離散裂隙網路數值模擬：以花蓮溪畔坑道花崗片麻岩體為例

本文以座落於台灣東部花崗片麻岩體的溪畔坑道為目標坑道，於南北兩坑道壁設置共六條測線及兩個採樣面，以人工採樣方式量測坑道壁上的裂隙資料，目的則在模擬出目標坑道中的三維離散裂隙網路(Discrete Fracture Network, DFN)，並評比本實驗室發展的DFN數值模擬工具程式(簡稱DFN_OPT)與商用軟體FracMan®之功能。分析裂隙資料的工具採用本實驗室自行發展的裂隙資料地質統計分析軟體，可針對裂隙位態、間距、軌跡長度、裂隙強度等資料進行資料校正及分析。溪畔坑道內的裂隙位態較接近Fisher或Bivariate Fisher分布，並可找出三組高角度及兩組中低角度裂隙，其中高角度與低角度的裂隙分別反應花崗岩體在生成過程中因岩漿穿入而產生之裂隙，以及花崗岩抬升至地表後產生之解壓節理。其餘分析結果顯示，裂隙軌跡長度與裂隙間距等均相當接近對數常態分布。利用裂隙位態資料，本文首先利用Miller法得出溪畔坑道中的裂隙資料具有均質性。DFN_OPT根據模擬燧火法(Simulated annealing, SA)原理，反覆擾動DFN模擬結果，直到最終的DFN分布特性接近現場觀測的裂隙分布特性，且在擾動的過程中發現，視窗採樣資料的裂隙強度(P21)及裂隙位態為控制擾動結果好壞的主要因素。FracMan®因並未建立類似的最佳化模組，故本文僅利用現場量測資料以FracMan®進行條件模擬。比較DFN_OPT與FracMan®的模擬結果發現，兩種模擬工具所得裂隙位態的模擬結果均相當接近現場裂隙的位態分布。雖然FracMan®提供相當友善的使用者圖形介面，並提供完整的DFN展示環境，但因DFN_OPT加入最佳化DFN的功能，故DFN_OPT所得的裂隙軌跡長度、裂隙間距、裂隙強度及其空間相關性等，均比FracMan®的模擬結果更接近現場資料的裂隙參數特性，代表SA的擾動方式能夠得到較佳的裂隙參數事後分布(Posterior distribution)。

關鍵字

DFN_OPT ；溪畔坑道；離散裂隙網路(DFN) ； FracMan®

並列摘要

Sipan tunnel, excavated in a granitic gneiss rock body situated in eastern Taiwan, was considered as our target tunnel. Fracture traces on the north and south tunnel walls were manually measured from six scanlines and two scanwindows, with the objectives of simulating the three-dimensional discrete fracture network (DFN) and comparing the performance of the DFN simulator, DFN_OPT, and the commercial software FracMan®. Fracture attitude, spacing, trace length, density were analyzed by a geostatistical code. Results showed that fracture attitude is close to Fisher or Bivariae Fisher distribution. Three high-angle and two medium-angle fracture sets were identified. The former and the latter sets are related to cooling joint and release joint that are commonly seen in granitic rocks. Generally, trace length and spacing tend to be log-normally distributed. Prior to DFN simulation, Miller’s method was used to analyze fracture attitude and concluded that the sampled fracture traces can be treated as coming from a homogeneous population. Simulated annealing (SA) is incorporated into DFN_OPT for repeatedly perturbing simulated DFNs until an optimal one that best reflects field conditions is obtained. It was found that fracture intensity (P21) and fracture attitude are the two key parameters that control the convergence of perturbation. As no global optimization scheme is taken into account in FracMan®, FracMan® simulations were obtained by conditioning DFN simulation results to field data. Fracture attitude derived from both DFN_OPT and FracMan® simulation results are consistent with those derived from field data. However, compared to FracMan® simulations, posterior distributions of other fracture parameters and the spatial correlation of P21 derived from DFN_OPT simulations are closer to the prior distributions analyzed from field data, manifesting the fact that SA is able to obtain a realization that can be better validated against field observations.

並列關鍵字

DFN_OPT ； Sipan tunnel ； Discrete Fracture Network ； FracMan®

參考文獻

Andersson, J. and B. Dverstorp, 1987, Conditional simulations of fluid flow in three-dimensional networks of discrete fractures, Water Resources Research, 23(10), 1876-1886.

Baecher, G., 1983, Statistical analysis of rock mass fracturing, Mathematical Geology, 15(2), 329-348.

Billaux, D., J. P. Chiles, K. Hestir, and J. Long, 1989, Three-dimensional statistical modelling of a fractured rock mass-an example from the Fanay-Augères mine, International Journal of Rock Mechanics, Mining Science & Geomechanics Abstracts, 26(3-4) , 281-299.

Call, R. D., Savely, J. P., and Nicholas, D. E., 1976, Estimation of joint set characteristics from surface mapping data, Proceedings of the 17th U.S. symposium on rock mechanics, Salt Lake City, Utah, 2B2.1-2B2.9

Deutsch, C. V. and A. G. Journel, 1998, Geostatistical software library and user’s guide, Oxford, 369p.

國際替代計量

離散裂隙網路數值模擬：以花蓮溪畔坑道花崗片麻岩體為例

未授權

主題瀏覽