基于灰色BP神经网络组合模型的深基坑周围地表沉降预测研究

Research on Deep Foundation Pit Surrounding Ground Settlement Prediction Based on Coupling Model of Gray-Neural Network

在深基坑工程施工过程中,基坑周围地表的沉降对周围建筑物、地下管线和支护体系的安全都会造成很大的影响,如何利用对影响基坑周围地表沉降的关联因素的研究并结合基坑周围地表沉降的监测数据来分析得到基坑周围地表沉降量的走势,已成为城市建设中的一个重要的安全课题.以天津市某换乘车站为例,分析影响该基坑周围地表沉降的关联因素,建立灰色预测和BP神经网络组合模型,在"小样本、贫信息"的情况下,得到的预测结果与实际监测值吻合度较高.利用该预测模型可对一些在开挖过程中监测天数相对较少的深基坑工程进行可靠而准确的预测.

During the construction of deep foundation pit,ground settlement around the deep foundation has big influenceon the building surrounding the pit or underground pipeline and supporting structure system.How to make use of the associatedfactors which affect the ground settlement around the foundation pit,and combined with the ground settlementaround the foundation pit monitoring data to get the amount of the ground settlement around the foundation pit,has becomean important safety issue in city development.This article,taking a transfer station of Tianjin as an example,analyzes thecorrelation factors that influence the ground settlement around the foundation pit,and establishment the grey forecast andBP neural network combination model.In the "limited sample,limited information "case,the prediction results is highertally with the actual monitoring value degree.Finally,this model provides a reliable and accurate prediction method formonitoring the excavation of deep foundation pit engineering within fewer days.