摘要
建立了以经验模态分解法(EMD)和果蝇算法(FOA)优化BP神经网络为基础的EMD-FOA-BP大坝变形预测模型,该模型首先利用EMD将大坝变形序列分解成相对平稳的分量,再根据各分量的特点构造不同FOA-BP模型并进行预测,叠加各分量预测值得到最终预测结果。结果表明,EMD-FOA-BP模型的自适应能力、学习能力及非线性映射能力较强,在大坝变形预测应用中能有效提高精度,预测精度较FOA-BP模型有所提高,且明显优于BP神经网络模型和GA-BP模型。
An EMD-FOA-BP dam deformation prediction model is established based on empirical mode decomposition(EMD)and BP neural network modified by fruit fly optimization algorithm( FOA). The EMD is firstly used to decompose the sequenceof dam deformation into relative stable components,then different FOA-BP models are constructed in accordance with thecharacteristics of each component,and finally the predictions can be obtained by overlaying the predicted value of eachcomponent. The model calculation results indicate that the EMD-FOA-BP model is good in adaptive capacity,learning abilityand nonlinear mapping ability,which can improve the accuracy of dam deformation prediction. The prediction accuracy ofEMD-FOA-BP model is slightly improved than that of FOA-BP model,and is obviously better than those of the BP neuralnetwork model and GA-BP model.
作者
黄军胜
黄良珂
刘立龙
谢劭峰
HUANG Junsheng;HUANG Liangke;LIU Lilong;XIE Shaofeng(Guangxi Water and Power Design Institute,Nanning 530023,Guangxi,China;College of Geomatics and Geoinformation,Guilin University of Technology,Guilin 541004,Guangxi,China;Guangxi Key Laboratory of Spatial Information and Geomatics,Guilin 541004,Guangxi,China)
出处
《水力发电》
北大核心
2019年第2期106-110,共5页
Water Power
基金
国家自然科学基金资助项目(41704027)
广西自然科学基金(2017GXNSFBA198139
2015GXNSFAA139230)
关键词
大坝变形
经验模态分解
果蝇算法
BP神经网络
dam deformation
empirical mode decomposition
fruit fly optimization algorithm
BP neural network
作者简介
黄军胜(1973—),男,广西灌阳人,高级工程师,研究方向为GPS技术与水利测绘;通讯作者:黄良珂.
