基于遗传-神经网络的深基坑变形实时预报方法研究  

REAL TIME PREDICTION METHOD BASED ON GENETIC ALGORITHM AND NEURAL NETWORK FOR DEFORMATION CAUSED BY DEEP EXCAVATION

作　　者：刘勇健[1] 李彰明[1] 张建龙[1] 易又庆[1] 

机构地区：[1]广东工业大学岩土工程研究所,广州510643

出　　处：《岩石力学与工程学报》2004年第6期1010-1014,共5页Chinese Journal of Rock Mechanics and Engineering

基　　金：广东省自然科学基金(20010055);广东工业大学重点学科基金(20216)

摘　　要：根据深基坑变形的基本特征,运用遗传-神经网络建立了深基坑变形的实时预报模型,并对刘屋洲泵站深基坑变形进行了建模预报,预报值与实测值基本吻合,预测精度高,证明该方法用于深基坑变形实时预报的有效性和实用性。Considering the basic characteristics of rock deformation caused by deep excavations, a real-time modeling predication method is established, based on genetic algorithm and neural network, in this paper. The predicted values agree well with the measured ones when this method is applied to the deep foundation pit project of the Liuwuzhou pumping station. The results indicate high prediction accuracy and the proposed method is a valid and useful tool for deformation prediction.

关 键 词：土力学 深基坑 遗传-神经网络 时间序列 变形 实时预报 

分 类 号：TU473.2[建筑科学—结构工程]