神经网络在高强粉煤灰混凝土强度预测及优化设计中的应用  

Application of Neural Networks in Strength Forecast and Optimal Design of High Strength fly Ash Concrete

作　　者：胡明玉[1,2] 唐明述[1] 

机构地区：[1]南京化工大学材料学院 [2]南昌大学土木学院 江西 南昌 330029

出　　处：《混凝土》2001年第1期13-17,共5页Concrete

摘　　要：首先介绍了神经网络应用中使用最为广泛的BP网络和RBF网络的模型及其学习算法，然后将其用于高强粉煤灰混凝土的强度预测和优化设计，并与线性回归进行了对比，结果表明神经网络方法具有较高的预测精度，在混凝土性能预测和优化设计中具有广阔的应用前景。In this paper,the models of BP(Error Back propa gation)and RBF(Radial Basis Function)networks,which are widely applied,and their learning algorithms are recommended.Then the two approaches are applied to stre ngth forecast and optimal design of high strength fly ash concrete.Furthermore,w e contrast them to the linear regression and the results suggest that neural net work approach have merit of high forecasting accuracy,so it’ll have broad pros pect of application in performance forecast and optimal design of concrete.

关 键 词：神经网络 配合比 高强粉煤灰混凝土 强度 优化设计 

分 类 号：TU528.2[建筑科学—建筑技术科学]