文献类型：期刊文献

中文题名：基于遗传神经网络的高强高掺粉煤灰材料设计

英文题名：Design of Products with High Strength and High Volume of Fly Ash Based on Genetic Neural Network

作者：周波[1];乔秀臣[1];宋兴福[1];汪谨[2];刘够生[1];于建国[1]

机构：[1]华东理工大学资源与环境工程学院,上海200237;[2]华东理工大学化工学院,上海200237

年份：2009

卷号：35

期号：5

起止页码：684

中文期刊名：华东理工大学学报（自然科学版）

外文期刊名：Journal of East China University of Science and Technology

收录：CSTPCD;;Scopus;北大核心:【北大核心2008】；CSCD:【CSCD2011_2012】；

基金：国家"863"计划项目(2007AA06Z128);上海市科委重大创新行动计划项目(07dz12027)

语种：中文

中文关键词：均匀实验设计;BP神经网络;遗传算法;罚函数

外文关键词：uniform design; BP neural network; genetic algorithm; penalty functions

摘要：应用均匀设计方法研究了高强高掺粉煤灰试块的配比关系。基于遗传算法与改进的BP神经网络建立了适用于抗压强度试块原料配比之间的数学模型。通过加入罚函数项并二次采用遗传算法对数学模型寻优求解得到最优配比,此配比下模型预测7 d和28 d强度分别达到35.61MPa和34.25 MPa,与实验结果35.89 MPa和34.10 MPa非常相近。
The influences of mix proportion of raw materials including fly ash, calcium hydroxide and chemical activator on the compressive strength were investigated by uniform design in order to produce high performance product with high volume of fly ash. A mathematical model described the relationship between compressive strengths and raw materials proportion was deduced using genetic algorithm and modified back propagation（BP） neural network. The optimal mix proportion was obtained by twice applications of genetic algorithm after introducing a penalty function in the model. The calculated compressive strengths at 7 d and 28 d were 35.61 MPa and 34.25 MPa, respectively, which were very close to the corresponding experimental results of 35.89 MPa and 34.10 MPa.