基于GA-BP神经网络的SAP内养护机制砂混凝土抗压强度预测.pdfVIP

2023年第5期（总第403期） 混凝土 理论研究 Number 5 in 2023(Total No.403) Concrete THEORETICALRESEARCH doi:10.3969/j.issn.1002-3550.2023.05.014 基于GA-BP神经网络的SAP 内养护机制砂 混凝土抗压强度预测 陈浩1，刘荣桂1，闫乾勋1，陈留兵² ，顾兵² ，缪卫国2 （1.江苏大学土木工程与力学学院，江苏镇江212013；2.苏中建设集团股份有限公司，江苏南通226600) 摘要：为了能够更好的预测SAP 内养护机制砂混凝土的抗压强度，利用遗传算法（GA)对BP神经网络的初始权值和阈值进 行优化，建立基于GA-BP神经网络的SAP 内养护机制砂混凝土抗压强度预测模型。结果表明：和BP神经网络相比，GA-BP神经 网络的预测精度有很大程度的提升，其最大相对误差为5.54%，平均相对误差为2.0369%，平均绝对误差为0.8581MPa，仅为BP 神经网络误差的40%，说明GA-BP神经网络能够更好的用于SAP 内养护机制砂混凝土抗压强度预测中。 关键词：机制砂混凝土；内养护；抗压强度；GA-BP神经网络；预测模型 中图分类号：TU528.064 文献标志码：A 文章编号：1002-3550(2023)05-0072-05 Based on GA-BP neural network prediction of compressive strength of machine-made sand concrete with SAP internal curing CHEN Hao, LIU Ronggui, YAN Qianxun, CHEN Liubing, GU Bing”, MIAO Weiguo? (1.Department of Civil Engineering, Faculty of Civil Engineering and Mechanics, Jiangsu University,Zhenjiang 212013, China; 2.Suzhong Construction Group Co., Ltd., Nantong 226600, China ) Abstract: In order to better predict the compressive strength of sand concrete with the curing mechanism in SAP,the initial weights and thresholds of the BP neural network are optimized by genetic algorithm (GA),and the SAP inner curing mechanism sand concrete based on GA-BP neural network is established.Compressive strength prediction model.The results show that compared with BP neural network,the prediction accuracy of GA-BP neural network has been greatly improved.The maximum relative error is 5.54%,the average relative error is 2.036 9%,and the average absolute error is 0.858 1 MPa, which is only BP.The 40% error of the neural network indicates that the GA-BP neural network can be better used in the prediction of the co