一個平行化耦合DSMC-NS模擬程式的精進與通用化及其應用之研究

Abstract

高超音速 (hypersonic) 的流動分析在應用上是相當重要的。相關應用，例如：銳緣 區(sharp leading edges)、鈍型物尾流 (blunt body wakes)、衛星姿態控制器噴嘴形成之擴 張氣流場(expanding reaction control system plumes)、直接物理氣相沉積( directed PVD)、 噴射式化學氣相沉積( jet-type CVD)等等。在過去也已經有許多人針對此類流場做過不少 相關的研究，由於流場中包含有連續及稀薄的區域，然而利用數值模擬這些流場依然是 相當困難及具挑戰性的工作。根據我們最近的研究成果 [J.-S. Wu* et al., Journal of Computational Physics, Vol. 219, No. 2, pp. 579-607, 2006.]，我們計劃要在朝這個方向繼續努力，加 強在此領域的研究深度。 在這次提報的三年計畫中，我們想要更進一步的改善我們之前的平行化DSMC-NS 偶合數值方法。我們會把我們以前針對三維空間平行化的DSMC 數值解法，由只能處理 純三維問題改善到能亦可處理二維以及二維軸對稱的問題。另外，我們會將利用DSMC 取樣技巧去研究過去文獻中所提出的失效參數( breakdown parameters)門檻值是否適 當。接著我們會修改及驗證我們的平行化DSMC-NS 偶合數值方法，以完成此計劃所面 臨的挑戰。最後，我們會在平行化DSMC-NS 偶合數值方法中加入網格自動重繪的技巧 (包含網格細切及合併)，進而增加數值模擬的準確性。我們會藉由模擬數個具有代表性 的流場來驗證這DSMC-NS 偶合數值方法的能力，包含RCS (reaction control system) nozzle plume impingement、high-altitude missile plume interaction 和 jet-type CVD。 每一年的具體目標陳列如下： 第一年， 我們會修改和驗證以前發展的DSMC，使它可以處理二維、二維軸對稱和三 維的稀薄氣體流場。除此之外，將利用DSMC 取樣技巧來研究出可接受的連續失效和熱 平衡失效參數的門檻值，我們會把重點放在靠近壁面區域，我們可能會提出一個新的失 效參數準則與指標。 第二年， 我們會結合修改過的DSMC 與NS 數值解法(UNIC-UNS code)，繼續發展及驗 証偶合的作法。我們會把重點擺在證實DSMC-NS 偶合數值方法証確性，尤其是二維軸 對稱流場的模擬驗證上。此外我們將研究新的失效參數指標對偶合次數和模擬時間及收 斂的影響。 第三年， 我們會在DSMC-NS偶合數值方法中加入平行化網格自動重繪的技巧( parallel adaptive mesh refinement and coarsening, PAMRC)進而改善模擬的準確性。在計畫結束前 我們會模擬出數個具有代表性的二維軸對稱及三維流場題目。這些題目最主要的困難處 為不適合單獨使用DSMC 或NS 求解。 本計劃的完成，我們有信心將使台灣成為偶合 DSMC-NS 數值方法研究的領先國家之 一，以此自勉。

Hypersonic flows of practical importance often involve flow fields having continuum and rarefied regions, e.g., blunt body wakes, sharp leading edges, and expanding reaction control system plumes, directed physical vapor deposition, and jet-type chemical vapor deposition, among others. Numerical solution of these important flows is very challenging, which has attracted numerous studies in the past. Based on our previous efforts [J.-S. Wu*, Y.-Y. Lian, G. Cheng, R. P. Koomullil and K.-C. Tseng, "Development and Verification of a Coupled DSMC-NS Scheme Using Unstructured Mesh, " Journal of Computational Physics, Vol. 219, No. 2, pp. 579-607, 2006.], we propose to continue our efforts in this direction. In this proposed two-year project, we intend to further improve and generalize our previous work in parallel hybrid DSMS-NS scheme using unstructured mesh. We shall generalize our previously devloped parallelized 3D DSMC code (PDSC) from 3D only to include both 2D and 2D-axisymmetric. In addition, we shall conduct a kinectic study of breakdown parameters using DSMC sampling technique to reinvestigate the criterion of continuum and thermal equilibrium breakdowns. Then, we shall modify and verify our hybrid DSMC-NS scheme to accommodate this new change. Finally, we shall combine the parallel mesh refinement and coarsening module embedded in the parallel NS code (UNIC-UNS) with the proposed hybrid DSMC-NS scheme to further improve the accuracy of the numerical simulation. We shall demonstrate the capability of this genealized DSMC-NS code by simulating several challenging flow probles, including RCS (reaction control system) nozzle plume impingment, high-altitude missle plume interaction, and jet-type chemical vapor deposition. Specific major tasks of each year are shown in the following for brevity. In the first year, we shall modify and verify a previously developed PDSC (3D) using unstructured mesh into a general DSMC code which can deal with 2D, 2D-axisymetric and 3D rarefied gas flows. In addition, a kinetic study which utlizes the DSMC sampling technique shall be used to reinvestigate the well-accepted criterion of the continuum and thermal-equilibrium breakdown parameters. We shall focus on the region near the solid wall. We may proposed new criterion for the breakdown parameters. In the second year, we shall combine the newly improved PDSC with newly available UNIC-UNS code (developed by the co-PI, a parallel NS code using unstructured mesh) following previously developed coupling procedures. We shall focus on verifying the hybrid DSMC-NS code, especially in the axisymmetric flow simulation, which has tremendous applications. In addition, we shall study the impact of the he new breakdown criterion to the hybrid code in terms of number of couplings and simulation time. In the third year, we shall combine the available parallel adaptive mesh refinement and coarsening (PAMRC) module with the hybrid DMSC-NS code to further improve the accuracy and applicability. We shall simulate several challenging 2D axisymmetric and 3D problems at the end of the project, which are impossible otherwise using DSMC or NS solver alone. Completion of this project shall provide a general-purpose tool for simulating several challenging and important hypersonic flows involving continuum and rarefied regions, which makes Taiwan in a leading position in this regard.