Algorithm for identifying wind power ramp events via novel improved dynamic swinging door

Abstract

With the rapid increase in the penetration of wind power in recent years, wind power ramp events (WPREs) have become the main factors affecting the safety and stability of electric power systems. Accurate detection of ramp events could help power systems better manage the extreme events and reduce economic losses. The previous ramp detection methods are either too complex to implement that influence the computing efficiency, or based on the value of points which cannot completely reflect the trend of data segments and lead to a decrease of accuracy. Based on the above problems and the on-site requirements, this paper proposes a novel improved dynamic swinging door algorithm (ImDSDA) to optimise the state-of-the-art in WPREs detection. Firstly, the swinging door algorithm (SDA) is used to extract ramp segments. Secondly, the dynamic programming method is used for ramp trend identification and segment combination. Finally, raw data obtained from three real-world wind farms in Hubei, China were applied to validate the performance of the proposed ImDSDA. The detection results show that the ImDSDA is more accurate and efficient than the traditional detection methods and could be a feasible option for WPRE detection in power systems.