地震砂土液化判别的灰色关联-逐步分析耦合模型  

Coupling model based on grey relational analysis and stepwise discriminant analysis for seismic liquefaction discrimination of sandy soil

作　　者：李波[1,2] 苏经宇[1] 马东辉[1] 王威[1] 

机构地区：[1]北京工业大学抗震减灾研究所,北京100124 [2]中国建筑科学研究院地基基础研究所,北京100013

出　　处：《中南大学学报（自然科学版）》2016年第1期232-238,共7页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(51208017);国家"十二五"科技支撑计划项目(2011BAJ08B05);PICC灾害研究基金资助项目(2013D32);北京工业大学博士科研启动基金项目资助(012000543114515)~~

摘　　要：通过分析灰色关联模型和逐步判别模型在单独进行砂土液化等级判别时所体现的优点及存在的问题,建立以两者为基础的耦合判别模型。利用实际样本结果,选取震级M、地面加速度最大值g_(max)、比贯入阻力P_s、标准贯入击数N63.5、平均粒径D_(50)、相对密实度Dr、地下水位d_w共7个实测数据作为砂土液化判别因子,对该模型进行验证。研究结果表明:BP神经网络、灰色关联分析、逐步判别分析这3种方法对样本的判别准确率分别为60%,80%和60%,耦合模型的判别准确率为100%。与单独使用2种基础模型相比,耦合模型的判别结果与实际结果更加吻合,表明该方法具有较高的准确性和良好的实用性。The advantages and disadvantages of grey relational model and stepwise discriminant model were analyzed when both of them were used alone for evaluation of seismic liquefaction of sandy soil, and coupling model was established based on both models. Seven sets of factors（magnitude M, the maximum ground acceleration gmax, specific penetration resistance Ps, standard penetration blow count N63.5, average particle size of D50, relative density Dr, water table dw） were selected for model verification. The results show that the accuracy of BP neural network model, grey relational model and stepwise discriminant model are 60%, 80% and 60%, respectively. The accuracy of coupling model is 100%, it indicates that the discriminant results are more compatible with the actual results compared with one analysis method alone, and that the coupling model has high accuracy and good practicability.

关 键 词：砂土液化 灰色关联分析 逐步判别分析 耦合判别模型 

分 类 号：TU441[建筑科学—岩土工程]