Automated Quantification of Wind Turbine Blade Leading Edge Erosion from Field Images

Abstract

Wind turbine blade leading edge erosion is a major source of power production loss and early detection benefits optimization of repair strategies. Two machine learning (ML) models are developed and evaluated for automated quantification of the areal extent, morphology and nature (deep, shallow) of damage from field images. The supervised ML model employs convolutional neural networks (CNN) and learns features (specific types of damage) present in an annotated set of training images. The unsupervised approach aggregates pixel intensity thresholding with calculation of pixel-by-pixel shadow ratio (PTS) to independently identify features within images. The models are developed and tested using a dataset of 140 field images. The images sample across a range of blade orientation, aspect ratio, lighting and resolution. Each model (CNN v PTS) is applied to quantify the percent area of the visible blade that is damaged and classifies the damage into deep or shallow using only the images as input. Both models successfully identify approximately 65% of total damage area in the independent images, and both perform better at quantifying deep damage. The CNN is more successful at identifying shallow damage and exhibits better performance when applied to the images after they are preprocessed to a common blade orientation.

Use of Artificial Intelligence in Design, Development, Additive Manufacturing, and Certification of Multifunctional Composites for Aircraft, Drones, and Spacecraft

“…The reported quasi-static multifunctional properties and the failure modes and locations, coupled with machine learning (ML), could help inform future iterations of the design, development, and advanced manufacturing with the hope of offering significant weight reductions and potentially replacing the bulky electrical wires in aerospace applications [13]. A novel automated quantification methodology was developed to evaluate the health and performance of the wind turbine blade leading edge erosion using the data from the 140 field images covering the varying blade orientation, resolution, aspect ratio, and lighting [14]. Supervised ML and unsupervised ML models were created for evaluating the automated quantification of the damage states obtained from the field images of the wind turbine leading edges.…”

“…Supervised ML and unsupervised ML models were created for evaluating the automated quantification of the damage states obtained from the field images of the wind turbine leading edges. While the supervised ML model applied convolutional neural networks (CNN) and learnt damage details/types as features that are found in the typical datasets of training of the 140 field images, the unsupervised ML model conducted a consolidation of pixel intensity thresholding with the calculation of a pixel-by-pixel shadow ratio to independently identify features within the 140 field images of the wind turbine leading edges [14]. It was found that the CNN model was better in the identification of shallow damage and yielded higher performance when used with the 140 field images after their preprocessing to common blade orientation [14].…”

“…While the supervised ML model applied convolutional neural networks (CNN) and learnt damage details/types as features that are found in the typical datasets of training of the 140 field images, the unsupervised ML model conducted a consolidation of pixel intensity thresholding with the calculation of a pixel-by-pixel shadow ratio to independently identify features within the 140 field images of the wind turbine leading edges [14]. It was found that the CNN model was better in the identification of shallow damage and yielded higher performance when used with the 140 field images after their preprocessing to common blade orientation [14]. While there have been various damage classification techniques using the supervised ML models for self-organizing maps in the structural health monitoring (SHM) systems, this study did not address the use of ML in the design, development, manufacturing, testing, and certification of the aircraft, UAS, and spacecraft using the digital data from across different entities such as accidents/incidents databases, service difficulty reporting, airlines, maintenance, repair and alteration, design-manufacturing-testing data, etc.…”