Hybrid Deep Learning for Wind Turbine Fault Detection

Wind turbine fault detection is crucial for maintaining efficient and reliable renewable energy systems. This paper introduces a novel hybrid deep learning architecture, LSTM-Attention-CapsNet, which combines Long Short-Term Memory networks, attention mechanisms, and Capsule Networks for time series-based fault detection in wind turbines. Our proposed model achieved unprecedented performance metrics when tested on a wind turbine dataset, attaining ۱ accuracy, F۱ score, precision, and recall. This exceptional performance marks a significant advancement in fault detection capabilities, potentially revolutionizing predictive maintenance strategies in the wind energy sector. Our findings herald a new era in wind turbine fault detection and condition monitoring, promising substantial improvements in the efficiency and reliability of wind energy production

کلیدواژه ها:

Wind Turbine ، Fault Detection ، Deep Learning ، LSTM ، Attention Mechanism ، Capsule Neural Networks