A Deep Learning Approach for Structural Health Monitoring of Wind Turbine Using CNN Algorithm

Renewable wind energy is generated by constructing wind turbines, which aim to provide economical, reliable, and environmentally friendly energy sources. The blade of a wind turbine is really important. If it gets damaged, it can cause a big breakdown. This damage is a major reason for the turbine to fail. Hence, detecting and addressing any damage to the blade as early as possible is advantageous. This research aims to create a deep learning technique that relies on data analysis to detect and monitor any damage in wind turbine blades using vibrational analysis. This study involved the use of ANSYS 2022 R2 software to generate a NACA 63-412 profile, and its accuracy was confirmed by comparing it with finding from a previous study. According to numerical simulations, the proposed method can identify two types of damage in wind turbine blades: (i) crack in the blade and (ii) loosening of bolts in the turbine. Acceleration data was gathered from various locations on the structure, while it was subjected to impact loading, and this data was analyzed in the frequency domain using the Fast Fourier Transform (FFT) technique, both in its healthy and unhealthy conditions. After the acceleration time-histories data were collected, they were transformed into scalogram images. These images were then used for the Convolutional Neural Network algorithm to classify and identify any damage present. With the deep learning technique, it is possible to accurately differentiate between healthy and various types of damage conditions with high accuracy. This demonstrates its effectiveness as an automation tool for Wind Turbine Structural Health Monitoring (SHM).