- 學位論文
不同氣動噪音預測方法應用於風車葉片之比較研究
A comparison of different approaches on the aeroacoustic noise radiated from airfoils for wind turbine blades
摘要
本文主要利用Lighthill聲學類比理論為基礎的寬頻帶噪音源模式(Broadband Noise Source Model, BNS)與FW-H(Ffowcs Williams and Hawkings)方程式，模擬氣流流經風車葉片時所產生的噪音，針對氣動力噪音部分進行研究，探討不同的入流速度、入流攻角及紊流強度對氣動噪音產生的影響。首先對於NACA64(3)-618、S809、S822三種翼剖面進行流場的驗證與比對，在得到正確的紊流場資訊後，方可進行聲場之分析。文中採用不同的氣動噪音預測方法，探討噪音產生之相關機制及影響因素，首先，使用雷諾平均法(Reynolds-averaged Navier-Stokes Equation)求解穩態流場分佈，而紊流之模擬則採用 模式，在數值分析中，將入流紊流強度分別增加至5%與10%，從結果中得知入流紊流強度的增加，會使得流場產生劇烈的變化，對於噪音更是重要的影響因素之ㄧ。在聲場分析中，採用寬頻帶噪音源模式，運用Proudman方程式可得到聲音能量密度在整個計算區域中所分佈的情形，以及使用Curle方程式可進行表面聲功率的預測。另外，本文利用大尺度渦漩模擬方法(Large Eddy Simulation)求解暫態流場，在聲場中以FW-H方程式來分析計算噪音的暫態響應，藉此預測其聲壓訊號，經由快速傅立葉轉換得到聲音壓力位準頻譜，並進而求得其聲功率。而此一方程必須以流場的紊流資訊作為音源項，為獲得正確的紊流場資訊，網格的分割需相當的細膩，因此計算量較為龐大。為簡化分析，在本文僅以小型風車葉片作為分析對象，累積相關分析經驗，以作為將來大型風車葉片噪音預測的參考。
並列摘要
The purpose of the research is to investigate the noises induced by flow over the wind blades. The noise analysis is conducted by the Broadband Noise Source Model and FW-H (Ffowcs Williams and Hawkings) Formula which are based on theory of Lighthill’s acoustic analogy. How the wind velocity, angle of attack as well as the inflow turbulent intensity influence the induced aerodynamic noise is discussed. First of all, the dynamic coefficients and flow field of three airfoils NACA64(3)-618、S809 and S822 were verified, and then the accurate information of turbulence was provided as the source to evaluate the sound energy distribution. Three types of noise models that provided different characteristics of the noise distribution were adopted in this wrok. Firstly Reynolds-averaged Navier-Stokes Equation with the k-e turbulent model was used to predict the turbulent flow field. When the inflow turbulent intensity was increased to 5% and 10%, it causes great changes to the flow field and obviously it is also one of the major facts to the flow induced noise. For aerodynamic noise analysis, Proundman’s BNS model was performed to get the acoustic energy density distribution over the entire calculating domain. Further, Curle’s Formula was adopted to predict the surface acoustic power along the solid boundary. In order to understand the details of flow induced noise one step further, the Large Eddy Simulation approach for the unsteady flow combined with the FW-H equation was used to predict the unsteady sound pressure signal. Then by Fourier Transformation the spectrum of the noise can be calculated and consequently the frequency distribution and the power output are achieved. It might be useful in reducing the flow induced aerodynamic noise. However, LES requires a very fine grid resolution to capture the large scale eddy. At this stage, our current computer resources are extremely difficult to satisfy the computational efforts. Therefore, only the small wind blades were taken as the analysis object in this study. This experience may be useful in large wind blade analysis in the near future.