Michael Pecht1 , Michael H. Azarian1, Ranjith S.R. Kumar1, Nishad Patil1, Anshul Shrivastava1
1CALCE, Center for Advanced Life Cycle Engineering, Department of Mechanical Engineering, University of Maryland, College Park, Maryland 20740, USA
With rising oil prices, the depletion of fossil fuel sources and changing geo‐political environment, there is
heavy investment worldwide in the development of alternate sources of energy. Wind provides a widely
available, non‐polluting, renewable source of energy which can help to meet the ever growing demand.
The success of wind energy as an alternative to fossil fuels hinges on the extent to which wind turbines
can provide a dependable and cost‐effective source of power.
Statistics show that the longest downtime for wind turbines is due to failures of the gearbox, followed by
the electrical system and the control system. Wear‐out failure of bearings caused the longest average
downtime among the gearbox failures, most of which demanded a complete change of the gearbox or
bearings. The circuitry used for conversion and transfer of power in wind turbines consists of
components such as power semiconductor devices (e.g., IGBTs, MOSFETs, diodes), as well as
transformers, inductors, and capacitors. Insulated gate bipolar transistors (IGBTs) are known to exhibit
latch‐up or parametric degradation due to aging. Liquid aluminum electrolytic capacitors have been
responsible for numerous costly and potentially hazardous equipment failures.
The availability of wind energy generation equipment can be improved by implementing techniques for
health monitoring, anomaly detection, and prognostics. Successful identification of degraded components
is critical to preventing deterioration of power quality or a potentially catastrophic short circuit. This
paper provides an overview of PHM (prognostics and health management) strategies for critical
components of wind energy systems. Specific examples from health monitoring studies of IGBTs,
electrolytic capacitors, and bearings provide an illustration of how these techniques can lead to real‐time
fault detection and life time estimation in wind turbines, enabling condition‐based maintenance and
reduction in life‐cycle costs.