Wind Turbine Fault Detection Based on Time Series Monitoring Data

Qingyi Guo, Xing Wang, Linzhi Wang, Guohua Liu

Keywords: anomaly detection, wind turbine, deep learning

TL;DR: This project proposes a context-aware framework integrating heterogeneous data and temporal networks to enhance the robustness of wind turbine fault diagnosis in complex environments.

Abstract: Wind energy has become a cornerstone of the global transition to renewable power. However, wind turbines often operate in harsh, remote environments, leading to high structural fatigue and frequent component failures. Traditional fault detection relies on analyzing isolated signals—such as vibration—in the time or frequency domains. While effective for localized issues, these methods struggle to distinguish between genuine faults and normal operational variability caused by fluctuating wind resources. This project proposes a context-aware framework that integrates heterogeneous data sources (temperature, electrical, and meteorological data) through advanced temporal networks, aiming to provide a more robust diagnostic tool.

Submission Number: 5