基于GA-BP模型的卵石混凝土抗折强度预测  

Prediction of Flexural Strength of Pebble Concrete Based on GA⁃BP Model

作　　者：翟来铮 高宇龙 李清富[3] 张华[3] ZHAI Laizheng;GAO Yulong;LI Qingfu;ZHANG Hua(Eastern Henan Water Conservancy Administration Bureau,Kaifeng 475002,China;The First Engineering Bureau of Henan Water Conservancy,Zhengzhou 450016,China;School of Water Conservancy Engineering,Zhengzhou University,Zhengzhou 450052,China)

机构地区：[1]河南省豫东水利工程管理局,河南开封475002 [2]河南省水利第一工程局,河南郑州450016 [3]郑州大学水利科学与工程学院,河南郑州450052

出　　处：《人民黄河》2022年第11期137-141,共5页Yellow River

基　　金：河南省水利科技攻关计划项目(GG202062)。

摘　　要：为了提高卵石混凝土抗折强度的预测精度,利用已有试验结果,建立了考虑水胶比、粉煤灰掺量、粉状矿渣掺量、砂率、卵石级配、龄期等6个因素的反向传播神经网络(BPNN)模型,并使用遗传(GA)算法优化传统BPNN模型的初始权值和阈值;此外,通过均方误差MSE和拟合优度R2对2个模型的预测结果与规范公式预测结果进行了比较。结果表明,GA-BP模型具有最高的预测精度,可作为卵石混凝土抗折强度的预测工具。In order to improve the prediction accuracy of the flexural strength of pebble concrete,a back propagation neural network(BPNN)model considering the effects of six factors such as water⁃cement ratio,fly ash admixture,pulverized slag admixture,sand rate,pebble grading,and age was developed,and the initial weights and thresholds of the conventional BPNN model were optimized by using a ge⁃netic algorithm(GA).In addition,the prediction results of the 2 models were compared with the canonical formula prediction results by MSE and R2.The results show that the GA⁃BP model has the highest prediction accuracy.Therefore,this model can be used as a prediction tool for the flexural strength of pebble concrete.

关 键 词：卵石混凝土 GA-BP模型 抗折强度预测 规范公式修正 

分 类 号：TV431.9[水利工程—水工结构工程]