應用於離岸風機之Jensen尾跡流修正模型

Abstract

本研究基於計算流體力學結果，開發新的非線性尾跡流模型，以預測風機跡流的動能損失以及速度分布，進而預測下游處風機的發電量。因應2050淨零排放(Net Zero)要求，臺灣政府也著重發展離岸風力發電產業。在密集的海洋風能開發政策之下，離岸風場的規劃與設計相當重要。由於風機跡流影響發電量甚巨，所以在設計風機位置時需要考慮跡流的影響，然而傳統的Jensen線性跡流模型預估速度分布與跡流邊界並不符合大型離岸風場實際情形。因此本研究提出新的非線性顯式模型，搭配動量理論的推導，分析複雜尾跡流現象。該模型假設風速、風向與轉速為定值，利用功率係數計算速度損失，進而分析下游風速與發電量。相較於Jensen模型使用單一擴張係數，新模型由四個定義流場特徵的參數組成，再以雷諾平均納維斯托克斯方程(RANS)的計算結果DTU 10MW風機的尾離情形，並進行新模型參數調校。在預估Horns Rev離岸風場中各別風機和整個風場發電量時，新模型使用極少量計算資源即比傳統Jensen模型更為接近觀測值。經結果顯示，在三種測試風向下，風向為270°誤差最大為13%，風向為222°誤差最大為16%，風向為312°誤差最大為16%；在風向為270°風向偏角±1甚至接近無誤差。新模型實現了任意風向、任意風速、任意風機架設位置等多重自由度，且展示了短時間預測大規模風場的實力。另外，本研究也利用誘導速度與推力的關係，提供了使用致動盤研究風機尾跡流的模型新的計算方法，將使結果更為準確。成果將可以更快速且精確地運用在建置風場時的可行性分析以及預估發電量。

This study develops a nonlinear wake flow model based on the result of computational fluid dynamics to predict kinetic energy losses and downstream velocity distribution. The offshore wind power industry in Taiwan targets to fulfill 2050 Net Zero emissions requirement. The intensive development of marine wind energy makes the planning and design of offshore wind farms crucial, with wake effects having a significant impact on power generation. However, the traditional linear wake model, e.g. Jensen, does not accurately predict velocity distribution and wake boundaries in large offshore wind farms. Hence, a new explicit nonlinear model is proposed to analyze complex wake flow phenomenon. This research assumes steady wind speed, wind direction and turbine rotation speed, and predicts velocity deficit using power curve, as well as analyze downstream wind speeds and power generation. Validation of Horns Rev offshore wind farm using the proposed model against shows high precision in the power prediction. The results show that, under three test wind directions, the maximum errors are as follows: 13% for the wind direction of 270 degrees, 16% for the wind direction of 222 degrees, and 16% for the wind direction of 312 degrees. In the case of a wind direction of 270 degrees with a deviation of ±1 degree, the error is nearly negligible. The model's versatility is exemplified by its ability to handle diverse wind directions, speeds, and turbine layouts, illustrating its efficacy in swiftly and accurately forecasting large-scale wind fields. In addition, this research introduces a computational methodology for analyzing turbine wake using the correlation between induced velocity and thrust. These will improve the accuracy of the researches using actuator disk model for wind turbine wake prediction. This approach provide a fast and accurate application in conducting feasibility analyses and estimating power generation during wind farm planning.