基于遗传-神经网络无粘结部分预应力高强混凝土梁的延性  

Research on Ductility of Unbonded Partially Prestressed Concrete Beam with High Strength Based on Genetic Algorithm and Neural Network

作　　者：胡立黎[1] 郑宏[1] 

机构地区：[1]长安大学建筑工程学院,陕西西安710061

出　　处：《江南大学学报（自然科学版）》2008年第4期479-483,共5页Joural of Jiangnan University (Natural Science Edition)　

摘　　要：利用建立的遗传-神经网络模型,对无粘结部分预应力高强混凝土梁的延性进行预报,并且利用网络模型分析影响此种梁的延性因素,得到各种位移延性比曲线.结果表明:随着预应力筋配筋指标、非预应力筋配筋指标、跨高比和截面高度的增大,梁的位移延性比呈逐渐减小趋势.据此提出了计算梁位移延性比的公式.In the paper, the genetic algorithm-neural network model is used to predict ductility of unbonded partially prestressed concrete beam with high strength. Meanwhile, kinds of ingredients of the beamg ductility are analyzed,and curves between kinds of ingredients and displacement ductility ratio are obtained. The result indicates that displacement ductility ratio of the beam gradually reduces with increasing unprestressed index, prestressed index, span depth ratio, and depth of section. The formula of the ductility is proposed.

关 键 词：遗传算法 神经网络 高强混凝土梁 预应力 延性 

分 类 号：TU378.2[建筑科学—结构工程]