摘要
在MTS电液饲服万能试验机上对64个哑铃形试件在应变速率10^-5～10^-0.3/s内进行单轴直接拉伸试验,研究了混凝土强度、含水量、应变速率对混凝土直接拉伸特性的影响.结果表明：随着应变速率的提高,混凝土动态拉伸强度明显增加；含水量较高的混凝土其拉伸强度提高更显著.针对混凝土动态特性影响因素众多的特点,提出利用人工神经网络方法来综合反映多种因素对混凝土动态特性的影响作用.与传统动态试验数据分析方法比较表明：利用人工神经网络来分析研究多种因素影响的混凝土动态特性,是一种简便、适用的新方法.
Direct uniaxial tensile experiments with 64 dumbbell shaped specimens were conducted on the MTS810 servo-hydraulic universal machine at strain rate 10^-5～10^-0.3/s. The effects of material strength, water content, and strain rate on the dynamic properties of concrete were investigated systenucaUy. The results snow that tensile strength rises with the increasing strain rate and this trend is more notable for the saturated concrete. On the basis of analysis of test results, artificial neural network approach is suggested to be employed to treat the mulitfactor dynamic properties of concrete. In comparison with the conventional data analysis method, it is found that this approach is simple and effective to handle such complex dynamic properties of concrete.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2006年第4期622-625,643,共5页
Journal of Harbin Institute of Technology
基金
国家自然科学基金资助项目(50139010)
关键词
应变率
单轴
拉伸强度
神经网络
混凝土
strain rate
uniaxial
tensile strength
neural network
concrete
