Abstract

In recent years, wind turbine condition monitoring based on Supervisory Control and Data Acquisition (SCADA) systems has attracted considerable scientific research interest. Frequently reported challenges include the fact that most wind turbine SCADA parameters are highly dependent on the operating conditions, such as wind speed, wind direction, and LV Active Power, along with the control actions imposed on the wind turbine. Thus, strict and effective data quality control of the SCADA data is crucial. Besides that, intelligent anomaly detection for wind turbines using artificial intelligence techniques has been extensively researched and yielded significant results. Likewise, the usage of machine/deep learning techniques is widely spread and has been implemented in the wind industry in the last few years. The development of sophisticated deep learning now allows improvements in anomaly detection from historical data. In this paper, we present a wind speed anomaly detection approach using LSTM (Long Short-Term Memory), CNN (Convolutional Neural Network), and GRU (Gated Recurrent Unit) to detect the minimum and maximum values of wind speed. The approach applies the information in supervisory control and data acquisition systems of Aeolian wind speed. This comprehensive approach offers a promising avenue for the precise anomaly detection of Aeolian wind speed, providing practicians with a reliable tool for accurate diagnosis, critical for timely intervention. Real cases from a wind farm have confirmed the feasibility and advancement of the proposed model, while also discussing the effects of various applied parameters.

Statistics