地铁保护区建筑物沉降预测方法研究

Study on Prediction Method of Building Settlement in Metro Protection Area

为准确掌握地铁保护区内建、构筑物变形趋势,保障建、构筑物安全,本文充分发挥经验小波变换(EWT)与Elman神经网络模型在信号分解、数据预测中的优势,构建新的EWT-Elman组合预测模型。该组合模型实现地铁保护区内建、构筑物沉降预测的主要途径为:首先,使用EWT方法对建、构筑物沉降变形序列进行自适应分解得到不同分量;其次,使用Elman神经网络模型对不同分量进行预测;最后,重构不同分量预测值得到最终预测结果。使用两组地铁保护区内建筑物实测沉降变形数据进行实验,结果表明,相较于单一的Elman神经网络模型,本文提出的组合预测模型的精度、适应性更高,为相关类沉降预测提供了有效借鉴。

In order to accurately grasp the deformation trend of buildings and structures in the metro protection area and ensure the safety of buildings and structures,this paper gives full play to the advantages of empirical wavelet transform(EWT)and Elman neural network model in signal decomposition and data prediction,and constructs a new EWT-Elman combined prediction model.The main ways to realize the settlement prediction of buildings and structures in the metro protection area by the combined model are as follows:first,the EWT method is used to adaptively decompose the settlement deformation sequence of buildings and structures to obtain different components;secondly,Elman neural network model is used to predict different components;finally,the final prediction result is obtained by reconstructing the prediction values of different components.Two groups of measured settlement and deformation data of buildings in the metro protection area are used for experiments.The results show that the combined prediction model proposed in this paper has higher accuracy and adaptability than the individual Elman neural network model,which provides an effective reference for the settlement prediction of related types.