基於臺灣東北季風觀測資料之離岸風機尾流模型參數研究

Abstract

臺灣海峽風能資源豐富，近年來離岸風電產業快速發展，而風機尾流效應對風場發電效率與功率預測準確性具關鍵影響，值得深入探討。由於臺灣冬季盛行東北季風，具風速強勁、風向穩定之特性，本研究選取東北季風期間觀測資料，以呈現冬季典型風場。工程上常以 Jensen 模型描述尾流行為，惟多數研究僅依單點或二維風速量測分析，難以全面反映尾流結構與變化。本研究利用丹麥風電開發商沃旭（Ørsted）於臺灣竹南海洋風場所建置之雙都卜勒雷達系統，結合多角度風速量測與 Jensen 模型，探討 6 MW 風機在不同空間配置與風況條件下之尾流擴散行為與模型參數校準，涵蓋單風機尾流、上下游風機互擾與海陸風條件差異等情境。分析採用公稱轉子直徑與以相對風速閾值法估算之實測尾流直徑兩種假設，並分別以尾流中心風速與尾流擴散半徑推估擴散係數 k。結果顯示，實測直徑假設更適用於近場與尾流重疊情境，能有效提高模型擬合精度。單風機情境下，實測直徑假設的 k 值約為 0.06 至 0.10，邊界法約 0.07 至 0.09，略高於公稱直徑假設，能更準確反映尾流中心風速衰減與橫向擴張行為。雙風機分析中，下游風機尾流之 k 值平均約 0.18，顯示尾流重疊導致混合顯著增加。海陸風條件比較顯示，海風條件下紊流活躍，中心風速法 k 約為 0.09，邊界法約 0.03；陸風條件下大氣穩定度較高，紊流活動受抑制，中心風速法 k 下降至約 0.05，邊界法上升達 0.04至 0.05，顯示尾流在穩定環境中雖恢復較慢，但橫向擴張更為顯著。綜合而言，尾流擴散係數 k 具明顯的情境依賴性，隨風機佈局、大氣穩定度、紊流強度及尾流度量方式而變化。本研究成果可供離岸風場佈置設計與發電效率評估參考。

The Taiwan Strait possesses abundant wind energy resources, and in recent years, the offshore wind power industry has developed rapidly. The wake effect of wind turbines has a critical impact on wind farm power generation efficiency and the accuracy of power output prediction, making it an important topic for further investigation. During winter, Taiwan is dominated by the northeast monsoon characterized by strong wind speeds and stable wind directions. This study utilizes observational data collected during the northeast monsoon period to represent typical winter wind field conditions. In engineering applications, the Jensen model is commonly employed to describe wake behavior; however, most studies determine its parameters based solely on single-point or two-dimensional wind speed measurements, which limits the comprehensive representation of wake structures and their variations.This research employs dual-Doppler radar systems installed by Ørsted, a Danish wind power developer, at the Zhunan offshore wind farm in Taiwan. By integrating multi-angle wind speed observations with the Jensen model, this study investigates the wake expansion characteristics and parameter calibration of 6 MW wind turbines under different spatial configurations and wind conditions, encompassing single-turbine wakes, upstream–downstream turbine interactions, and sea–land breeze conditions. Two assumptions are adopted: one based on the nominal rotor diameter and another derived from measured wake diameters estimated using the relative wind speed threshold method. The wake expansion coefficient k is estimated using both the wake centerline wind speed and wake boundary radius approaches.The results indicate that the measured-diameter assumption is more suitable for near-wake and wake-overlapping conditions, effectively improving model fitting accuracy. For single-turbine cases, the k values based on the measured-diameter assumption range from approximately 0.06 to 0.10 using the centerline method and 0.07 to 0.09 using the boundary method, slightly higher than those obtained under the nominal-diameter assumption, and better capturing the wake centerline wind speed decay and lateral expansion behavior. In two-turbine analyses, the downstream turbine exhibited an average k value of about 0.18, indicating that wake overlap significantly enhances mixing intensity. The comparison under sea–land breeze conditions shows that during sea-breeze periods, turbulent activity is strong, with k values of about 0.09 and 0.03 for the centerline and boundary methods, respectively; whereas under land-breeze conditions, higher atmospheric stability suppresses turbulence, resulting in k decreasing to around 0.05 for the centerline method and increasing to 0.04–0.05 for the boundary method. These results indicate that under stable atmospheric conditions, wake recovery is slower but lateral expansion is more pronounced. Overall, the wake expansion coefficient k exhibits clear dependence on environmental conditions and varies with turbine layout, atmospheric stability, turbulence intensity, and wake evaluation method. The findings of this study can serve as valuable references for offshore wind farm layout design, wake effect assessment, and power generation efficiency evaluation.