集水廊道附近地下水流數值模式率定方法建立與應用—以林邊溪為例

Abstract

本研究旨在建立一有系統之集水廊道附近地下水流數值模式率定方法，並將此一模式率定方法應用於林邊溪集水廊道之實際案例中，以彰顯所建立方法之可用性。 本研究建立之集水廊道附近地下水流數值模式率定方法，包含地下水系統之水平衡計算、建立數值模式及其參數檢定優選模式三部份。首先，透過地下水系統之水平衡計算，可充分瞭解其河床垂向入滲補注、上游流入及流出量以及集水廊道之集水量，該方法首先透過地下水系統之觀測水位及河床沖積層辨識成果，計算邊界入流或出流量以及蓄水量變化率，並代入連續方程式反推河床垂向入滲量；其次，建立數值模式及其參數檢定優選模式，求解地下水系統最佳之土壤參數組合、河道入滲強度以及降雨入滲強度佔總降雨強度之比例。上述之參數檢定優選模式，係經較有效率之試誤方法迭代求解，直至達到目標函數值最小為止。集水廊道處理方面則以刪除網格的方式，將集水廊道概念化為一明渠之型式，刪除的網格即表示其中空部份，左右兩面及底面給定為混凝土之材質，頂面則視濾層種類及阻塞程度而有所不同，並在混凝土材質內側邊界上之節點以給定水頭邊界條件以模擬集水廊道受重力驅動之水力特性，在模擬完成後，將這些節點上之流量進行加總，即得集水廊道之集水量。 本研究將所建立之方法實際應用於林邊溪集水廊道附近之地下水系統，其地下水平衡分析計算結果顯示2010年1月1日至2010年9月5日間，總補注量為4.19千萬噸，其中河床垂向入滲量佔98.9%，上游邊界補注佔1.1%；總流出量為4.13千萬噸，其中下游邊界流出量佔35.6%，集水廊道集水量則佔64.4%。 林邊溪集水廊道附近地下水流數值模式率定結果顯示，地下水系統於2010年1月1日至2010年9月5日觀測井C旱季與雨季均分根誤差(RMSE)分別為0.17公尺及0.08公尺、效率係數(CE)分別為0.93及0.98、相關係數(CC)分別為0.97及0.99；觀測井D旱季與雨季RMSE分別為0.27公尺尺及0.14公尺、CE分別為0.67及0.87、CC分別為0.93及0.94；觀測井E旱季與雨季RMSE分別為0.33公尺及0.41公尺、CE分別為0.83及0.92、CC分別為0.97及0.99；集水廊道集水量旱季與雨季誤差分別為RMSE分別為15,827(噸/日)及8,971(噸/日)、CE分別為0.83及0.72、CC分別為0.96及0.85。總補注量4.28千萬噸，其中河床垂向入滲量佔99.2%，上游邊界補注佔0.8%；總流出量為4.29千萬噸，其中下游邊界流出量佔41%，集水廊道集水量則佔59%。

This thesis aims to establish a numerical model calibration method for groundwater flow near infiltration gallery, and apply this method to an in-situ case of infiltration gallery in Lin-Bien River; in order to highlight the usage of the method had built. The numerical model calibrating method established by this thesis includes calculating groundwater balance near infiltration gallery, building numerical model and formulating optimization model for parameter verification. First, by calculating system groundwater balance, we will realize the ground surface boundary recharge, upstream inflow quantity, downstream outflow quantity and water intake of infiltration gallery. The method computes the quantity of recharge through every system boundary, the rate of change of system water storage and calculates river and rainfall recharge by continuity equation. Second, build a numerical model and its optimization model for parameter verification; solve the best combination of soil parameter, depth/infiltration recharge of river and rainfall infiltration proportion. Above-mentioned optimization model for parameter verification will solve by iterative method with a more efficient trial and error method, until objective function is minimum. On the other hand, delete the mesh to simulate the space in the infiltration gallery, and conceptualize it into a form of open channel, the right, left hand and bottom side are concrete material, and top side is decide as different filter layer. Given head condition are set up in the concrete region to simulate the gravity controlled condition, after the simulation is finished, summation the flux on each node to get the infiltration gallery discharge. This thesis was applied to the process to groundwater system near infiltration gallery of Lin-Bien River. According to calculating system groundwater balance result, during 2010/01/01~2010/09/05, the total recharge is 41.9 million tons, river and rainfall infiltration recharge are 98.9% among them, upstream boundary recharge are 1.1% among them. Total out flow is 41.3 million tons, downstream outflow are 35.6% among them, water intake of infiltration gallery are 64.4% among them. According to the result of calibrating groundwater flow model near infiltration gallery of Lin-Bien River, during dry and rainy season in 2010/01/01~2010/09/05, RMSE of observing well C are 0.17 m and 0.08m respectively, CE are 0.98 and 0.998 respectively, CC are 0.99 and 0.999 respectively. RMSE of observing well D are 0.27 m and 0.14m respectively, CE are 0.89 and 0.98 respectively, CC are 0.98 and 0.99 respectively. RMSE of observing well E are 0.112 m and 0.114m respectively, CE are 0.98 and 0.99 respectively, CC are 0.98 and 0.996 respectively. RMSE of observing water intake of infiltration gallery are 5,465(cmd) and 2,958(cmd) respectively, CE are 0.94 and 0.99 respectively, CC are 0.97 and 0.995 respectively. The total recharge is 4.28 million tons, river and rainfall infiltration recharge are 99.2% among them, upstream boundary recharge are 0.8% among them. Total out flow is 42.9 million tons, downstream outflow are 41% among them, water intake of infiltration gallery are 59% among them.