Machine Learning Analysis of Non-Destructive Evaluation Data from Radar Inspection of Wind Turbine Blades

Wind Turbines are vital contributors to powering the world with renewable energy. As the wind energy sector grows, the reliability and resilience of wind turbine systems becomes increasingly important. One of the most important components, the wind turbine blades, are typically inspected with visual analysis. This is insufficient for providing detailed, consistent, and readily accessible surface and sub-surface analysis for Structural Health Monitoring (SHM). In this paper we present a novel method of Non-Destructive Evaluation (NDE) for wind turbine blades by utilizing Frequency Modulated Continuous Wave (FMCW) radar sensing with machine learning analytics. By utilizing machine learning models on FMCW radar return signal amplitude (RSA) collected from different turbine blade samples, our results demonstrate that we can classify blade types by composition, and diameter differentials of 3 millimeters with over 95% classification accuracy. Thus, our methodology presents an insight to a promising SHM-NDE solution for surface and subsurface characterization of wind turbine blades and other composite structures.