基于ANN的深基坑变形预测方法研究(新修改)

城市地区深基坑工程愈来愈多，为了确保深基坑施工安全，对基坑进行监测和预测，实现信息化施工证明是有效的方法。国内外学者以往的研究大多侧重于基坑施工地面沉降预测方面，有关深基坑围护结构桩体整体水平位移变形的预测建模鲜有报道。结合某深基坑工程，以桩体水平位移实际监测数据为样本，建立BP神经网络时间窗口预测模型，采用MATLAB平台编写程序，预测围护结构桩体水平位移，预测值同监测值和设计计算值吻合，表明了该预测方法的可行性，研究结果可供同类工程项目参考借鉴。

Research of the Deformation Prediction in deep foundation pit construction

There are more and more project for deep foundation pit in current urban areas. In order to guarantee the security of deep foundation pit and take implementation of informationalized construction,the deformation monitoring and prediction of deep foundation pit is necessary and was proven to be an effective method. The former research of domestic and foreign scholars mostly focus on the prediction of ground settlement for deep foundation pit excavation,while the prediction of level displacement of the whole supporting structure for deep foundation pit were few reported.Aimed at specific examples of project and taking the primitive data of the level displacement of supporting structure in deep foundation pit as samples, this thesis take the priority of time series prediction model based on BP neural network using the MATLAB language to write program,in which the level displacement of the whole supporting structure is predicted. In the model，the predicted result is much close to the measured data and designed control values which proved that the model is feasible. The studies presented in the thesis will provide valuable references for similar engineering.