Intelligent ice detection on wind turbine blades using semantic segmentation and class activation map approaches based on deep learning method

Hacıefendioğlu, Kemal; Başağa, Hasan Basri; Yavuz, Zafer; Karimi, Mohammad Tordi

doi:10.1016/j.renene.2021.10.025

Cited by 42 publications

(24 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…At present, deep learning has continuously achieved better and better results in the field of image classification and recognition [93]. Deep Learning is a field of machine learning that is very close to artificial intelligence, which is motivated by building, simulating the human brain for analytical learning, and neural networks that can extract features of complex objects more intelligently.…”

Section: Based On Deep Learning Methodsmentioning

confidence: 99%

A Review of Methods and Techniques for Detecting Frost on Plant Surfaces

Song

et al. 2021

Agriculture

View full text Add to dashboard Cite

Severe frost usually has adverse impacts on agricultural production, resulting in crop freeze injury, poor crop yield, and crop quality reduction. Timely and accurate detection of frost plays an important role in cold damage warnings, prevention, and control. Current frost detection methods mostly use physical properties such as light, electricity, and heat, or the judge and quantify using environmental factors such as temperature and wind speed. However, it is difficult to detect and accurately identify the frosting phenomenon in real time during field trials because of the complex environment, different plant types, and interference by many factors during observation. To provide an overview of the analytical tools for scientists, researchers, and product developers, a review and comparative analysis of the available literature on frost mechanisms, correlations, and characteristics are presented in this study. First, the mechanisms of the frost formation process, frost level, and the significance of detection, are introduced. Then, the methods and techniques used to measure frost on plant surfaces are synthetically classified and further compared. Moreover, the key points and difficulties are summarized and discussed. Finally, some constructive methods of frost detection are proposed to improve the frost detection process.

show abstract

Section: Based On Deep Learning Methodsmentioning

confidence: 99%

A Review of Methods and Techniques for Detecting Frost on Plant Surfaces

Song

et al. 2021

Agriculture

View full text Add to dashboard Cite

show abstract

“…The central idea of Grad-CAM is to use the gradient information coming into the final convolutional layer of the CNN to identify the significance of each spatial position in the input picture for a particular class; in other words, a given class prediction is used to create a heatmap of the most relevant areas of the image. 173,174 Grad-CAM first computes the gradients of the output class score concerning the activations of the final convolutional layer to generate the heatmap. These gradients represent the extent to which each feature map in the last convolutional layer contributes to the target class prediction.…”

Section: Concept Activation Vectors (Cav)mentioning

confidence: 99%

Potential of Explainable Artificial Intelligence in Advancing Renewable Energy: Challenges and Prospects

Nguyen,

Tarełko,

Sharma

et al. 2024

Energy Fuels

View full text Add to dashboard Cite

Modern machine learning (ML) techniques are making inroads in every aspect of renewable energy for optimization and model prediction. The effective utilization of ML techniques for the development and scaling up of renewable energy systems needs a high degree of accountability. However, most of the ML approaches currently in use are termed black box since their work is difficult to comprehend. Explainable artificial intelligence (XAI) is an attractive option to solve the issue of poor interoperability in black-box methods. This review investigates the relationship between renewable energy (RE) and XAI. It emphasizes the potential advantages of XAI in improving the performance and efficacy of RE systems. It is realized that although the integration of XAI with RE has enormous potential to alter how energy is produced and consumed, possible hazards and barriers remain to be overcome, particularly concerning transparency, accountability, and fairness. Thus, extensive research is required to address the societal and ethical implications of using XAI in RE and to create standardized data sets and evaluation metrics. In summary, this paper shows the potential, perspectives, opportunities, and challenges of XAI application to RE system management and operation aiming to target the efficient energy-use goals for a more sustainable and trustworthy future.

show abstract

“…Previous studies (Zhang et al, 2018;Liu et al, 2019;Yuan et al, 2019;Jiang and Jin, 2021;Kreutz et al, 2021;Kemal et al, 2022) have utilized image data from automated optical cameras to train machine learning models (e.g., decision trees) as well as leveraged DL techniques for ice detection. They have all demonstrated near-perfect accuracy with regards to the degree of detecting ice, including near and far views of a blade, ice density, and light (Alvela Nieto et al, 2023).…”

Section: Direct Approachesmentioning

confidence: 99%

Domain-invariant icing detection on wind turbine rotor blades with generative artificial intelligence for deep transfer learning

Chatterjee

Nieto

Gelbhardt

et al. 2023

Environ. Data Science

View full text Add to dashboard Cite

Wind energy’s ability to liberate the world from conventional sources of energy relies on lowering the significant costs associated with the maintenance of wind turbines. Since icing events on turbine rotor blades are a leading cause of operational failures, identifying icing in advance is critical. Some recent studies have utilized deep learning (DL) techniques to predict icing events with high accuracy by leveraging rotor blade images, but these studies only focus on specific wind parks and fail to generalize to unseen scenarios (e.g., new rotor blade designs). In this paper, we aim to facilitate ice prediction on the face of lack of ice images in new wind parks. We propose the utilization of synthetic data augmentation via a generative artificial intelligence technique—the neural style transfer algorithm to improve the generalization of existing ice prediction models. We also compare the proposed technique with the CycleGAN as a baseline. We show that training standalone DL models with augmented data that captures domain-invariant icing characteristics can help improve predictive performance across multiple wind parks. Through efficient identification of icing, this study can support preventive maintenance of wind energy sources by making them more reliable toward tackling climate change.

show abstract

Intelligent ice detection on wind turbine blades using semantic segmentation and class activation map approaches based on deep learning method

Cited by 42 publications

References 28 publications

A Review of Methods and Techniques for Detecting Frost on Plant Surfaces

A Review of Methods and Techniques for Detecting Frost on Plant Surfaces

Potential of Explainable Artificial Intelligence in Advancing Renewable Energy: Challenges and Prospects

Domain-invariant icing detection on wind turbine rotor blades with generative artificial intelligence for deep transfer learning

Contact Info

Product

Resources

About