EMBEDDED SENSITIVITY FUNCTIONS FOR IMPROVING THE EFFECTIVENESS OF VIBRO-ACOUSTIC MODULATION AND DAMAGE DETECTION ON WIND TURBINE BLADES

In structural health monitoring, it is desirable to select sensor locations in order to minimize the number of sensors required for and the cost associated with an on-board monitoring system. When using a frequency response-based structural health monitoring technique, data measured at sensor locations which exhibit the greatest change in frequency response function (FRF) due to damage are expected to maximize the effectiveness of the chosen technique. In this work, an embedded sensitivity function is presented which identifies the sensor locations at which the maximum differences in FRFs due to damage at a known location will be observed. The formulation of the embedded sensitivity function is based on FRFs measured on a healthy structure in the frequency range in which the damage detection technique will be applied. The effectiveness of the embedded sensitivity functions in predicting the most effective sensor locations is demonstrated by applying a vibro-acoustic modulation (VAM) damage detection method to a residential-scale wind turbine blade. First, data from the healthy blade is acquired and the embedded sensitivity functions are calculated. Then, the blade is damaged and the VAM method is applied using several sensor locations. The data acquired using sensor locations identified by the embedded sensitivity functions as being most effective are shown to most clearly identify the damage on the blade.