Abstract:In order to understand the influence of different dosages of sodium silicate on the mechanical performance of Rubber Lightweight Concrete(RLC), sodium silicate was added into rubber lightweight aggregate concrete(RLC)and 20% fly ash was used instead of cement. The concrete blocks with content of 0, 2%, 4%, 6% and 8% sodium silicate were prepared. Through NMR to study the microscopic change and strength formation mechanism. The logarithmic function curve prediction and BP neural network strength prediction were carried out to judge the reliability and compare the advantages and disadvantages. The results show that the compressive strength of RLC increases first and then decreases with the addition of sodium silicate, and the best content of sodium silicate is 2%. The mechanical properties of 80-mesh RLC are almost the same as the optimal group(20-mesh RLC group)with 28 d, and the mechanical properties can be improved by adding appropriate amount of sodium silicate. After the addition of sodium silicate, the porosity shows a tendency of first decreasing and then increasing, which indicates that the addition of appropriate content of sodium silicate can optimize the pore structure of concrete and improve the strength of concrete. The prediction models of curve fitting and BP Neural Network are established. Both models can be used to predict the concrete age strength, but the stability, reliability and comprehensiveness of parameters of the BP-neural network prediction model are better than those of curve fitting models.