Operational modal analysis for damage detection in a rotating wind turbine blade in the presence of measurement noise

Abstract

The present paper demonstrates the feasibility of detecting and quantifying crack-type damage features in rotating wind turbine blades by Operational Modal Analysis (OMA) under noisy measurement conditions. The excitation of blade vibrations is achieved through the rotationally sampled turbulence spectrum of the inflowing wind field. Accelerations are sampled at 30 locations along the blade. Given the high sensitivity of OMA-reconstructed mode shapes to the presence of noise a six-part methodology has been developed. It is demonstrated that small cracks can be detected reliably down to signal-to-noise ratios (SNR) at the blade tip of 40 dB, well below the SNR values achievable with commercial accelerometers and wind-excited acceleration amplitudes. Calibration curves for crack lengths and center positions have been constructed. The proposed methodology holds the potential for the implementation of robust and low-cost structural health monitoring systems for wind turbines at locations with difficult access, such as in an offshore environment.