Damage Identification of HAWT Blade using Ordinary Linear Kriging Method and Variation of Blade’s Modal Parameters

Abstract

Wind turbine blades operate in a harsh environment causing them to always be susceptible to damage. Variable wind loading, debris impact, and thermal gradient, among other factors, can cause damage to the blades. Detection of blade damage at early stages can prevent massive cost associated with turbine down-time and blade replacement. In this work, a vibration-based method is presented to detect damage at early stages. The presented method takes advantage of the effect of crack on modal parameters of the blades vibration. Finite element model (FEA) is constructed for both healthy and damage blade to study that effect. Power spectral density (PSD) plots of the blade’s vibration before and after damage are compared and the changes in the resonant modal amplitudes frequencies are identified. To minimize the number accelerometers needed to monitor the health of the blade and without compromising the accuracy of damage predictions, ordinary kriging method is used to predict cracks in the blade’s structure. Kriging uses modal parameter data, experimental or otherwise, to estimate damage location on the blade. It creates a map of damage predictions throughout the region use measurements from far less sensors than common techniques. Damage characteristics estimates using the proposed method showed damage attributes predictions with accuracy greater than 93 %. Simulation is used to validate the proposed method and the results are discussed.