. A lack of a coherent maintenance strategy can result in the loss of individual items of plant, a reduction in operational safety, and a heavy burden of capitalized losses. There is a growing realization that condition monitoring of plant can substantially reduce direct costs, and also help to alleviate the problem associated with the shortage of suitably skilled personnel. By condition monitoring, we mean the continuous assessment of the performance and health of plant throughout its useful operating life.
Maintenance can be categorized as one of three types, breakdown, preventative and predictive. Breakdown maintenance is self explanatory. The machines damaged by power accidents may result in extremely expensive cost for breakdown maintenance. Preventative maintenance is scheduled out irrespective of machine condition. Intervals should be scheduled to prevent catastrophic failures but still achieving reasonable component usage. Predictive maintenance is the indication of maintenance actions and planned by an assessment of machine condition. Saving of up to 40% of maintenance budgets have been achieved by several large corporation. These saving can be made by carefully integrating planned and condition based maintenance and effectively integrating condition monitoring into plant maintenance activity. Machine functions require to be assessed critically to determine the relative indication of the importance of any one machine. If a points system is used to perform this task, each assessment heading can be subdivided into low risk through to high risk and subsequently the points aggregated. The assessment is then considered against financial savings, the complexity of suitable condition monitoring techniques and finally the practicality of monitoring the required parameters.
Recently, the techniques and computer softwares have been developed to predict and monitor electrical machine behavior for a range of disturbance conditions. However, some of these ways can only monitor some kinds of machine faults and are not practical in industrial applications. The more practical methods developed are able to monitor the behavior of a induction motor in two ways: First, the monitoring is of use for line current and secondly, for the axial leakage flux.
This thesis is principally concerned on the monitoring of axial leakage flux, which may avoid to involving the effects of voltage and current so as to omit the consideration of safety and isolation factor. In addition, the vibration measurement of the induction motor is also implemented in a practical experiment of 50 HP induction motor in thesis. The results obtained is presented in power spectrum density method so that various different types of faults cna be compared and recognized which fault it is.
Installation of the equipments, study of the computer software and practical experimental tests is finished respectively. In this thesis various faults which may occur in the induction motor are simulated below:
no load
locked rotor test
loss of phase
unbalance supply
unbalance rotor
misalignment
mechanical relaxation
The results presented are taken a comparison and analysis for the characteristic of these faults and provide engineer with a basis and reference to predict maintenance for electrical machine in industrial application.
|
|---|
