论文部分内容阅读
提出了一种基于人工神经网络(ANN)技术的加筋挡墙设计高度预测方法。通过分析挡墙失效的原因,确定了7个主要因素作为网络的输入神经元。收集23组挡墙离心模型试验数据,2组足尺试验数据,1组实际工程的破坏数据,共26组样本作为训练及检验样本,建立了可用于加筋挡墙设计高度预测的径向基函数网络(RBFN)及误差反传网络(BPN)模型。结果表明径向基函数网络在学习速度,预测准确性及网络推广能力方面均优于BP网络,本文方法可用于加筋支挡结构的设计参考。
A new height prediction method for reinforced retaining wall based on artificial neural network (ANN) is proposed. By analyzing the causes of failure of the retaining wall, seven main factors were identified as input neurons of the network. Twenty-three groups of centrifugal wall model test data, two sets of full-scale test data and one set of actual engineering damage data were collected. A total of 26 samples were used as training and testing samples to establish a highly predictable radial basis for reinforced retaining wall design Function Network (RBFN) and Error Back Propagation Network (BPN) model. The results show that the radial basis function network is superior to BP network in terms of learning speed, prediction accuracy and network promotion ability. The proposed method can be used as a reference for the design of stiffened support structures.