2009-08-012024-05-17https://scholars.lib.ntu.edu.tw/handle/123456789/679638摘要：隨著全球氣候的變遷，由洪水引發的自然災害，例如土石流、山崩，泥流及洪災等，發生益加頻繁，且正嚴重威脅著我們的日常生活。從物理上，我們可以將這類自然流(natural flows)現象簡化為流固體之間交互作用的二相顆粒流(two-phase granular flow)。不過，其動態流動的特性-- 顆粒組成及流場隨著時間、位置產生變化，使得整體行為相當複雜。本研究藉由引入四個重要的無因次參數，即雷諾數(Reynolds number)、斯多克斯數(Stokes number)、巴格諾爾德數(Bagnold number)，及薩維基數(Savage number)，系統化地將自然流分類，以輔助模擬模式開發與自然流現象之間的對應。 本研究提出三種模擬二相顆粒流的模式，分別為離散顆粒模式、二相混合流模式，及二相流模式。離散顆粒模式主要依據離散元素法(Discrete Element Modelling)的原理，藉由適當地調整元素間碰撞時的彈性係數及阻尼係數，來反應出流體對固體顆粒的效應。因為不用計算流場，效率為三種模式中最佳，不過，也因此僅適用於流體效應較小的自然流現象。不同於離散顆粒模式，另外二種模式都必須求解流場。其中二相混合流模式是依循混合理論(mixture theory)，分別平均流體和固體的運動，然後以流固體之間的作用力將二相聯結在一起，因此如何計算交互作用力是此模式的關鍵，過去已有相當多的研究從理論、實驗，和數值等方面提出計算的方法，本研究將探討前人所提出的方法，討論不同方法所適用的自然流分類。二相流模式的概念則是直接求解流場，理論上，計算結果最精確，不過根據前人研究，當顆粒數較多的時候，此模式的效率最差。因此，本研究將探討現有的數值計算方法，包括Lattice-Boltzmann及Immersed Boundary Method等方法，來改善計算的效率，且將此模式計算流固體交互作用力的結果回饋給二相混合模式，以驗證、修訂，及擴充前人的理論。 本研究的模擬結果，將會與F. L. Yang （子計畫四）和 H. Capart（子計畫二）所主持的其他子計畫的實驗結果作驗證，並與D. L. Young（子計畫一）的計算結果作比較，以探討各模擬模式的適用性及實用性，並推演至實際尺寸的自然流模擬。<br> Abstract: Flow-induced natural hazards such as debris flows, rock avalanches, mudflows, and water floods have posed a great threat to the society. Many efforts have been made to study these complex phenomena of natural flows. However, their dynamic characteristics poses a great challenge to conduct a model for capturing their motion completely. This research aims to adopt four dimensionless variables, namely Reynolds number, Stokes number, Bagnold number, and Savage number, to characterize the behavior of particles and mixtures for various natural flows. Pertinent to the proposed classification, three types of simulation schemes will be developed herein. They are dry grains with wet spring, two-fluid model with empirical interactions, and two-phase model, respectively. The scheme of dry grains with wet spring is conceptually simple: we assume that the liquid effects can be properly lumped to the classical spring and dashpot models in discrete element modeling (DEM). Such treatment is considered to be the most efficient scheme since it does not have to estimate fluid field which usually spends most computational time and resources in two-phase problems. Nevertheless, this scheme may not be feasible for the phenomena in which fluid fields dominate. In contrast, the other two schemes stress on the technologies of estimating the fluid field. The theoretical challenge of the two-fluid model with empirical interactions scheme hinges on deriving analytically or numerically of the functional forms of empirical interaction when calculating the fluid-solid interaction. The two-phase model scheme will be developed through directly solving fluid fields. Novel methodologies and numerical technologies will be developed to speed up this scheme. The numerical results obtained from this subproject will be compared with the laboratory experiments conducted by F. L. Yang (subproject 4) and H. Capart (subproject 2) as well as numerical results by D. L. Young (subproject 1). Finally, the proposed simulation schemes will be extended to model large scale simulation of natural flows.二相流離散元素晶格波茲曼內嵌邊界混合理論土石流two-phase flow modelmixture theorydiscrete elementlattice-boltzmannimmersed boundarydebris flow