Gauss-Newton-ANN算法在滑坡位移预测中的应用  

Application of Gauss-Newton-ANN Algorithm in Prediction of Landslides Displacement

作　　者：梁万金[1] 王永[2] 

机构地区：[1]河海大学水利水电工程学院,江苏南京210098 [2]河海大学环境科学与工程学院,江苏南京210098

出　　处：《水电能源科学》2009年第4期67-69,共3页Water Resources and Power

基　　金：国家自然科学基金资助项目(50579011)

摘　　要：针对滑坡位移随时间变化具有非线性及难以预测预报的特征,基于人工神经网络的基本原理提出了基于Gauss-Newton的人工神经网络优化算法,并以清江隔河岩库区墓坪滑坡为例进行了计算。结果表明,采用该方法计算的预测位移与实测位移变化特征基本吻合,可供类似滑坡位移预测参考。Because the features of landslides displacement performing is nonlinear with time and hard to predict, ANN optimal algorithm based on Gauss-Newton is mentioned in the paper. This algorithm is based on ANN fundamental principle, and the landslides within Geheyan reservoir area is calculated as an example. The results reveal that the predicted displacements coincide well with actually observation, and the method could be referred in the similar landslide prediction.

关 键 词：Gauss-Newton算法 人工神经网络(ANN) 优化 滑坡 位移预测 

分 类 号：TU311.2[建筑科学—结构工程] TP183[自动化与计算机技术—控制理论与控制工程]