Search

Existing earthquake damage reconnaissance reports indicate that many tunnels have suffered noticeable damage. During seismic excitation, ignoring the rock–concrete interaction may affect the accuracy of research results. In this study the influence of rock–concrete interaction on the response of a tunnel under seismic excitation was systemically investigated. A finite-difference method (FDM) model of the tunnel was constructed and full dynamic time history calculations were performed. The rock–tunnel interface was built using the FDM, with interface elements defined by a linear Coulomb shear strength criterion. The reliability of the numerical simulation of the interface was verified by comparisons with direct shear test results. Based on a real engineering case (Xianglushan water conveyance tunnel), two models with different rock–tunnel interactions were established and their seismic responses evaluated. The mechanical properties of the interface were compared. The seismic responses of the tunnel were assessed in terms of displacements, stresses and lining dynamic internal forces. The results indicated that rock–concrete interaction may have a significant effect on a tunnel's seismic response, and that effect has a certain pattern. The influences of the mechanical parameters of the interface on the seismic response of tunnels are complex and the stiffness of the interface plays a significant role. [ABSTRACT FROM AUTHOR]

This study, proposes the use of a novel rubber-sand concrete (RSC) material, which comprises rubber particles, sand, and cement, as an aseismic material in practical engineering construction. The uniaxial compressive strength (UCS) of damping materials is an important factor that directly affects the seismic activity in underground structures. To predict the UCS of RSC, artificial intelligence model back propagation neural network (BPNN), which is optimized through four swarm intelligence optimization (SIO) algorithms: particle swarm optimization algorithm (PSO), fruit fly optimization algorithm (FOA), lion swarm optimization algorithm (LSO), and sparrow search algorithm (SSA), is used. The dataset for the prediction models was obtained from uniaxial compression tests in the RSC laboratory. The performances of the four hybrid intelligence models were evaluated using six performance indicators: the root mean square error (RMSE), correlation coefficient (R), determination coefficient (R2), mean absolute error (MAE), mean square error (MSE), and sum of square error (SSE).The prediction capability of these models was graded based on these indicators using a ranking system. The results show that the prediction ability of the LSO-BPNN hybrid model is better than that of the three other hybrid models, with RMSE of (1.0635, 1.2352), R of (0.9887, 0.9713), R2 of (0.9776, 0.9165), MAE of (0.7257, 0.8243), MSE of (1.1352, 1.5256), SSE of (64.7074, 36.6151), and ranking score of (24, 24) in the training and testing phases, respectively. Therefore, the LSO-BPNN hybrid model is an efficient and accurate method for predicting the UCS of RSCs. Sensitivity analysis showed that rubber and sand were the most important elements that affected UCS prediction, followed by cement, with the lowest relative importance being RPZ. This study provides guidance for the extension and application of RSC materials to underground seismic engineering. [ABSTRACT FROM AUTHOR]