Ice Detection on Rotor Blades of Wind Turbines using RGB Images and Convolutional Neural Networks
Markus Kreutz, Abderrahim Ait Alla, Anatoli Eisenstadt, Michael Freitag, Klaus‐Dieter Thoben
Abstract
Ice formed on wind turbines not only causes economic damage but also poses a risk to nearby humans. Countermeasuring ice formation requires reliable detection systems. Existing ice detection solutions have improved with time, but are still not sufficiently accurate. This paper investigates the use of different convolutional neural networks (CNNs) and RGB images taken from cameras installed on the nacelle of wind turbines to detect ice on the rotor blades. In our research, the VGG19 model achieved the best performance with an accuracy and an f1-score of (96 ± 2) %.
Topics & Concepts
Convolutional neural networkNacelleWind powerRGB color modelRotor (electric)Computer scienceArtificial intelligenceTurbineMarine engineeringRemote sensingGeologyEngineeringAerospace engineeringMechanical engineeringElectrical engineeringIcing and De-icing TechnologiesSmart Materials for ConstructionFire Detection and Safety Systems