摘要
利用反向传播神经网络(BP神经网络)的模拟能力来代替传统的方法,建立灌芯砌体的本构模型,对比模拟结果,并与传统方法进行比较.从模型的预测值和试验值可以看出,训练后的模型具有较高的学习精度和较好的泛化能力,可以保证在结构问题中的应用.
The BP model masonry instead of traditional of artificial neural network is applied for the construction model of grouted block methods in the paper. Comparative study suggests that using the BP model to model the behavior of grouted block masonry results in greater study accuracy and better generalizing competence. The predicted results are in the good agreement with the measured ones.
出处
《厦门理工学院学报》
2006年第3期48-51,共4页
Journal of Xiamen University of Technology
关键词
BP神经网络
本构模型
灌芯砌体
BP Neural Networks
constitutive model
grouted block masonry
