Speech content is below
 

Tutorial M.A.1:  <em>Designing health diagnosis and prognosis based on vibration signals</em> 

Vibration signals are of critical importance for diagnosing and predicting the health condition of mechanical systems. A great deal of researches have been conducted on developing diagnostic and prognostic methods implemented in various engineering fields like energy, transportation, defense, etc. This talk presents how to design a formal procedure for health diagnosis and prognosis based on vibration signals, including data acquisition system design, vibration signal preprocessing (noise reduction, filtering, etc.) and processing (feature extraction, health classification, supervised and semi-supervised learning, statistical inference, etc.), and health prognosis. For the demonstration purpose, this tutorial presents two engineering case studies: (a) health classification for steam turbine rotor systems with multiple failure modes and (b) health diagnosis and prognosis for wind turbine drivetrain systems. A sample code for diagnosing the health condition of an RK4 rotor will be provided to the attendee.