Condition monitoring needs smart technologies to diagnose faults and prognose failures, which this paper will investigate. The paper starts with the description of the fundamentals of system identification and neural network approaches. The proposed models will be tested and validated with measurement data. Issues related specially to continuous online monitoring will be addressed for these models. The work demonstrates that the proposed techniques can be potentially applied to online condition monitoring and therefore health assessment of power plant generators.
Conference name: The 24th International Congress on Condition Monitoring and Diagnostics Engineering Management (COMADEM2011), Stavanger, Norway, 30th May - 1st June 2011