摘要
采用PSO和GA算法优化BP神经网络,建立了3种神经网络预测模型(PSO-BP、GA-BP和BP),通过将不同锂渣掺量、橡胶掺量和养护时间作为神经网络输入层参数,水泥砂浆抗压强度作为输出层参数,并选择RMSE、相关性系数和平均误差来对比分析3种模型对锂渣-橡胶水泥砂浆抗压强度的预测性能。试验结果表明,PSO-BP和GA-BP模型的预测性能均优于BP神经网络,并且PSO-BP预测性能更优异。总体而言,2种算法都可以很好的提高BP模型的预测精度。
PSO and GA algorithms were used to optimize BP neural network,and three neural network prediction models of PSO-BP,Ga-BP and BP were established.Different lithium slag content,rubber content and curing time were taken as input layer parameters of neural network,and cement mortar compressive strength was taken as output layer parameters.RMSE,correlation coefficient and mean error were selected to compare and analyze the prediction performance of the three models on the compressive strength of lithium slag-rubber cement mortar.The experimental results show that the prediction performance of PSO-BP and GA-BP model is better than that of BP neural network,and the prediction performance of PSO-BP is better.In general,both algorithms can improve the prediction accuracy of BP model.
作者
胡艳香
陈昊
薛凯喜
王升福
李明东
高嘉仪
李志勇
周铭诚
HU Yanxiang;CHEN Hao;XUE Kaixi;WANG Shengfu;LI Mingdong;GAO Jiayi;LI Zhiyong;ZHOU Mingcheng(School of Civil and Architectural Engineering,East China University of Technology,Nanchang 330013,China;State Key Laboratory of Nuclear Resources and Environment,East China University of Technology,Nanchang 330013,China)
出处
《混凝土》
CAS
北大核心
2023年第10期59-63,共5页
Concrete
基金
国家自然科学基金(42167024)
江西省地质环境与地下空间工程研究中心开放基金(JXDHJJ2021-007)。
关键词
人工神经网络
锂渣砂浆
橡胶
抗压强度
artificial neural network
lithium slag mortar
rubber
compressive strength
作者简介
第一作者:胡艳香(1982-),女,硕士,讲师,主要从事新型建筑材料领域的教学科研工作。联系地址:南昌经开区广兰大道418号(330013)联系电话:17779138257;通讯作者:陈昊(1999-),男,硕士研究生,主要从事新型建筑材料领域的研究工作。联系电话:18942352455。
