桂林理工大学学报2017，Vol.37Issue(4)：641-646,6.DOI:10.3969/j.issn.1674-9057.2017.04.014
基于EMD-PSO-BP网络模型的大坝变形预测
Dam deformation forecast based on EMD-PSO-BP neural network model
摘要
Abstract
Dam deformations have lots of characteristics such as non-stationary,non-linear.In this paper,we take Empirical Mode Decomposition (EMD) and Particle Swarm Optimization (PSO) into BP neural network,and establish EMD-PSO-BP model to forecast dam deformative values.In this model,EMD is used to decompose complex deformation data into finite and relatively stable components,and to predict each component by BP neural network optimized by Particle Swarm Optimization model,then to stack and reconstruct components as the final results.The experimental results show that the EMD-PSO-BP model has a better nonlinear mapping ability,learning ability and adaptive ability which can improve the predicted accuracy of deformations effectively.Also,the accuracy is obviously better than BP neural network model,a little better than PSO-BP model.关键词
变形预测/EMD/PSO/BP神经网络Key words
deformation prediction/EMD/PSO/BP neural network分类
天文与地球科学引用本文复制引用
秦旭元,刘立龙,陈军,陈发德,黄良珂,谢劭峰..基于EMD-PSO-BP网络模型的大坝变形预测[J].桂林理工大学学报,2017,37(4):641-646,6.基金项目
国家自然科学基金项目(41541032) （41541032）
广西自然科学基金项目(2015GXNSFAA139230) （2015GXNSFAA139230）
广西空间信息与测绘重点实验室项目(14-045-24-03 （）
14-045-24-10) （）
广西“八桂学者”岗位专项经费项目 （）
研究生教育创新计划项目(YCSZ2015163) （YCSZ2015163）
