早期受冻损伤混凝土强度预测模型研究及其配比优化  

Research on Strength Prediction Model and Ratio Optimization of Concrete Damaged by Early Freezing

作　　者：徐存东 徐慧 陈家豪 李准 赵志宏 王海若 任子豪 XU Cun-dong;XU Hui;CHEN Jia-hao;LI Zhun;ZHAO Zhi-hong;WANG Hai-ruo;REN Zi-hao(School of Water Conservancy,North China University of Water Resources and Electric Power,Zhengzhou 450046,China;Key Laboratory for Technology in Rural Water Management of Zhejiang Province,Hangzhou 310018,China;Henan Provincial Hydraulic Structure Safety Engineering Research Center,Zhengzhou 450046,China)

机构地区：[1]华北水利水电大学水利学院,河南郑州450046 [2]浙江省农村水利水电资源配置与调控关键技术重点试验室,浙江杭州310018 [3]河南省水工结构安全工程技术研究中心,河南郑州450046

出　　处：《水电能源科学》2023年第3期144-148,共5页Water Resources and Power

基　　金：国家自然科学基金项目(51579102);河南省高校科技创新团队支持计划(19IRTSTHN030);中原科技创新领军人才支持计划(204200510048);河南省科技攻关项目(212102310273);河南省高等学校重点科研项目计划(20A570006);浙江省重点研发计划(2021C03019);平原河网多尺度水动力调控对河湖水生态影响研究(LZJWD22E090001)。

摘　　要：针对严寒地区水工混凝土建筑物在浇筑施工时易受早期冻害,并可能影响后期健康服役的问题,为研究早冻混凝土强度损伤规律并对其配比进行优化,利用RSM响应面法Box-Behnken(RSM-BBD)优化试验设计,并建立以水胶比、粉煤灰掺量、引气剂掺量为变量的RSM响应面模型;同时,构建一种GA-BPNN强度预测模型,实现早冻混凝土强度的精确预测。结果表明,与RSM模型相比,GA-BPNN模型有更精确的预测性能,且能更高效地优化配比设计;GA-BPNN强度预测模型的拟合优度R^(2)和平均相对误差eMRE分别为0.998 5、2.13%,最优配比强度预测值与试验值的相对误差约为1%,应用该模型可实现混凝土受冻强度推演及其配比的高效优化。In view of the problem that hydraulic concrete buildings in cold regions are vulnerable to early freezing injury during pouring construction and may affect the later healthy service, in order to study the strength damage law of early frozen concrete and optimize its ratio, Box-Behnken(RSM-BBD) response surface method was used to optimize the experimental design. The RSM response surface model was established by taking water binder ratio, fly ash content and air entraining agent content as variables. A GA-BPNN strength prediction model was constructed to predict the strength of early-frozen concrete accurately. Compared with the RSM model, the results show that the GA-BPNN has more accurate prediction performance and can optimize proportion design more efficiently. The goodness of fit R~2 and average relative error eMREby the GA-BPNN strength prediction model are 0.998 5 and 2.13%, respectively. The relative error between the predicted value of the optimal strength ratio and the experimental value is about 1%. The application of GA-BPNN strength prediction model can realize the efficient optimization of concrete freezing strength and its ratio.

关 键 词：响应面法 遗传优化 BP神经网络 配比优化 早冻混凝土 

分 类 号：TV431[水利工程—水工结构工程]