Dynamic Wake Modelling for Offshore Wind Farms: Evaluation, Comparison and Scenario Adaptability Analysis

Abstract The wake expansion in offshore wind farms significantly affects wind turbine power generation, so the accuracy of wake models is crucial for the operational optimization of wind farms. Existing studies on wake models mainly focus on constructing and improving the models themselves, whereas comparative analyses of these models across diverse scenarios (e.g., wind turbine layouts and incoming wind conditions) remain insufficient. Based on operational data from the Supervisory Control and Data Acquisition (SCADA) system of an offshore wind farm, this study compares power generation predictions derived from simulated wake results of the modified Park models (such as Park-Max, Park-L) and the 3DJSG model, and conducts verification and analysis of their applicable scenarios. Case analysis shows that: in terms of the average bias of predicted power generation across all scenarios, the Park-Max model exhibits the largest average bias (5.8%); although it performs well in wake modeling for front-row turbines, the bias increases rapidly for downstream turbines. The Park-L model has the smallest average bias (1.6%) and also demonstrates the most stable wake modeling performance under varying wind speeds and turbine row positions. The average bias of the 3DJSG model falls between the two Park models (5.0%), but its performance is unstable: varying wind speeds and turbine row positions can significantly alter its model bias, even though it can achieve optimal simulations under certain operating conditions. These three models each possess scenario-specific adaptability: Park-L model delivers the best and most stable performance across different scenarios; Park-Max model is largely only applicable to front-row turbine scenarios; the simulation performance of the 3DJSG model across scenarios is unstable and varies largely with environmental factors such as wind conditions and turbine row number. Therefore, sufficient preliminary research is required before the model is used. The differences in model performance essentially stem from the selection of key parameters and formulas, including wake expansion coefficient and wake superposition methods. Through cross-model evaluation across multiple operating conditions, the research results provide an empirical basis for model selection and offshore wind farm operational optimization.