標題: 最佳降低變異數分析應用於地下水監測井網規劃之應用
Application of optimal multiple-point variance reduction analysis to the groundwater monitoring network planning
作者: 林君怡
Lin, Chun-i
張良正
Chang liang-cheng
土木工程學系
關鍵字: 最佳化;optimal
公開日期: 1995
摘要: 本研究為利用最佳降低變異數分析法(OMPVR)來解決區域性地下水觀 測井網規劃問題。本模式係根據地質統計方法中的降低變異數分析法( VRA)繼續發展,加入多變數函數之非線性規劃理論,成為一具有最佳化特 性的地下水觀測井網規劃模式,可以增進VRA規劃方式之效率及提高其可 信度。 VRA模式必須將研究區域以格網表示,佈井的位置亦必須在格 點上,因此限制了可能設井的位置,並且由於計算量的限制,在此位置離 散化(格點化)的情況下,計算所得之佈井方案理論上並不是真正最佳的佈 井方案。本研究發展的OMPVR模式,不須對佈井位置進行空間上的離散化 ,並應用非線性規劃的理論求取最佳佈井位置,因此能求得較VRA模式更 佳的方案。根據本研究初步成果顯示,兩者之差距隨著井網密度增加而減 少。 本模式以MATLAB軟體發展其程式模組。文中以一個假設的研究場 址,先對模式進行驗證的工作。之後於實例的演算中,選擇濁水溪沖積扇 之彰化、雲林兩縣為研究場址,進行分析工作,決定最佳之井網規劃結果 。 This research applies a new algorithm, Optimal Multiple- Point VarianceReduction Analysis(OMPVR) to solve a problem of groundwater monitoring networkplanning. This model is the extension of Variance Reduction Analysis(VRA)combined with non- linear optimal procedures. The above newly developedprocedures present a series of highly effective algorithms for optimalgroundwater monitoring network design. VRA model has selected the additional wells from dense potential pointswhich may be pool. This study uses sparser potential points, and presents anew effective algorithm which not only considers the issue of optimaladditional monitoring well locations, but also avoids the spatialdiscretization. These optimal procedures increase the possibility ofsignificant cost savings. The model is developed by the MATLAB language. And these methods areapplied to both hypothetical and field cases. The first case is forverification, and the field case obtains the optimal monitoring network designof Choshuichi Fan.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT840015027
http://hdl.handle.net/11536/59979
Appears in Collections:Thesis