Experimental and numerical model for mechanical properties of concrete containing fly ash: Systematic review

被引:20
作者
Fasihihour, Nazanin [1 ]
Abad, Javad Mohebbi Najm [2 ]
Karimipour, Arash [3 ,4 ]
Mohebbi, Mohammad Reza [5 ]
机构
[1] Univ Politecn Torino, Dept Telecommun & Elect, Turin, Italy
[2] Islamic Azad Univ, Dept Comp Engn, Quchan Branch, Quchan, Iran
[3] Univ Texas El Paso, Dept Civil Engn, El Paso, TX 79968 USA
[4] Ctr Transportat Infrastruct Syst CTIS, El Paso, TX 79968 USA
[5] Univ Passau, Dept Comp Sci, D-94032 Passau, Germany
关键词
Artificial neural networks; Deep neural network; Fly ash; Mechanical properties; Multilayer perceptron; Radial basis function; Support vector regression; HIGH-STRENGTH CONCRETE; HIGH-VOLUME; DURABILITY PROPERTIES; COMPRESSIVE STRENGTH; AGGREGATE CONCRETE; RECYCLED AGGREGATE; TENSILE-STRENGTH; SILICA FUME; PERFORMANCE; RESISTANCE;
D O I
10.1016/j.measurement.2021.110547
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
Increasing use and need for cement have forced researchers to build alternative building materials that are environmentally friendly in nature and help manage less waste. Using waste materials to produce the structural concrete needs theoretical models for implement application for construction. Fly ash as a byproduct of coal-fired electric generating plants showed many advantages in improving the properties of concrete due to its pozzolanic feature. Therefore, this study intends to review previous studies and propose new models to determine the elastic moduli, compressive and tensile strengths of concrete produced by fly ash as a replacement of cement. For this aim, wide-range experimental results were evaluated and provided from previous studies. The disadvantage of the previous study is not covering all elastic moduli, compressive and tensile strengths together. Therefore, a total of 263 concrete mixtures were also produced in this study and the gap of the previous investigations was filled. Therefore, highly accurate machine learning models were used in MATLAB software to predict the mechanical properties of concrete produced by fly ash including radial basis function, multilayer perceptron, support vector regression, adaptive-network-based fuzzy inference system and deep neural network. Additionally, the experimental results were compared with existing models and new highly accurate models were developed to determine the modulus of elasticity, compressive and tensile strengths of concrete produced by various fly ash contents. Experimental results showed that the optimal content for fly ash was obtained by 10% in terms of the maximum mechanical properties of concrete. In addition, used Artificial Neural Networks, particularly deep neural networks showed a highly accurate prediction. Moreover, the proposed models by a high agreement with experimental results (R-2 > 0.98) could be used as highly efficient and accurate tools to determine the mechanical properties of fly ash concrete.
引用
收藏
页数:30
相关论文
共 114 条
[1]   Evaluation of modulus of elasticity of ultra-high performance concrete [J].
Alsalman, Ali ;
Dang, Canh N. ;
Prinz, Gary S. ;
Hale, W. Micah .
CONSTRUCTION AND BUILDING MATERIALS, 2017, 153 :918-928
[2]  
[Anonymous], 2010, ACI 363R 10
[3]  
[Anonymous], 2007, JSCE 07
[4]  
[Anonymous], 2008, JCI-08
[5]  
[Anonymous], 1992, NS 3473
[6]  
[Anonymous], 2014, ACI 318--14, DOI DOI 10.14359/51716937
[7]  
Antonio GulliS.P., 2017, Deep learning with Keras
[8]  
Arena P, 1998, NEURAL NETWORKS MULT, P234
[9]  
Arezoumandi M, 2015, ACI MATER J, V112, P559
[10]   Effect of fly ash replacement level on the bond strength of reinforcing steel in concrete beams [J].
Arezoumandi, Mandi ;
Looney, Trevor J. ;
Volz, Jeffery S. .
JOURNAL OF CLEANER PRODUCTION, 2015, 87 :745-751