Vibration-based ice monitoring of composite blades using artificial neural networks under different icing conditions

Abstract

Cold climates pose significant challenges for wind turbines, primarily due to icing complications that influence electrical energy production. Precise methods are needed to identify and predict ice distribution on blades. Thus, enhancing prediction of ice accumulation based on the blade’s frequency response. The study involves using glass fiber reinforced plastic composite rotor blades equipped with actuators and accelerometers to measure the response of the blade subjected to icing, with a total of 1700 measurements. Small-scale icing experiments are conducted inside a climate chamber at temperatures from −10 ◦C to −20 ◦C with seven icing distribution profiles on the blades. The gathered data are analyzed for the effects of icing on the frequency response of the blades. Additionally, we propose the use of optimized artificial neural networks to predict the accumulated ice thickness on rotor blades with a weighted mean absolute percentage error of 5.1 %, and ice volume and ice mass with an error of 5.7 %, based on the frequency response. Overall, this paper investigates the relation between icing, with regard to ice mass, ice location, and ambient temperature, and frequency response of wind turbine blades, along with proposing a high-performance method for ice detection and monitoring during operation.