AN APPROACH TO CONSTRUCTING WIND FARM POWER CURVES: A CASE STUDY USING WINDPRO

Offshore wind power is becoming a vital part of the global energy mix, increasing the need for accurate modeling of wind farm output. Traditional single turbine power curves fail to capture the aggregated behavior of full wind farms, where wake effects, wind speed variation, and asynchronous control responses lead to significant deviations. This thesis proposes a simulation-based approach to construct an aggregated power curve (APC) that reflects the physical and statistical characteristics of offshore wind farms. Using WindPRO, a 100-turbine offshore wind farm with a regular 10 × 10 layout was modeled. Ten years of high-resolution reanalysis wind data (NOW-23) served as inputs, with five meteorological points used for spatial interpolation. The simulation incorporated key physical mechanisms, including the Jensen wake model, the Forsting blockage model, and high wind ride through (HWRT) turbine behavior. Time-varying mode was adopted to simulate independent responses under realistic wind conditions. The resultant APC was constructed using wind speed binning and Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) fitting. Compared to single turbine curves, the APC shows smoother growth in the mid-wind range and delayed curtailment at high wind speeds. A sensitivity analysis was conducted by rotating the wind farm axis at 30°, 45°, and 60°. Capacity factor differences across orientations were all below 0.01, indicating strong robustness of the APC structure. This modeling approach avoids reliance on historical power data and offers high interpretability and reliability. It provides a useful tool for grid planning, wind power integration studies, and performance assessment of large-scale offshore wind farms.