| 基于EMD-NAR神经网络的大坝变形预测 |
| |
| 引用本文： | 杨诚,王维钰.基于EMD-NAR神经网络的大坝变形预测[J].北京测绘,2020(3):386-390. |
| |
| 作者姓名： | 杨诚 王维钰 |
| |
| 作者单位： | 海南地质综合勘察设计院;海南钰成测绘科技有限公司 |
| |
| 摘 要： | 为了使大坝变形的预测精度更高,针对大坝形变量的时间序列中存在着非平稳和非线性等曲线特性,使用一种经验模态分解(EMD)和非线性自回归动态神经网络(NAR)相结合的EMD-NAR模型对大坝形变时间序列进行预测。以某大坝实测的时间序列数据为算例,分别使用BP模型、NAR模型和EMD-NAR模型进行实验对比,结果表明,BP、NAR、EMD-NAR模型预测的均方根误差(RMSE)分别为0.9449,0.6993,0.4678;模型预测的平均相对误差(MRE)分别为0.1492,0.1065和0.0688,从三种模型预测结果对比可知,组合的EMD-NAR模型预测精度最高且稳定性最好,为时间序列的大坝形变预测提供一种新的参考思路。
|
| 关 键 词： | 大坝变形 经验模态分解(EMD) 非线性自回归(NAR) 神经网络 时间序列 |
Dam Deformation Prediction Based on EMD-NAR Neural Network |
| |
| YANG Cheng,WANG Weiyu.Dam Deformation Prediction Based on EMD-NAR Neural Network[J].Beijing Surveying and Mapping,2020(3):386-390. |
| |
| Authors: | YANG Cheng WANG Weiyu |
| |
| Institution: | (Hainan Geological Survey and Design Institute, Haikou Hainan 571100, China;Hainan Yucheng Surveying and Mapping Technology Company Limited, Haikou Hainan 571100, China) |
| |
| Abstract: | In order to improve the prediction accuracy of dam deformation,there are non-stationary and nonlinear curve characteristics in the time series of dam-shaped variables,using an empirical mode decomposition and nonlinear autoregressive dynamic neural network(EMD-NAR)model predicts dam deformation time series.Taking the time series data measured by a dam as an example,using BP,NAR and EMD-NAR models predict the dam data respectively,The results show that the root mean square errors predicted by BP,NAR and EMD-NAR models are 0.9449,0.6993,0.4678;the average relative error is 0.1492,0.1065 and 0.0688,respectively.The comparison of the three models shows that the combined EMD-NAR model has the highest prediction accuracy and the best stability,which provides a new reference for time series dam deformation prediction ideas. |
| |
| Keywords: | dam deformation emperical mode decomposition(EMD) nonlinear autoregressive(NAR) neural network timeseries |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
