应用GRNN网络预测高强混凝土强度

为了更好地解决混凝土强度的预测问题,提出了应用GRNN(广义回归网络)预测混凝土28d强度的方法。该网络是建立在数理统计基础之上的一种新型的神经网络,能够综合考虑影响混凝土强度的各种因素,具有良好的函数逼近效果。利用试验所得数据,建立了GRNN网络模型和BP网络模型,并对两种网络进行了比较。     

神经网络在高强粉煤灰混凝土强度预测及优化设计中的应用

首先介绍了神经网络应用中使用最为广泛的BP网络和RBF网络的模型及其学习算法,然后将其用于高强粉煤灰混凝土的强度预测和优化设计,并与线性回归进行了对比,结果表明神经网络方法具有较高的预测精度,在混凝土性能预测和优化设计中具有广阔的应用前景。     

高强混凝土强度预测的支持向量机模型及应用

高强混凝土的强度受多种因素的影响,其强度的预测是一个动态性的系统工程。采用支持向量机理论,建立了高强混凝土的强度预测的支持向量机预测模型。并将该模型计算结果与实测混凝土28 d抗压强度、BP网络计算的强度、RBF径向基函数神经网络计算的强度、线性回归模型计算的强度、非线性回归模型计算的强度进行比较。研究表明:预测结果与实测结果吻合较好,较线性回归和神经网络预测精度高,为高强混凝土的强度预测提供了一条新途径。     

混凝土强度预测的人工智能方法

针对混凝土强度预测中存在的问题,提出了一种人工智能模型——人工神经网络预测方法,用人工神经网络对16种配比的混凝土进行了28 d强度预测,结果表明该模型可靠度很高,可用于优化混凝土的试配,从而节约大量的人力、物力和时间。     

高强混凝土强度预测的未确知聚类预测优化模型

高强混凝土的强度受多种因素的影响,其强度的预测是一个动态性的系统工程。建立了高强混凝土强度预测的未确知聚类的预测优化模型。并将未确知聚类预测优化模型计算的强度结果与实测混凝土28 d抗压强度进行比较。研究表明,预测结果与实测结果吻合一致,说明该预测模型具有较高的预测精度,为以后高强混凝土强度预测提出一种新方法和一条新途径。     

高强混凝土强度预测的高斯过程机器学习模型

高强混凝土强度与其影响因素之间存在着复杂的非线性关系,针对现有方法的局限性,建立了基于高斯过程机器学习的高强混凝土强度预测模型.该模型采用高斯过程机器学习方法对少量试验样本的测试结果进行学习,就可以建立高强混凝土强度与其影响因素之间的复杂非线性映射关系.研究结果表明:高强混凝土强度预测的高斯过程机器学习模型是可行的,具有预测精度高、适用性强、参数自适应化且易于实现的优点.     

基于BP神经网络预测高强钢筋与高强混凝土的粘结锚固强度

应用Matlab语言及其神经网络工具包,通过BP神经网络对高强钢筋与高强混凝土间的粘结锚固强度进行模拟和预测,结果表明该法能较准确地预测高强混凝土间的粘结锚固强度。     

高强混凝土的试件强度及检验

本文分析了影响高强度混凝土试件强度检测结果的主要因素,试件强度与构件混凝土强度的相关性;提出了构件混凝土强度检验中存在的问题和措施。     

养护方法对高强混凝土强度的影响

高强混凝土材料用量大,其水化速度取决于周围环境的温湿度,影响混凝土增强度的增长,将高强混凝土试件放入标准养护水中,盖遭,民塑料薄膜,空气中和涂养护剂六种不同环境中,通过试验和分析比较可知,养护条件好,混凝土强度早期增长速度就,抗碳化能力就强,因此要加强高强混凝土早期养护。     

三种神经网络在混凝土强度预测中的应用

通过建立改进的BP和RBF两种神经网络模型，对混凝土的强度进行预测，将预测值与常规BP神经网络模型预测结果进行了比较，研究表明，改进的BP和RBF的神经网络模型能够充分考虑影响混凝土强度的各种因素，在强度预测中具有广泛的应用前景。     

混凝土强度预测的两种非线性模型比较研究

混凝土的强度的预测是一个复杂的问题,受多种因素的影响.采用两种先进的非线性算法支持向量机与小波支持向量机,建立 了混凝土强度预测的两种非线性预测方法.研究结果表明:两种方法的预测结果与实测结果吻合较好,小波支持向量机的预测精度较支持向量机精度高,在混凝土的强度预测中具有较好的适应性.     

