灰色-BP神经网络在深基坑变形预测中的应用研究  

Study on Application of Gray-BP Neural Network to Deformation Prediction of Deep Foundation Pit

作　　者：黄永红[1] HUANG Yonghong(Chongqing Survey Institute, Chongqing 400020, Chin)

机构地区：[1]重庆市勘测院,重庆400020

出　　处：《四川理工学院学报（自然科学版）》2016年第5期18-22,共5页Journal of Sichuan University of Science & Engineering(Natural Science Edition)

摘　　要：随着城市化进程的推进,向地下空间发展将是趋势,由此带来大量深基坑工程。深基坑的稳定问题不仅关系上部结构的安全,而且还会影响周边建筑物的变形,因此深基坑变形研究越来越受到重视。将改进灰色预测模型与BP神经网络组合模型结合,通过在南京某大型深基坑的变形模拟和预测效果对比,改进后的灰色模型在一定程度上提高了拟合精度与预测精度,而灰色-BP组合模型在此基础上通过MATLAB平台又进一步地提高了拟合与预测的精度。该研究可为工程施工提供较好的指导意义,并为同类工程提供借鉴。With the advance of the urbanization process, more and more underground space was developed, which brings a lot of deep foundation pit engineering. The stability of deep foundation pit is not only related to the safety of the upper structure, but also affects the deformation of the surrounding buildings, so the research of the deformation of deep foun- dation pit is more and more important. The combination of grey prediction model and BP neural network combination model is improved, by comparing the deformation simulation and prediction effect of a large deep foundation pit in Nanjing, it is shown that the improved grey model improves the precision of fitting and forecasting in a certain extent, based on which the grey-BP combination model through MATLAB platform further improves the precision of fitting and prediction. This research can pro- vide a good guide for engineering construction, and provide reference for similar projects.

关 键 词：灰色-BP神经网络 深基坑 变形预测 

分 类 号：TU473[建筑科学—结构工程]