Classification of Wind Turbine Defects based on the SqueezeNet Neural Network

The article presents a study on classifying wind turbine defects using the SqueezeNet neural network. Wind turbines are critical for renewable energy, but defects such as corrosion, erosion, and cracks can significantly reduce their efficiency. The study proposes using image classification techniques with neural networks, particularly SqueezeNet, to automate defect detection. The compact nature of SqueezeNet makes it suitable for real-time applications with limited computing resources. Through training and testing, the network achieved an accuracy of 89%, demonstrating its potential to improve wind turbine maintenance and reduce operational costs.