|
位置：首页 >> 期刊文章
|
|
摘要：
|
|
提出一种基于支持向量机(SVM)的粉煤灰混凝土氯离子渗透性新型智能预测模型,并给出了相应的步骤和算法。通过该模型分析了水胶比、水泥用量、用水量及粉煤灰掺入量4个因素对粉煤灰混凝土渗透性的影响。在此基础上,能利用有限的试验数据方便地由不同的配比对所浇注混凝土渗透性进行预测,有助于准确认识混凝土耐久性随配比参数的变化规律。通过具体实例及与BP预测方法效果对比表明,该模型构造速度高2~3个数量级,预测精度高5倍左右。因此该模型在混凝土性能预测和优化设计中具有广阔的应用前景。
|
|
A novel prediction model for chloride permeability in fly ash concrete based on support vector machine(SVM)was proposed.The design steps and learning algorithm were also hiven.Four main factors of the water and cement ratio,complex scoria,cement and water quantity,which influence the chloride permeability in fly ash concrete,were,analyzed with the proposed model.Based on this,the chloride permeability in fly ash concrete by different proportioning can be conveniently predicted with limited test data.The practical experiment results show that the speed of this SVM model is 10 or 100 times,while the prediction errors are 80% less than,that of BP model,therefore,this model has bright application prospect in prediction of the performance and optimal design of concrete.
|
|
基金项目：
|
|
国家自然科学基金(70272032);;
江西省教育厅科技项目(2007328)资助
|
|
作者简介：
|
|
|
|
参考文献：
|
|
[1]赵铁军.高性能混凝土的渗透性研究[D].北京:清华大学,1997.
[2]T Zhang,O E Gjorv.Diffusion behavior of chloride ions inconcrete[J].C.C.R.,1996,26(6):907-917.
[3]马保国.海洋高性能混凝土(MHPC)的研究[D].武汉:武汉理工大学,2000.
[4]R K Dhir,M R Jones,S L D Ng,Prediction of total chloridecontent profile and concentration/time-dependent diffusioncoefficient for concrete[J].Magazine of Concrete Research,1998,(1):37-48.
[5]P S Mangat,B T Molloy.Prediction of long term chlorideconcentration in concrete[J].Materials and Structures,1994,27:338-346.
[6]P B Bamforth.The deviation of input data for modeling chlo-ride ingress from eight-year UK coastal exposure[J].Mag-azine of Concrete Research,1999,(2):87-96.
[7]张兰芬,赵兴君,岳瑜.用BP网络预测粉煤灰混凝土的渗透性[J].粉煤灰综合利用,2002,(4):37-38.
[8]V N Vapnik.The nature of statistical learning theory[M].New York:Spring-Verlag,1999.
[9]V N Vapnik.An overview of statistical learning theory[J].IEEE Transaction Neural Networks,1999,10(5):988-999.
|
|
服务与反馈：
|
|
【文章下载】【加入收藏】
|
|
|
|
|