边坡稳定性神经网络评价模型中断裂构造的定量化处理(摘要).doc

边坡稳定性神经网络评价模型中断裂构造的定量化处理 祝传兵1，徐世光（1.云南省地质环境监测，昆明 650216；2. 云南，昆明 65001 3.云南省地震局形变测量中心，昆明 650041；） Kv应采用实测值，当无条件取得实测值时，可用岩体体积节理数（Jv）按表2确定对应的Kv值。 表1 Kv与岩体完整程度的对应关系 Table 1 the relation of Kv and the coefficient of rock integrity Kv 0.75 0.75~0.55 0.55~0.35 0.35~0.15 0.15 完整程度 完整 较完整 较破碎 破 碎 极破碎 Table 2 Contrast with Jv and Kv Jv（条/m3） 3 3~10 10~20 20~35 35 Kv 0.75 0.75~0.55 0.55~0.35 0.35~0.15 0.15 岩体体积节理数（Jv）可以很好的反映岩体的完整性指数，对Jv与KvY＝0.000335x2-0.0313x+0.837Kv值。 选用云南元江~磨黑高速公路50个典型边坡的工程实例随机取40个边坡样本作为网络模型的学习样本对神经网络进行训练，而将余下的10个边坡作为评价样本。作者：祝传兵（1978，男，工程师，硕士。主要从事地质环境与地质灾害防治等方面的工作。 单位：云南省地质环境监测院 电话：0871-（办）手机） 通讯地质：昆明市王大桥云南省地质环境监测院 邮编：650216 E-mail: Quantification of Fault influence for Analysis of Slope Stability based on Artificial Neural Networks ZHU Chuanbing1, XU Shiguang2, WANG Lingli3 1 Yunnan Monitoring Institute of Geological Environment, Kunming 650216, 2 Yunnan Geologic and Mineral Bureau of Exploration Exploitation, Kunming 650051, 3 Center of The Deformation Survey of Yunnan Seismological Bureau, Kunming 650041 Abstract: Although Analysis of slope stability based on artificial neural networks (ANN) has gained some significant results, it is basically limited in the stage of theory. So far it has not obtained a quite unified understanding about the selection and quantification of factors. Fault influence is ignored in most cases for it is hard to evaluate quantitatively. In this paper, Splintering Degree of Rock was taken into account in the ANN stability evaluation model of slope to evaluate the Fault influence, has definite geological basis. It is found that this method can solve the question of quantitative analysis for Fault influence and improve the research of slope stability based on ANN. Key Words: Slope Stability Artificial Neural Network Splintering Degree of Rock Fault 窗体顶端 窗体底端 - 1 - 图1 Jv与Kv相关性回归分析图 Fig.1 the result of regression analysis with Jv and Kv