基于非线性优化理论的混凝土材料配合比研究

建立了高强混凝土的强度预测的非线性优化模型。并将该模型计算结果与实测混凝土28 d抗压强度进行比较。改用十进制遗传算法在训练过程中搜索最优超参数,形成遗传-组合核函数高斯过程回归算法,并编制了相应的计算程序,研究结果表明:与单一核函数高斯过程回归算法和支持向量回归(SVR)算法相比,提出的遗传-组合核函数高斯过程回归算法显著提高了预测精度,预测结果与实测结果吻合较好,具有较高的预测精度,为高强混凝土的强度预测提供了一条新途径。     

Strength properties and durability aspects of high strength concrete using Korean metakaolin

Metakaolin is a cementitious material used as admixture to produce high strength concrete. In Korea, the utilization of this material remained mainly limited to fireproof walls but began recently to find applications as a replacement for silica fume in the manufacture of high performance concrete.In order to evaluate and compare the mechanical properties and durability of concrete using metakaolin, the following tests were conducted on concrete specimens using various replacements of silica fume and metakaolin; mechanical tests such as compressive, tensile and flexural strength tests, durability tests like rapid chloride permeability test, immersion test in acid solution, repeated freezing and thawing test and accelerated carbonation test.Strength tests revealed that the most appropriate strength was obtained for a substitution rate of metakaolin to binder ranging between 10% and 15%. It was observed that the resistance to chloride ion penetration reduced significantly as the proportion of silica fume and metakaolin binders increased. The filler effect resulting from the fine powder of both binders was seen to ameliorate substantially the resistance to chemical attacks in comparison with ordinary concrete. Durability tests also verified that concrete using metakaolin bore most of the mechanical and durability characteristics exhibited by concrete using silica fume. The tests implemented in this study confirmed that metakaolin constitutes a promising material as a substitute for the cost prohibitive silica fume.     

早期强度推定法在商品混凝土生产中的应用

阐述了阳煤集团宏厦建筑三公司混凝土原材料的使用现状,并通过试验,对混凝土29 h快速养护强度与混凝土28 d标准养护强度进行了对比,结果表明混凝土快速养护检测技术应用于混凝土生产中取得了很好的效果。     

基于人工智能技术的超声-回弹综合法检测混凝土强度的研究

在试验研究的基础上,建立了混凝土强度超声-回弹综合法检测的人工神经网络模型.探索并应用了神经网络的改进算法,其中包括附加冲量法,自适应学习算法及S型函数输出限幅算法等,以保证建立的神经网络的快速有效.与传统的回归算法相比,采用人工神经网络建立的混凝土强度评定模型推测出的混凝土强度具有更高的精度.     

Strength of slender concrete filled high strength steel box columns

The use of thin walled steel sections coupled with concrete infill has been used on various building projects with great advantage. The currently available international standards for composite structures are limited to the design of concrete filled steel columns with compact sections. However, there is limited research work in the literature available which is concerned with slender concrete filled thin-walled steel columns. This paper presents a comprehensive experimental study of thin walled steel sections utilising high strength steel of a thin walled nature and filled with normal strength concrete. A numerical model is developed herein in order to study the behaviour of slender concrete filled high strength steel columns incorporating material and geometric non-linearities. For this analysis, the equilibrium of the member is investigated in the deformed state, using the idealised stress–strain relationships for both the steel and concrete materials, considering the elastic and plastic ranges. This paper presents both an experimental and theoretical treatment of coupled local and global buckling of concrete filled high strength steel columns sometimes termed interaction buckling. The experimental results of columns with high strength steel casings conducted herein by the authors are used for comparison. The effect of the confined concrete core is also addressed and the method shows good agreement with the experimental results of concrete filled steel columns with compact sections. The behaviour of concrete filled steel slender columns affected by elastic or inelastic local buckling is also investigated and compared with relevant experimental results. The paper then concludes with a design recommendation for the strength evaluation of slender composite columns using high strength steel plates with thin-walled steel sections, paying particular attention to existing codes of practice so as not to deviate from current design methodologies.