| Because of the wide application of chemical additives and mineral admixtures in concrete the relations between the composition and the performance of modem concrete are complicated increasingly. So it is urgent to make use of the new idea, new methodology and new technology in probing into the regularity and predicting the performance of modern concrete.The approach of fuzzy neural network (FNN) is applied to research on the problems of fly ash or ggbs in suppressing expansion due to alkali -aggregate reaction (AAR) and AAR in concrete and to explore the relations of many factors with expansion due to AAR. The approaches of FNN and artificial neural network (NN ) are applied to predict strength and to design the mix proportion of concrete. Moreover, the reducing dimensional mappings based on error back propagation (BP) and radial basis function (RBF) networks have been developed for determining the optimal operational region of concrete mix proportion.Based on the experimental results it is drown that the coefficient of K2O in the calculation formula of sodium equivalent used worldwide now is too large, the region of suitable coefficient of K2O is computed, and the mechanism of which K2O and Na2O cause different AAR expansion is explored preliminarily.The progresses are as follows :1. It is one of the disquisitional stresses for this document to study the problem of fly ash in suppressing expansion due to AAR. For the purpose of study 3 kinds of Portland cement made in Northwest, Central Plains, East China and 9 kinds of fly ash collected from Northeast, Southeast, Southwest, East China are used to carry out a series of experiments from which 196 groups of sample data are obtained. We constructed respectively 4 types of FNN models and 2 types of regressive models to calculate these sample data and based on the results of calculation the essential laws of fly ash in suppressing expansion due to AAR was explored in an all-round way. The study shows that in all of the factors influencing expansion due to AAR the most important one is the amount of fly ash-replacement. The expansion due to AAR decreases as the amount of fly ash-replacement increases, and so long as the enough fly ash is used, the expansion due to AAR can be controlled by less than 0.1% even though the alkali content in cement and fly ash or alkali content fromAbstractexternal environment or the CaO content in fly ash are high. The alkali added in concrete has a quite effect on expansion due to AAR, that is, the more the added alkali is , the more it causes expansion due to AAR. Both the alkali from cement and the alkali from fly ash also has effect on expansion due to AAR, in other words, the more the alkali from them, the more expansive the concrete will be, but it is the alkali in cement that has stronger influence than the alkali in fly ash does though the alkali in fly ash has some effect on expansion due to AAR. The content of CaO or the value of CaO/(SiO2 + Al2O3 + Fe2O3) in fly ash has nearly same effect on expansion due to AAR and the conclusion is that there is the increasing tendency to expansion due to AAR as the CaO content or the value of CaO/(SiO2 + A12O3 + Fe2O3) in fly ash increases, however, the content of CaO or the value of CaO/(SiO2 + A12O3 + Fe2O3) in fly ash is not a decisive factor on the expansion due to AAR.Because the above conclusions are drawn based on the chemical composition of the sample data and the method of the test, it will have much work to do to know if these conclusions are right in any conditions or in what conditions which of these conclusions are not right.2. It is the other stress for this document to calculate the suitable coefficient region of K2O in the calculation formula of sodium equivalent. It is discovered and proved that the coefficient of K2O, which is 0.658, in the calculation formula of sodium equivalent used worldwide now is too large by using the approaches of fuzzy rule analysis, regressive analysis and graphing analysis to compute the 196 groups of sample data obtai...
|
PDF Full Text Request