论文部分内容阅读
深基坑开挖必然引起地表沉降,地表沉降监测数据不可避免要受到施工及周边环境的干扰,使沉降数据真实性受到极大的影响。以武汉深基坑工程的大量监测数据为基础,提出一种小波分析法与径向基神经网络的混合建模方法,对深基坑地表变形进行沉降预测分析。首先运用小波分析对实测数据进行去噪处理,提取反映实际变化的沉降数据作为径向基神经网络输入的特征向量,构建小波网络W-RBF预测模型,采用滚动预测方法对地表沉降进行预测。工程应用结果表明,W-RBF模型预测性能,要优于带有噪声构造的原始数据预测结果,具有较高的预测精度,可满足深基坑工程的信息化施工要求。
Deep foundation pit excavation will inevitably cause surface subsidence, surface subsidence monitoring data will inevitably be affected by the construction and the surrounding environment, so that the authenticity of the subsidence data will be greatly affected. Based on a large amount of monitoring data of deep foundation pit engineering in Wuhan, a hybrid modeling method based on wavelet analysis and RBF neural network is proposed to predict and analyze the surface deformation of deep foundation pit. Firstly, wavelet analysis is used to denoise the measured data and the subsidence data reflecting the actual changes are extracted as the eigenvectors input by RBF neural network. The W-RBF prediction model of wavelet neural network is constructed, and the rolling forecast method is used to predict the ground subsidence. The results of engineering application show that the prediction performance of W-RBF model is better than that of original data with noise structure, and it has high prediction accuracy and can meet the information construction requirements of deep foundation pit engineering.