摘要
采用Matlab基于Bayesian Regularization算法建立动态神经网络模型,对基坑支护桩顶面位移进行变形分析。通过实测数据对比分析:利用动态神经网络对基坑支护桩水平位移监测数据进行预测,预测精度较高,方法可行。
The dynamic neural network model is established by Matlab based on Bayesian Regularization algorithm to analyze the displacement of the top surface of foundation pit supporting pile. Through the comparative analysis of the measured data: the dynamic neural network is used to predict the horizontal displacement monitoring data of foundation pit supporting pile. The prediction accuracy is high and the method is feasible.
作者
程功
Cheng Gong(Surveying Institute,Kunming Metallurgy College,Kunming,Yunnan 650000)
出处
《江西建材》
2021年第2期44-45,共2页
Jiangxi Building Materials
关键词
动态神经网络
时间序列
位移监测
预测
Dynamic neural network
Time series
Displacement monitoring
Forecast