DOI: 10.3390/ma19132721 ISSN: 1996-1944

Shear Behavior and Predictive Model of Desert Sand Concrete Beams Subjected to Freeze–Thaw Cycles

Chao Huang, Meng Wu, Zhiqiang Li, Yingsheng Dang, Jian Li

To explore the shear behavior and evolutionary pattern of desert sand concrete beams (DSCBs) subjected to freeze–thaw cycles, 16 DSCBs were subjected to rapid freeze–thaw cycling and shear tests, with desert sand replacement ratios (0%, 20%, 40%, and 60%) and numbers of freeze–thaw cycles (0, 25, 50, and 75) considered as the main variables. The failure mode, diagonal crack development, diagonal cracking load, shear capacity, and load–stirrup strain curves of DSCBs were tested and analyzed. The results indicate that all specimens exhibited typical shear-compression failure. The diagonal crack development pattern of DSCBs was similar to that of ordinary concrete beams, whereas freeze–thaw cycles accelerated the initiation and propagation of cracks. Freeze–thaw cycling significantly reduced both the diagonal cracking load and shear capacity. After being exposed to 75 cycles of freezing and thawing, the ultimate shear capacity of test pieces with desert sand replacement proportions of 0%, 20%, 40%, and 60% decreased by 15.6%, 12.9%, 13.9%, and 13.8%, respectively, while the corresponding stirrup strains increased by 47.2%, 34.1%, 37.1%, and 53.7%, respectively. An appropriate desert sand replacement ratio can improve the shear performance of concrete beams. Among all specimens, the beam with a 20% replacement ratio exhibited the best overall mechanical performance, achieving a maximum increase of 6.0% in shear capacity and a maximum reduction of 26.8% in stirrup strain compared with conventional concrete beams. Finally, by introducing modification coefficients related to the desert sand replacement ratio as well as the freeze–thaw cycling times, predictive equations for the diagonal cracking load and shear capacity of DSCBs under freeze–thaw conditions were established. The numerical predictions achieve a high consistency with measured data.

More from our Archive