基于广义回归神经网络的β型磷建筑石膏强度预测  

Prediction of β-phosphogypsum Strength Based on General Regression Neural Network

作　　者：张文佳[1] 赵志曼[1] 全思臣 朱伟明[1] 姚毅惠 郜峰[1] 

机构地区：[1]昆明理工大学建筑工程学院,昆明650100 [2]云南昆钢结构有限公司,安宁650300

出　　处：《硅酸盐通报》2016年第7期2166-2170,共5页Bulletin of the Chinese Ceramic Society

基　　金：国家自然科学基金资助项目(51264017);与云南昆钢结构有限公司合同科研项目

摘　　要：本文利用工业废弃物磷石膏制备β型磷建筑石膏,并确定了影响β型磷建筑石膏强度的因素及特点,在此基础上,建立了β型磷建筑石膏强度预测的广义回归神经网络(General Regression Neural Network,GRNN)模型,利用实验室中制备β型磷建筑石膏的15组统计数据作为学习样本,通过网络拟合训练和预测分析,得到了较高精度的预测结果,证明了GRNN的非线性映射能力、容错性和自学习性用于β型磷建筑石膏强度预测是非常有效的,避免了大量盲目的配比试验及资源浪费,提高了实验水平和效率。Phosphogypsum from industrial waste is used to preparing β-phosphogypsum. The General Regression Neural Network （GRNN） model for the prediction of β-phosphogypsum strength is established by determining the influent factors and characteristics of β-phosphogypsum. The network is trained with 15 groups laboratorial data as learning sample, which gets a good fitting effect. With the high precision of predicted results, the ability of GRNN which includes nonlinear mapping function, fault tolerance and self-study, is demonstrated efficaciously in predicting β-phosphogypsum＇ s strength. By using GRNN model , large amounts of repeated proportioning test and resource waste can be avoided, improve economic benefits.

关 键 词：β型磷建筑石膏 广义回归神经网络 强度预测 

分 类 号：TU526[建筑科学—建筑技术科学]