【摘要】：針對Elman神經(jīng)網(wǎng)絡(luò)收斂速度慢、容易陷入局部極小等問題,建立了人工蜂群算法(ABC)與Elman神經(jīng)網(wǎng)絡(luò)組合的大壩變形監(jiān)控模型。應(yīng)用于某混凝土重力壩的結(jié)果表明,單純Elman神經(jīng)網(wǎng)絡(luò)建模方法預(yù)測的相對誤差和標(biāo)準(zhǔn)差分別為3.50%和0.131,ABC-Elman(人工蜂群算法與Elman神經(jīng)網(wǎng)絡(luò))模型預(yù)測的相對誤差和標(biāo)準(zhǔn)差分別為1.98%和0.063。從各影響因子對大壩變形的貢獻(xiàn)上看,水壓分量占27.9%,溫度分量占62.3%,時效分量占9.8%。ABC-Elman模型在建模效率、預(yù)測精度等方面均有一定的優(yōu)勢,較適合于大壩變形的建模分析,并可推廣于大壩滲流、應(yīng)力等監(jiān)控模型中。
[Abstract]:In order to solve the problems of slow convergence and easy to fall into local minima of Elman neural network, a dam deformation monitoring model based on artificial bee swarm algorithm (ABC) and Elman neural network is established. The results applied to a concrete gravity dam show that the relative error and standard deviation of simple Elman neural network modeling method are 3.50% and 0.131, respectively, and the relative error and standard deviation of ABC 鈮，

本文編號：2501985