Leveraging Artificial Intelligence to Promote Sustainability: Employing Machine Learning Techniques in Planning Predictive Maintenance of Wind Energy Systems

As industrial activities increasingly impact the environment, the need for resource conservation and long-term energy security becomes paramount, positioning sustainable electricity production as a critical focus area. In this regard, the application of Machine Learning, a branch of Artificial Intelligence, has revealed considerable promise in optimizing various dimensions of energy generation, especially within renewable sectors such as wind energy. This research presents a case study that showcases the implementation of Machine Learning techniques in forecasting the failures and establishing predictive maintenance plans for wind energy systems. In this study, we employed Logistic Regression, Decision Tree, Bagging, Random Forest, Gradient Boosting, Adaboost, and XGBoost to analyze more than 40,000 rows of data from 40 different system parts. At the end of the analysis, Tuned XGBoost model has turned out to be the best model having the highest Recall value that correctly predicts the Failures. Through the application of diverse methodologies, we demonstrate the effectiveness of these techniques in anticipating potential failures and prioritizing maintenance or replacement of parts, which in turn facilitates substantial resource savings in terms of labor, costs, and time, thereby advancing sustainability and strategic planning.