On the Condition Monitoring of Induction machines

摘要: Tuomo Lindh Lappeenranta University of Technology Acta Universitatis Lappeenrantaensis 174 On the Condition Monitoring Induction machines 2003 ISBN 951-764-841-3, ISSN 1456-4491 951-764--843-X (PDF) The research condition monitoring electric motors has been wide for several decades. and development at universities in industry provided means predictive monitoring. Many different devices systems are developed widely used industry, transportation civil engineering. In addition, many methods reported scientific arenas order to improve existing automatic analysis faults. methods, however, not as a part systems. main reasons are, firstly, that presented papers but their performance conditions is evaluated, secondly, include parameters so case specific implementation system using such would be far from straightforward. this thesis, some these evaluated theoretically tested with simulations drive laboratory. A new method bearing fault detection introduced. first work generation originating signal explained its influence into stator current concerned qualitative quantitative estimation. verification feasibility measurement indicator experimentally running 15 kW induction motor. second concentrates on vibration signal. micromachined silicon accelerometer chip conjunction envelope spectrum cyclic tested. Furthermore, creation feature extractors classification researched an classifier multivariate statistical discrimination fuzzy logic It often important on-line integrated industrial communications infrastructure. Two types sensor solutions thesis: one calculation capacity example production spectra; other can collect data memory another device read via field bus. requirements highly depend type solution selected. If already analysed needed only results case, all need transferred. complexity great if management level computer, made itself, analyses must simple due restricted capacity.