43. PGMON Meeting in Paris (October 6/7, 2011)

TOP 9: Use of advanced techniques for diagnosis and simulation of maintenance (Enrico Zio, University of Milan)

The ultimate goal of Condition Monitoring (CM) is to timely plan appropriate maintenance actions, based on information on the health state of the components. Such information can be of different forms (in-service data, sensor measurements, expert opinions, physical models) and requires different modeling techniques and methods of analysis. The successful resolution of the modeling and methodological issues associated to the task of CM is strongly beneficial for establishing effective condition-based maintenance (CBM) strategies capable of optimally coping with the production availability targets and the safety requirements of Power Plants (PP).

In the presentation, by way of real examples statistical models and Advanced Pattern Recognition (APR) methods have been presented for the establishment of effective CBM of components, including:

• the timely detection of abnormal operation conditions (fault detection);
• the identification of the causes of abnormality (fault identification or diagnosis); 
• the prediction of the remaining useful life in the given abnormal conditions (fault prognosis);
• the informed decision on the proper time and type of maintenance to be performed in order to recover normal operation.

back...