火灾后新老砼粘结加固劈拉性能的神经网络模拟  

Simulation to Splitting Tensile Strength of New-old Concrete Bonding Reinforcement After Fire with Neural Network

作　　者：郭进军[1] 张雷顺[1] 李平先[1] 杨光煜[2] 

机构地区：[1]郑州大学环境与水利学院,河南郑州450002 [2]天津财经学院信息系,天津300222

出　　处：《郑州大学学报（工学版）》2003年第4期55-58,共4页Journal of Zhengzhou University（Engineering Science）

基　　金：国家自然科学基金资助项目(59778045)

摘　　要：遭受火灾的混凝土建筑物采用混凝土修补加固后的性能因受许多不定量因素的影响而变得非常复杂.在试验的基础上,采用人工神经网络对新老混凝土粘结面的劈拉强度进行了模拟和预报,程序计算结果与试验结果吻合很好,进而对不同因素组合下的粘结强度进行了预报.结果表明:采用人工神经网络方法对该课题的模拟预报是准确有效的,得到了多种因素对新老混凝土粘结劈拉强度的影响规律,以利于对试验结果的补充与分析.The property of the strengthened concrete buildings after fire with concrete material is very complex due to many non-quantitative factors' influences. Based on the experiment, the paper makes use of artificial neural network to simulate and predict splitting tensile strength of new-old concrete bonding interface.The calculation results of the program are in good agreement with the test. Furthermore, the bonding strengths under some different factors combination are predicted. The results indicate that the neural network method is accurate and effective on simulation and prediction. By neural network, some influencing regularities of several factors on new-old concrete bonding splitting tensile strength are obtained, which is helpful to supplement and analyze the test results.

关 键 词：火灾 混凝土 人工神经网络 劈拉性能 粘结强度 

分 类 号：TU528.0[建筑科学—建筑技术科学] TP183[自动化与计算机技术—控制理论与控制工程]