High-dimensional wind farm wake modeling and efficiency optimization using a hybrid modeling approach

Abstract

The performance optimization of wind turbines faces challenges due to wake interactions, leading to overall power losses in wind farms. Wake steering through the yaw control strategy has strong potential; however, conventional algorithms are hindered by high computational costs for large-scale real-time deployments. This study presents a fast and scalable approach for the performance optimization of wind farms under different environmental conditions through wake steering. The methodology integrates the Gauss–curled hybrid wake model with a serial refined to investigate how physically resolved wake mechanisms influence yaw control optimization while maintaining computational efficiency. The approach reduces high-dimensional optimization problems into sequential upstream to downstream turbine search for optimal yaw angles by reducing the computational cost and improving the physical accuracy. The framework was first applied to onshore wind farm turbines and then extended to an offshore wind farm comprising 96 wind turbines. The results indicate that wake interaction significantly affects the downstream performance owing to the wake overlap. The proposed method effectively models and mitigates wake effects through integrated wake modeling and an optimized yaw control strategy. The results demonstrate that dynamic wake interaction, yaw-induced wake deflection, and wake recovery enable physically consistent and scalable wake steering optimization under different environmental conditions. The agreement between the results obtained from the second wind farm and the monitoring data from supervisory control and data acquisition confirmed the reliability of the proposed framework for real-time applications. This wake steering technique verified by monitoring data provides an efficient tool for managing wake control across large-scale wind farm arrays.