优化灰色BP神经网络组合模型在基坑沉降监测中的应用

Application of optimized grey BP neural network combination model in foundation pit settlement monitoring

基坑沉降监测是基坑工程施工中的一个重要环节.为准确预测基坑沉降,采用优化灰色GOM(1,1)模型与BP神经网络组合预测.GOM(1,1)模型是在GM(1,1)模型原理上对其参数进行优化,使拟合值接近观测值,BP神经网络具有较强的非线性数据处理能力,优化组合模型结合两者优点应用于基坑沉降监测,得出优化后的组合模型对基坑沉降数据预测优于普通组合模型.

Foudnation pit settlement monitoring is an important link in foudnation pit construction.In order to accurately predict the foundation pit settlement,the optimized grey GOM(1,1)model and BP network are combined to predict.GOM(1,1)model is based on the principle of GM(1,1)model to optimize its parameters,so that the fitting value is close to the observed value,BP neural network has strong nonlinear date processing ability,the optimized combination model combined with the advantages of the two,applied to the foundation pit settlement monitoring,it is concluded that the optimized combination model is better than the commmon combination model for foundation pit settlement number measurement.