High Volume Fly Ash High Strength Concrete

In this paper, the ratio of water to cementitious material, the mass of fly ash, silicon fume and superplasticizer are selected for four factors of experiment. After the orthogonal experiment, the material ingredients of C50 and C60 high-strength concrete with 60 percent fly ash are determined. Through variance analysis, the influence of every factor to compressive strength is known. The ratio of strength to cost is introduced to discuss the problem in which whether high volume fly ash can reduce the cost of concrete. In the multi-variables linear regression analysis, the 28d cube compressive strength is looked as function and the four factors of experiment is viewed as independent variables. The linear equation between compressive strength and four factors is concluded. The optimum regression equation is deduced by stepwise method. Neural network method is applied to the strength prediction. The ratio of water to cement material, the mass of fly ash and the silicon fume are regarded as network inputs and the 28d strength is the target. The inputs and target are used to train a three layers back-propagation network. The network that has been trained can be successfully used to infer the strength of concrete with a mass of fly ash.