【作者】 贾少杰；

【导师】 杨梅忠；

【作者基本信息】 西安科技大学 ， 地质工程， 2011， 硕士

【摘要】 太白县地质灾害较严重,其灾点类型多、成片成带分布,规模差异较大,影响和受控因素多,发生频率较高,小型及中型灾点多,其中危害属重大级的灾点较少;由于其复杂的地质条件、地形地貌条件、气象水文条件和强烈的新构造运动及人为工程活动,使得本区成为地质灾害多发县区之一。本文结合2009年“陕西省太白县地质灾害详细调查”任务,对太白县地质灾害详细调查及室内试验等工作的基础上,总结了区内滑坡隐患点的发育特征及分布规律;并对区内滑坡灾害点进行了危险性评价,研究的主要内容有:(1)分析了区内滑坡隐患点的类型、规模、发育特征及形成条件等;在对滑坡隐患点调查资料进行分析研究的基础上,根据区内滑坡隐患点的特征,以规模、稳定程度、威胁人数、威胁财产数四个危险性制约因子运用层次分析法构造训练样本集。(2)应用MATLAB工具箱,调用newff()函数建立神经网络,对动量BP算法与SCG算法的训练结果进行比较,确定SCG算法为训练方法。(3)应用LibSVM工具箱来实现支持向量机模型,通过对多项式核函数、径向基核函数RBF和Sigmoid核函数的样本训练结果及误差分析确定了径向基RBF为模型核函数。应用两种方法计算研究区内81处滑坡隐患点危险性指数,并对结果进行分析比较,与以往隐患点危险性定性评价结果基本一致,本研究成果对于太白县地质灾害隐患点防治及人工智能应用于地质灾害隐患点危险性评价具有指导意义。更多还原

【Abstract】 Taibai more serious geological disasters, the disaster point of many types and dist-ribution into the film into a zone, quite different scale, impact and controlled by man-y factors, high frequency, small and medium-sized disaster points, which are significant to level against disaster points less; because of its complex geological conditions, ter-rain conditions, meteorological and hydrological conditions and a strong new tectonic movement and human engineering activities, making the area into one of geological disasters counties.In this paper, 2009 "Shaanxi Taibai detailed investigation of geological disasters, " the task of Taibai geological disasters such as a detailed investigation and laboratory test based on the work, summed up the landslide hazard area and development characteristics of point distribution; and the region Point landslide risk assessment carried out to study the main contents are:(1) Analysis of the regional landslide hazard point type, size, development characteristics and formation conditions; landslide hazards in the analysis of survey data point-s based on the study, according to the characteristics of regional landslide hazard poin-ts to the scale, stability , the number of threats, threats to property, restricting the nu-mber of four risk factors AHP construct the training sample set.(2) application of MATLAB toolbox, called newff () function of neural networks, BP algorithm with momentum SCG training algorithm results were compared to deter-mine the SCG algorithm for training methods.(3) application LibSVM toolbox to implement support vector machine model, by t-he polynomial kernel function, radial basis function RBF and Sigmoid kernel function of the sample training results and error analysis to determine the radial basis kernel function RBF as a model.Application of two methods in the study area 81 point landslide hazard risk inde-x, and the results were analyzed and compared with the previous qualitative assessment of risk of hidden point was consistent, the results of this study point to Taibai risks of geological disasters prevention and artificial intelligence Applied to geological disa-sters have hidden point risk assessment guidance.更多还原