| 摘要: | 本論文旨在使用計算流體力學(Computational Fluid Dynamics, CFD)方法模擬旋轉效應對葉片的影響。有鑑於過去風機葉片氣動力負載計算以葉片元素動量理論(Blade Element Momentum, BEM)為主要方法，該方法須仰賴二維翼型的升力與阻力係數實驗數據，並以理論公式修正，以符合實際三維流場的影響。本研究使用商用軟體ANSYS Fluent，SST k-ω紊流模型，以移動座標方式(Moving Reference Frame, MRF)模擬計算風機葉片氣動力負載，結合計算流體力學流場可視化的優點，可直接解析三維流場現象，無需依賴實驗數據。分析的條件參考NREL Phase VI 風洞實驗，條件為3度節距角(pitch angle)，7 m/s、10 m/s、13 m/s、15 m/s、20 m/s、25 m/s六種風速，模擬結果與實驗結果實施比對，並探討葉片的氣動力負載，分析項目包括推力、扭矩、葉片流線、壓力係數、推力係數、扭力係數、升力係數、阻力係數。結果顯示模擬的旋轉現象與理論相符，即翼根處有失速延遲(stall delay)的現象；在翼尖處有翼尖損失(tip loss)的現象。但由於模擬的升力係數較低，阻力係數較高，使得扭矩值平均小於實驗值約20%。此外，風速10 m/s至15 m/s，流場處於不穩定狀態，扭矩計算誤差最大可達28%。本文另以滑移網格(Sliding Mesh Method, SLM)實施暫態模擬，風速10 m/s的扭矩計算誤差可修正至7%。透過本次研究可得結論如後，使用計算流體力學方式可以模擬葉片三維旋轉效應，但是靠近葉片壁面的流場，如分離與迴流，仍然有誤差。;Nowadays the main approach to the aerodynamic loading calculation for wind turbines is BEM (Blade Element Momentum) method. However, it heavily relies on the experimental data of the lift and drag coefficients which need further modification to match the actual three-dimensional flow around a wind turbine. This paper presents a computational fluid dynamics simulation for the rotational effects on the wind turbine blade. Therefore, there is no need to collect the lift and drag data by experiments. A commercial software, ANSYS Fluent, with MRF (Moving Reference Frame) method is used to simulate the aerodynamic load under uniform wind conditions as to combine the advantages of CFD flow field visualization to further understand the three-dimensional flow phenomena. The analysis conditions refer to the NREL Phase VI wind tunnel experiments, considering the pitch angle 3o, upwind speeds from 7 m/s to 25 m/s with no yaw angle. Six wind speeds are compared with the experimental results and the aerodynamic loads of the blade are discussed. The analysis items include thrust, torque, streamline, pressure coefficient, thrust force coefficient, torque force coefficient, lift coefficient, and drag coefficient. The results show that there is a stall delay phenomenon at the inboard of the wing and a tip loss phenomenon at the wing tip. These phenomena are consistent with the theory. The simulated torque is less than the experimental value by about 20% on average because the simulated lift coefficient is lower while the drag coefficient is higher than their experimental counterpart. When the wind speed is between from 10 m/s to 15 m/s, the flow field is unstable, the relative error from calculated torque is up to 28%. Compared with the result from MRF method, the relative error of torque is reduced to 7% by using the sliding mesh method. It is concluded that the CFD method can simulate the 3-D effects on the blade, but it requires further research effort to reduce the errors for the separation and reverse flow near the blade surface. |
