固結岩體中裂隙主導地下水儲存、流動與傳輸，但相對於岩體，裂隙佔據較小體積，因此當分析區域擴大時，在模擬域適用區大小的選擇上及以裂隙尺度分析大區域水流與污染傳輸上遇到困難。為求執行上之可行性，離散裂隙通常透過生成等效連續體參數分析大區域裂隙水流與污染傳輸問題。以上方法已成為現今重要的分析方法，但在計算上仍需改善，使其有效減少計算量提高評估效益。本研究提出利用現地資料分析結果，直接判斷模擬區域內裂隙水力連通機率的大小，做為升尺度模擬適用區的參考，並且利用假想裂隙資料做水力連通性分析。除此之外，本研究也利用前人研究調查數據建立數值模式，透過藉由裂隙特性做判斷的Oda方法及以水流逆推的Block法，進行升尺度參數轉換及差異性評估，希望建立最符合效益之計算方法。根據研究結果顯示，透過敏感度分析後可量化出不同裂隙參數的連通機率，並提出最小連通裂隙強度做為模擬區域尺寸選擇之參考指標；透過前人現地資料模擬結果，可畫分出兩種升尺度方法之適用範圍，在塊體網格小於裂隙片網格時適用Oda方法，而塊體網格遠大於裂隙片網格且連通性較少時適用Block方法，則可達到模擬最佳效益。透過本研究所建立之升尺度判斷標準，可透過現地之裂隙參數，升尺度網格與裂隙片網格尺寸進行篩選，提供未來升尺度模擬之參考。;The fractures in the rock matrix are fundamental units that control flow and contaminant transport in fractured formations. Because of complex network connectivity and relative small volumes of the fractures, the simulations of flow and transport in large-scale fractured formations have become challenge tasks to resolve flow and transport in 3D discrete fracture networks (DFNs). The flow parameter upscaling is one of the typical approaches to account for the fracture flow behavior in large-scale problems. The upscaled parameters (also call equivalent parameters) rely on accurate calculations of fracture connectivity and the representative elementary volume (REV) sizes (i.e., numerical cell sizes) for upscaling. The objectives of this study are to quantify the effects of fracture distribution parameters on fracture connectivity for different REV sizes, and to assess the upscaled flow parameters based on different upscaling techniques. A synthetic fracture distribution was used for estimating statistics of the fracture connectivity. The comparison of Oda and Block upscaling techniques in FracMan software were conducted to assess the efficiency and accuracy of the estimated equivalent flow parameters for specified REVs. In this study a field fracture distribution was used for the comparison. The results show that the fracture connectivity mainly controlled by the fracture intensity. The statistics of the fracture connectivity can be the important index to decide a suitable REV size for 3D DFNs. The comparison of two different upscaling techniques showed that the accuracy of the upscaled flow parameters might relevant to the cell size for resolving 3D fractures in the FracMan software. Because of the conceptual differences employed in Oda and Block upscaling techniques, the test cases showed that the Oda technique is efficient for high fracture density location 3D DFNs and the Block technique can result in relatively accurate flow parameters for low fracture intensity locations in 3D DFNs. The results of the study are important references for future works in selecting suitable simulating cells for upscaling 3D DFNs in large-scale flow and transport problems.
