摘要
分析了影响植生型多孔混凝土抗压强度的主要因素,选取目标孔隙率、水胶比、胶凝材料用量、粗骨料用量、水用量、粗骨料平均粒径、粗骨料比表面积、粗骨料堆积孔隙率及浆骨比作为植生型多孔混凝土抗压强度的影响指标,分别建立了BP多层前馈神经网络预测模型和采用遗传算法优化的BP神经网络预测模型(GA-BP)。收集国内外文献中146组植生型多孔混凝土试验数据,以其中116组数据作为训练样本,并采用其余30组数据作为试验样本与BP、GA-BP神经网络模型预测值、线性回归方程抗压强度计算值进行比较分析,结果表明:BP、GA-BP神经网络模型计算精度与离散性更优,且较线性回归方程计算结果更接近于样本试验值,更能够准确地预测多孔混凝土的抗压强度值。
The main factors affecting the compressive strength of the plant porous concrete were analyzed.The target porosity,water-cement ratio,the amount of cementitious material,the amount of coarse aggregate,the amount of water,the average particle size of coarse aggregate,the specific surface area of coarse aggregate,the accumulation porosity of coarse aggregate and the ratio of cement to bone were selected as the influence indicators of compressive strength of plant porous concrete.BP multilayer feedforward neural network prediction model and BP neural network prediction model optimized by genetic algorithm were established respectively.146 groups of experimental data of plant porous concrete in domestic and foreign literature were collected,.In the 146 groups of experimental data,116 groups of data were training samples,and 30 groups of data were test samples.BP,GA-BP neural network model prediction value,linear regression equation calculation value of compressive strength were compared by the 30 groups of data.The results show that,BP and GA-BP neural network model has better calculation accuracy and discreteness,and more close to the sample test value than the linear regression equation,the compressive strength of plant porous concrete can be predicted more accurately.
作者
焦楚杰
谭思琪
崔力仕
何松松
彭兰
JIAO Chujie;TAN Siqi;CUI Lishi;HE Songsong;PENG Lan(School of Civil Engineering,Guangzhou University,Guangzhou 510006,China)
出处
《混凝土》
CAS
北大核心
2022年第1期7-10,16,共5页
Concrete
基金
国家自然科学基金项目(51778158,51478128)
广东省水利科技创新重点项目(2017-32)
住房和城乡建设部科研开发项目(07-K4-5,2010-K3-27,2010-K4-18)
广州大学重点产学研项目(2018-14)。
