基于BP神经网络-实时跟踪算法的海堤沉降预测  

Prediction of seawall settlement based on BP neural network- realtime tracing algorithm model

作　　者：沈跃军 秦鹏[2,3] 

机构地区：[1]浙江省钱塘江管理局勘测设计院,浙江杭州310016 [2]浙江水利水电学院水利系,浙江杭州310018 [3]河海大学水利水电学院,江苏南京210098

出　　处：《人民长江》2013年第22期100-102,共3页Yangtze River

基　　金：浙江省教育厅科研项目(Y201223037);浙江省水利厅科研项目(RC1302;RC1311);浙江水利水电专科学校校级重点课题(xkyzd201201)

摘　　要：对海堤工程进行实时监控对保障堤防安全具有重要意义。以浙江省某海堤工程沉降数据为研究对象,建立基于MATLAB的BP神经网络-实时跟踪组合预测算法模型,并用原型观测数据对其进行了校核和检验,最后将该算法结果与灰色理论中的GM(1,1)模型结果进行比较。结果表明,用BP神经网络-实时跟踪组合算法预测海堤工程变形具有较高精度和抗噪性,为海堤工程变形沉降预测提供了新的研究思路。It is of great importance to carry out real - time monitoring on seawall for guaranteeing its operation safety. Based on BP neural network theory and Reahime Tracing Algorithm, we attempt to set up a forecasting model by MATLAB, and further the presented model is verified by monitoring data of the seawall. Through comparison of the results obtained by the presented model and the results calculated by GM （ 1,1 ） method, it is shown that the proposed model has satisfying accuracy and anti - noise ability, thus provide a new research idea for settlement prediction of seawall project.

关 键 词：BP神经网络 实时跟踪算法 沉降预测 海堤 

分 类 号：TV86[水利工程—水利水电工程]