Machine Learning in the context of structural health monitoring becomes more and more important. A paper has been accepted for publication in Sensors journal describing radar-based damage detection under fatigue loading using convolutional neural networks (CNNs).
Abstract:
This paper reports on a convolutional neural network (CNN) based damage detection approach for radar-based structural health monitoring of wind turbine blades. Subsequent radar measurements are transformed into an image-type representation being used as CNN-input. In contrast to conventional approaches that require a compensation of temperature and loading effects, the proposed framework inherently learns all the required information during a training phase. The damage detection performance, i.e. intact vs. damaged condition, is demonstrated using measurements from multiple embedded radar sensors during a fatigue test of a 31m long wind turbine blade. The achieved F1-score for correct damage classification is between 91% and 100% for the unloaded as well as the loaded blade.
Reference:
Streser, E. ; Alipek, S. ; Rao M ; Simon, J. ; Moll, J. ; Kraemer, P. and Krozer, V., Radar-Based Damage Detection in a Wind Turbine Blade using Convolutional Neural Networks: A Proof-of-Concept under Fatigue Loading, Sensors, 2025 (accepted in May 2025)
