【作者】 李娅；

【导师】 谢强；

【作者基本信息】 西南交通大学 ， 地质工程， 2004， 硕士

【摘要】 本文利用西南交通大学岩石边坡课题组二十年来所调查的全国不同地区的铁路、公路岩石边坡近200个工点资料，用范例推理的方法对结构性硬质岩石边坡的稳定性评价进行研究。本文的主要研究内容为： 1．提取结构性岩质边坡，对该类边坡通过统计分析及试算，确定影响其稳定性的关键因素为块体稳定性系数K、可动块体数N和结构面综合长度L。 2．应用基于欧氏距离、曼哈顿距离及模糊相似优先的范例推理方法对边坡稳定性进行评价。并对三种方法作以比较。 3．根据变权的思想确定边坡稳定性影响因素权重的计算方法。 4．为了对众多的边坡数据进行科学的管理以及方便计算，对DOS操作系统下的边坡数据库系统进行了WINDOWS界面的移置，并编写了基于范例推理的边坡稳定性分析程序。 通过研究，以K、N和L作为结构性岩质边坡的属性参数，利用三种范例推理方法对提取出的154个结构性岩质边坡稳定性进行回判，结果表明，基于欧氏距离的范例推理方法总符合率达85.06％，基于曼哈顿距离的范例推理方法总符合率达83.77％，基于模糊相似优先范例推理方法总符合率达85.71％。三种方法评价结果一致的符合率达82.47％。 将基于K、N和L三参数范例推理的边坡稳定性评价方法应用于内(江)—昆(明)铁路边坡稳定性分析，结果与现场调查一致，表明利用块体稳定性系数K、可动块体数N和结构面综合长度L三个参数对结构性岩石边坡应用基于范例推理的方法来进行稳定性评价是合理的，且使用方便，具有实用价值。更多还原

【Abstract】 Based on 200 slopes data collected from different areas of highway and railway in China by research group of rock slope of Southwest Jiaotong University, the stability of structured rock slope is researched with method of Case-Based reasoning. The main research of this paper is briefly mentioned as follows:1. The structured rock slope is chosen from slope database and the key stability influencing factors of this type of slopes are determined as block stability coefficient K, movable block number N and synthetical joint length L by statistic analysis and pilot calculation.2. The stability of structured rock slopes is evaluated by Case-based reasoning method based on Euclidean distance, Manhattan distance, Fuzzy analog optimum and the results are compared each other.3. Considering the sensitivity of attribute weights to the environment, the algorithm of attribute weights is set up on the basis of the concept of changeable weights.4. To manage and analyze the large numbers of slope data scientifically and conveniently, the slope database system based on DOS operation system is turned into WINDOWS interface. The slope stability analysis program based on case-based reasoning is added in the database system.Based on K, N and L as the attribute parameters of structured rock slope, the 154 rock slopes are re-evaluated with case-based reasoning. The results show that the accuracy ratios of re-evaluation based on Euclidean distance, Manhattan distance, Fuzzy analog optimum are 85.06%, 83.77% and 85.71% respectively. The accuracy ratio that the result is the same with three methods is 82.47%.The results using case-based reasoning method based on parameters K, N and L on slope stability analysis in Neijiang-Kunming railway show that this method is rational and convenient.更多还原