Experimental Investigation of Effect of a Transverse Crack on Wind Turbine Blade Using Vibration-Based Method for the Development of SHM

This study attempts to analyze the health condition of a wind turbine blade under the influence of a transverse crack using a vibration-based technique. A commercially available wind turbine is used to study the effect of transverse cracks on the blade using a fast Fourier transform (FFT) analyzer. Initially, frequency spectrum of a wind turbine with an intact blade is recorded as an indication of healthy condition. A transverse crack has been induced on one of the blades to record the frequency spectrum of the wind turbine, considering the occurrence of damage. The frequency spectrum is recorded for various locations of cracks at different depths. It is observed that, an imbalance in the system is generated due to the presence of cracks. Also, as the crack approached the free end of the blade, the frequency decreased; indicating a reduction in stiffness. The introduction of cracks reduced frequency considerably. However, for an increase in crack depth, hardly any reduction in frequency is observed. A Python program is constructed to predict the location of a crack using a database prepared from experimentation. Various machine learning algorithms have been used to form the basis for the structural health monitoring (SHM) system. K nearest neighbour (KNN) algorithm showed promising results out of all the algorithms used. It can be concluded that the constructed program can identify the occurrence of cracks at a particular location.