Fault detection in wind turbine based on machine learning

Abstract

Renewable energy in general and wind energy in particular are receiving increasing attention with the goal of reducing greenhouse gas emissions and producing clean energy. In recent years, wind farms and plants have significantly increased, driving wind energy to become an immensely potential energy source. However, due to the unpredictable nature of wind energy, ensuring the safe operation of wind turbine systems, minimizing downtime due to malfunctions plays a crucial role in optimizing production costs and enhancing the system's reliability. In this study, the authors propose using machine learning models to detect issues occurring in wind turbine systems. Operational parameters measured from SCADA are used as input data for the machine learning models. The results indicate that machine learning models can detect issues in the wind turbine system with an accuracy of over 99%, with model training taking only tens of milliseconds.