BP神经网络在深基坑监测预报中的应用与精度分析

Application and Accuracy Analysis of BP Neural Network in Monitoring and Forecasting of Deep Foundation Pit

以珠海市横琴某深基坑项目为例,考虑到影响基坑产生位移的内因和外因有很多,在不考虑基坑其他因素对监测数据预报带来的影响的情况下,结合BP神经网络算法,提出了基于BP神经网络的预报方法研究,从基坑监测数据本身出发探寻相应的变化规律。使用此方法对所获得的基坑变形数据进行线性内插后建模,并对比分析线性内插前后的预报精度。结果表明, BP神经网络用于深基坑变形监测分析与预报是可行的,并能为后续施工提供有效的参考。

The paper takes a deep foundation pit project of Hengqin in Zhuhai City as an example. As we all known, there have many the internal and external factors which affecting the displacement of thefoundation pit, the paper without considering the influence of other factors of foundation pit on monitoring data forecasting, by combining with BP neural network algorithm, the research of forecasting method based on BP neural network is proposed. The law of corresponding changes is found from the foundation pit monitoring data itself.Using this method, the obtained deformation data of the foundation pit is linearly interpolated and modeled, and the prediction accuracy before and after linear interpolation is compared and analyzed. The results show that BP neural network is feasible for the analysis and prediction of deep foundation pit deformation monitoring, and can provide an effective reference for subsequent construction.