Embedded Structural Health Monitoring
and Predictive Modeling of Wind Turbine Blades

  • Figure 1

  • Figure 2

  • Figure 3


Click here to return to the projects summary page.


Sponsors / Collaborators: Los Alamos National Laboratories

Recent developments in structural health monitoring (SHM) have provided the tools and algorithms needed to identify damage in airfoil-based structures, such as aircraft wings and wind turbine blades. These tools include active sensing methods using piezoelectric actuator/sensors, and the algorithms include a host of pattern recognition techniques rooted in detection theory and signal processing. We are implementing certain of these algorithms on electronic hardware that is capable of acquiring and processing relevant data from a wind turbine blade to determine the presence and progression of damage in the blade. That information will further be incorporated into a model (Figure 2 - Non linear displacement field) of the blade that will predict blade performance in terms of load transmission and ultimate power transfer in order to make informed decisions concerning control of the blade and the overall wind turbine maintenance schedule. Models capable of directly incorporating damage information (e.g. crack propagation) are far too complex to run in real time, requiring millions of degrees of freedom. Simple beam models are inadequate for rotating wind turbine blades, which undergo significant geometric nonlinearities. We are collaborating with researchers at Los Alamos National Laboratory to apply a nonlinear beam code being developed there to predict the global structural state given an indication of damage. That prediction will be refined and validated in real time using kinematic data collected by the same sensor node that conducts the active sensing measurements for SHM.

To that end, we are conducting series of experiments to monitor both the failure and the normal operation of wind turbine blades using our prototype embedded system, the Wireless Active Sensing Platform (WASP) (Figure 3). The first experiment is the fatigue testing to failure of a 9-meter CX-100 wind turbine blade (Figure 1). This test will provide valuable information from which to develop damage models for the blade, and with which to validate the model’s ability to predict the effects of damage caused by fatiguing the blade. The second experiment will consist of a series of flight tests of two Whisper 500 wind turbines. The Whisper 500 has a span of 4.5 meters with blades 2.1 meters long, and it is primarily intended for residential installations. This experiment will provide data to validate the geometrically nonlinear beam models used to model the larger blades. Multiple turbines will enable us to observe the effect of one turbine’s wake on the other turbine’s wind loading. The third experiment will be a flight test of the CX-100 blade with the sensor node embedded in the blade. This experiment will demonstrate the capability of the embedded system to monitor and predict blade performance in real time.