應用計數資料模式於公路長隧道風險分析之研究; A Study on Risk Analysis of Long Road Tunnel Using Count Data Model

NCUIR > Engineering College > Graduate Institute of Civil Engineering > Electronic Thesis & Dissertation > Item 987654321/60415

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/60415

Title:	應用計數資料模式於公路長隧道風險分析之研究;A Study on Risk Analysis of Long Road Tunnel Using Count Data Model
Authors:	李君睿;li,Chun-jui
Contributors:	土木工程學系
Keywords:	公路長隧道;卜瓦松迴歸模式;負二項迴歸模式;風險分析;Long Road Tunnel;Poission regression model;Negative Binomial regression model;Risk Analysis
Date:	2013-07-25
Issue Date:	2013-08-22 11:36:45 (UTC+8)
Publisher:	國立中央大學
Abstract:	臺灣為地狹人稠之地形，且山地面積占全島三分之二以上，因此需以隧道形式建設公路路網，但隧道具有封閉性、不可及性及救援困難等特性，一旦發生事故往往造成嚴重之人身財產損失。目前隨著國內八卦山隧道及雪山隧道的開通，公路長隧道之議題逐漸受到重視，因此隧道安全之相關研究亦成為近期極重要之課題。本研究以隧道安全性為研究主軸，參考世界各先進國家相關技術及研究，提出一套公路長隧道風險分析程序以評估公路長隧道之安全性，研究首先應用計數資料模式中之卜瓦松迴歸模式及負二項迴歸模式，針對國內、外隧道構建死亡人數及事故次數模式，但由於國內隧道死亡人數資料樣本數不足，故僅構建事故次數模式，再根據模式找出最適模式及相關顯著影響因素進行探討，研究結果顯示主要影響國內隧道事故次數之因素為車輛數、隧道長度及速限等三項變數；接後利用風險分析之概念，結合第一部分迴歸分析之結果，透過情境設定計算各隧道於不同情境之風險值，分析比較後提出相關交通控管措施，以降低隧道風險，提高隧道安全性。 Narrow terrain but densely populated, Taiwan island owns more than two-thirds mountainous area. Therefore, when facing the road network design, the form of tunnel is bound to be considered. Along with the opening of the Baguashan tunnel and the Hsuehshan Tunnel, 「Long road tunnel」 is gradually taken seriously. With features of closure, unreachable, hard-to-rescue and so on, any event of an accident happened in tunnel will certainly result in serious personal and property losses. Hence the issue of the tunnel safety is getting emphasized recently. Based on tunnel safety, with related skills and researches of advanced country as reference, this study proposed a risk analysis procedure to assess the safety of long road tunnel. Poission regression of count data model and Negative Binomial regression model were applied to construct the model of number of fatalities and number of accident respectively in domestic and international tunnels. Due to the insufficient samples of fatalities, only the model of the number of accident was built. According to the discussion of the optimization model and related significant factors, the result shows that three variables include the number of vehicle, tunnel length and the speed limit mainly affect the number of accident. Then, the concept of risk analysis was combined with the regression analysis to calculate the value-at-risk through different scenario settings. By analysis and comparison, several traffic control measures can be taken to reduce the risk and improve the safety of tunnal.
Appears in Collections:	[Graduate Institute of Civil Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	819	View/Open

社群 sharing

Loading...