摘要
本文基坑变形预测共包含了两个过程,即一次非线性预测和二次非线性预测。其中,一次非线性预测是利用多种回归模型对基坑的变形进行回归预测,探讨不同回归模型的预测效果,并选取较优的回归结果进行组合预测;二次非线性预测是利用混沌RBF神经网络对组合预测的搜索误差序列进行二次预测,进一步减少预测误差,提高预测精度。结果表明:本文的预测精度较高,该方法在基坑变形预测中具有较高的有效性和可行性。
In this paper, the deformation prediction of foundation excavation contains two processes, namely the first time nonlinear and the two nonlinear prediction. Among them, a nonlinear prediction is used to predict the deformation of foundation excavation by using multiple regression models, and to explore the effect of different regression models and select the best regression results to predict by the combination forecasting. The two nonlinear prediction is used to predict the error sequence of combination forecasting by using chaotic RBF neural network, and to further reduce the prediction error and improve the prediction accuracy.
作者
马还援
杨振兴
王显彪
陈飞飞
Ma Huanyuan Yang Zhenxing Wang Xianbiao Chert Feifei(Hydrogeological and Geothermal Geological Key Laboratory of Qinghai Province （Hydro Geology and Engineering Geology and Enviromental Geology Survey Institute of Qinghai Province）, Xining, Qinghai 810008, China)
出处
《施工技术》
CAS
北大核心
2016年第19期73-77,共5页
Construction Technology
关键词
基坑
回归模型
组合预测
混沌神经网络
研究
foundation excavation
regression model
combination forecasting
chaotic neuralnetwork
research
