- 學位論文
結合空間資訊與統計推論之隧道地質剖面機率式預測方法
Probabilistic prediction of tunnel geological profiles by applying geographic information system and statistic inference
摘要
地質剖面為隧道地質調查整合既有地質資料以及地表地質調查、地球物理探查與鑽孔探查等結果的具體展現，亦為隧道工程設計與施工的關鍵資訊。目前隧道地質剖面多係地質師主觀判斷測製而得，結果受到個人經驗影響甚大，且多以定率型式展現，雖可提供設計施工基本所需，卻無法展現隧道沿線不同施工風險的程度，難以判斷前進探查方法實施的時機，亦不利於風險管理與分擔、施工預警與損害防阻等措施的推動。 本文結合空間資訊系統與統計推論途徑，自地質調查資料中求得各類資料的分佈型式與對應的特性參數，提出一隧道地質剖面機率式預測方法，並應用ESRI ArcScene軟體編成人性化界面程式。主要的程序包括：(1)將數值地形模型(Digital Terrain Model, DTM)資料轉換為三角化不規則網格(Triangulated Irregular Network)，建構高精度地形立體模型；(2)透過ArcScene軟體內建VBA(visual basic for application)開發地質調查資料統計模組，估計待推估資料的分佈型式與對應的特性參數；(3)依據現地調查與統計推估所得參數，預測空間中分佈特性與機率；(4)透過繪圖功能輸出隧道地質剖面所需地質資料與機率值。本文將介紹所提方法，並以台灣北部一隧道實例，說明所提方法的應用性。
並列摘要
In a tunnel geology survey, a geology section map is a substantial expression of an integration that incorporates existing geological data with results of surface geology survey, geophysics exploration and borehole survey, key information to tunnel engineering and construction. Currently, tunnel geology section maps are mostly surveyed and drafted based on objective judgments of geologists, results of which are affected by personal experience and are expressed in certain forms of ratios that may have provided basic necessities for the design and construction tasks but are insufficient to manifest the extent of different risks along the tunnel route. It is therefore difficult to determine the timing for applying an advanced survey accordingly, and it adversely affects the implementation of measures for risk management and sharing, constructional pre-alarm and damage mitigation. This paper combines spatial data system (Geomatics) with statistical reduction methods for finding the distribution pattern of versatile data and their corresponding parameters, so as to derive predictive probabilities of tunnel geology sections via geology survey data and compose a user-friendly interface program using the ESRI ArcScene software. The main procedures comprise: (1) Converting Digital Terrain Model, DTM into Triangulated Irregular Network for constructing a high-precision 3D terrain model; (2) using the built-in VBA (visual basic for application) of the ArcScene software, develop a statistical module of geology survey data to estimate the distribution pattern and probability of the examined data; (3) predicting the characteristic and probability of the spatial distribution based on parameters attained from the site survey and the statistical estimation; (4) outputting geological data and probability required for drafting the tunnel geology section map via the plot function. The paper introduces all these methods and explains their applications using an example of a tunnel in northern Taiwan.