文献类型：期刊文献

中文题名：骨料品质特性对RAC弹性模量预测的影响研究

英文题名：Effect of aggregate characteristics on elastic modulus prediction of recycled aggregate concrete

作者：梁超锋[1,2];顾烽[1];段珍华[2];王星[3]

机构：[1]绍兴文理学院建筑工程系,浙江绍兴312000;[2]同济大学建筑工程系,上海200092;[3]上海市环境工程设计科学研究院有限公司,上海200232

年份：2021

期号：7

起止页码：90

中文期刊名：混凝土

外文期刊名：Concrete

收录：CSTPCD、、北大核心、北大核心2020

基金：绍兴市科技局创新计划项目(2018C30007);上海城投科技创新课题(CTKY-ZDXM-2018-011)。

语种：中文

中文关键词：再生骨料;再生混凝土;人工神经网络;骨料特性;弹性模量

外文关键词：recycled aggregate;recycled aggregate concrete;artificial neural network;aggregate characteristic;elastic modulus

中文摘要：利用人工神经网络(ANN)方法,对从5篇文献中收集的与再生混凝土性能有关的290组数据进行训练与测试,评估了再生骨料的品质特性对混凝土弹性模量预测模型的重要性,并进一步分析再生混凝土ANN预测模型中输入参数的选择问题。研究结果表明:再生骨料的品质特性参数对再生混凝土的弹性模量有较大影响,与以往只考虑配合比和取代率的ANN模型相比,含再生骨料品质特性参数的ANN模型具有更高的预测精度,相关系数(R2)都超过了0.996,且平均绝对百分比误差(MAPE)值均非常小,介于2%-5.98%。ANN可以用来预测含不同品质再生骨料混凝土的弹性模量,但需合理选择输入参数进行建模。

外文摘要：The artificial neural network(ANN)is used to train and test 290 sets of data related to the properties of recycled concrete collected from 5 literatures.The importance of the quality characteristics of recycled aggregate to the prediction model of concrete elastic modulus is evaluated,and the selection of input parameters in the ANN prediction model of recycled concrete is further analyzed.The results show that the characteristic parameters of recycled aggregate with different quality have great influence on the elastic modulus of RAC.Compared with the previous network model which only considers the mixture ratio and replacement ratio,the new model containing the characteristic parameters of recycled aggregate has higher prediction accuracy.In the prediction examples of different data sets,the correlation coefficient(R2)of the new model exceeds 0.996 and the average absolute percentage error(MAPE)value is also very small which ranges from 2%to 5.98%.ANN is suitable for prediction of elastic modulus of concrete containing recycled aggregates of different quality,but it is still necessary to reasonably select key parameters for modeling.