基于人工神经网络的混凝土实时强度影响因素敏感性分析  

Sensitivity analysis on influence factors to concrete real-time strength based on artificial neural network

作　　者：杨晓明[1] 吴天宇[1] 时丹[1] 

机构地区：[1]辽宁工程技术大学建筑工程学院,辽宁阜新123000

出　　处：《混凝土》2014年第11期15-18,23,共5页Concrete

基　　金：国家自然科学基金资助(51008148);辽宁省教育厅年杰出青年学者成长计划资助(LJQ2013037)

摘　　要：准确地预测出混凝土材料在使用过程的实时强度对于正确评估结构安全性有着重要的意义。影响混凝土材料实时强度的主要因素包括龄期、环境类别、水灰比、胶凝材料用量等等。采用人工神经网络进行混凝土实时强度影响因素敏感性分析。首先,对影响混凝土实时强度的各类因素进行分析,确定敏感性因素。其次,针对龄期敏感性因素,建立两个神经网络,一个神经网络的输入变量包含龄期,另一个不包含龄期,将训练好的两个神经网络用同组数据进行测试,比较两组测试结果,以此来确定龄期因素对混凝土强度的敏感性。采用上述方法分别对环境类别、水灰比、胶凝材料用量等因素进行敏感性分析。最后,通过比较确定混凝土龄期、环境类别、水灰比为影响混凝土实时强度的敏感性因素。It is significant for correct condition assessment of concrete structure to accurately predict the real-time strength of concrete during its service life.A lot of factors, such as age, environment type, water cement ratio, amount of cementitious material and so on, affect the real-time strength of concrete.The artificial neural network （ANN）was used to carry out the sensitivity analysis of influence factors to concrete real-time strength.First, all kinds of influence factors to concrete real-time strength were analyzed and the possible sensitive fac- tors were obtained.Second, two ANNs with and without age as one of input variables were built, trained and tested.The sensitivity of age to concrete real-time strength was found according to the change of test results of two ANNs.The sensitivities of other factors were obtained using the above method.Finally, it was found that the age, environment type and water cement ratio were sensitivity factors to concrete re- al-time strength.

关 键 词：混凝土 实时强度 人工神经网络 敏感性分析 预测 

分 类 号：TU528.01[建筑科学—建筑技术科学]