全域可壓縮流場計算方法之開發應用於氣動聲學之前期研究

Abstract

隨著生活品質的提昇，噪音問題越益受到重視，其中又以流體所誘導的氣動噪音最為複雜。由於計算聲場時，需先計算流場以獲得相關的聲源資訊，因此本研究將發展全域可壓縮流場計算方法，作為將來聲場計算之流場部份的基礎。為了獲得精確度極高之聲源資訊，首先改善過去將流體區分為可壓縮與不可壓縮流的缺失，利用Preconditioning法使得程式適用於全域範圍。除此之外，由於邊界條件對於流場與聲場的影響甚鉅，因此計算方法需輔以非反射性邊界，此法不僅提高準確度，更可節省計算空間提昇效率。在紊流模擬方面，本研究更開發出隱式紊流模式：dissipative model，此法利用數值耗散當作紊流模式，不需要額外的方程式描述小渦流運動，計算效能大幅提昇。除此之外，也適用於曲線座標，因此實用性更加廣泛。最後，由於本程式計算過程極為複雜且需要大量的時間，因此發展OpenMp與GPU(Graphics processing unit)平行化計算，大幅提昇計算效能。綜上所述，本研究所發展的流體力學計算方法相當成熟，因此可作為氣動聲學流體計算部份的基礎。

With the improvement of the life quality, the noise induced by aeroacoustic becomes one of the most important and complicated problems of the noise control. Before resolving the problems of the aeroacoustics, due to the necessary of defining the aeroacoustic source, the aim of this study will develop a computational method for all speed regions. In order to obtain the accurate information of the aeroacoustic source, the preconditioning method is adopted to make the program is suitable for all speed regions. Besides, due to the serious effect of the boundary conditions on fluid and acoustic fields, the computational method should be matched with the non-reflection boundary condition. With the non-reflection boundary condition, the accuracy can be increased and the efficiency can be improved. In the turbulence modeling, the dissipative model is created to generalize the computational method. The extra equations for the motions of small scales can be elimated and the efficiency can be improved. Besides, this model is also available in the curvilinear coordinate without any modifications so it can be used more general in real applications. Finally, because the tremendous computational time is demanded, parallel computation by OpenMp and GPU are built to improve the efficiency. Therefore, the computational method developed in this study can be potentially a basement in the computational aeroacoustic.