This research presents an integrated intelligent framework for the early detection of wind turbine failures by combining physical modeling with artificial intelligence techniques, with the aim of improving system reliability and reducing maintenance costs. The significance of this work lies in addressing the challenges associated with a lack of real-world data and the limited ability of traditional models to generalize in complex operating environments. To achieve this, a dynamic model was developed using MATLAB/Simulink to generate signals representing thermal and mechanical behavior under various operating conditions. Several AI models were then applied, including SVM, ANN, Autoencoder, and the hybrid CNN–LSTM model. The results demonstrated the superiority of the CNN–LSTM hybrid model in terms of accuracy, achieving an accuracy of 99.84% with a recall value of 1.0000, reflecting its high ability to detect all failure cases. However, the results showed a relative decrease in precision, indicating the presence of some false alarms. This research provides a simulation-based framework that can support predictive maintenance research. However, further validation using real-world data is required to confirm its applicability in practical wind turbine systems.
Ahmed et al. (Fri,) studied this question.