Introduction
In cold regions, concrete is widely used as a material in hydraulic structures. Its durability is crucial for ensuring the stability of concrete structures over service life. As a porous medium, concrete undergoes freeze–thaw cycles where the water in the pores freezes and expands, causing water migration and pore pressure. This phenomenon results in internal structural damage and thus degradation of concrete performance. Therefore, it is important to analyze the relationship between temperature, humidity, and pore pressure of concrete during the freeze–thaw cycles. Scholars have proposed various mechanism theories, and a general knowledge is given that the thermo–hydro coupled behavior of concrete imposes the major machinal loads inducing the freeze–thaw damage. But the dominant mechanism is still not clear. The hydrostatic pressure theory and the crystallization pressure theory explain the freeze–thaw damage of concrete from the perspective of the effective stress principle of fluids. The theory of poroelasticity explains the freeze–thaw damage of concrete from the perspective of internal pressure and deformation generated by the pores inside the concrete during the formation and thawing of ice crystals. Hence, it is believed that an additional attention should be paid to the combined mechanical effect of solid phase expansion and the load contribution of the fluid phase. By conducting a detailed analysis of pore pressure and ice expansion deformation during the freeze–thaw process, the dominant mechanism of freeze–thaw damage can be further analyzed.
Methods
The spatial discretization method of flow lattice network (FLN) considering the concrete mesoscopic heterogeneity is adopted to describe the heat transfer and mass transport behavior. Within the framework of FLN, a thermo–hydro coupled model for concrete during the freeze–thaw cycles is proposed and its mechanical mechanism are discussed based on the poroelasticity and effective stress principle. The proposed model implemented into UEL subroutine and solved by standard/implicit solver in ABAQUS. The proposed model is used to simulate the freeze–thaw process of concrete under uniform and non-uniform temperature change.
Results and discussion
Firstly, the numerical simulations of concrete under freeze–thaw conditions are conducted based on existing concrete ice content and the corresponding deformation experiments. As the temperature decreases, the relative ice content predicted by the model begins to increase at 0 ℃, and the ice crystallization first increases and then decreases. At –55 ℃, the relative ice content reaches 0.42. Subsequently, due to the consideration of freeze–thaw hysteresis, the relative ice content does not decrease with the increase in temperature, but starts to gradually decrease only when the temperature rises to the melting temperature (–15 ℃). The ice completely thawed at 0 ℃, which is in good agreement with the experimental results. In the simulation with constant and temperature-dependent thermal expansion coefficient, the total deformation and deformation trend calculated by the freezing stage model are in good agreement with the freeze–thaw deformation test data. However, during the thawing stage, due to the damage (cracking strain) in the experimental, the model calculation results do not match the experimental results. The strain of εΣm and εΔρ show expansion, and εth shows contraction deformation. εΔρ dominates the over deformation. Additionally, through the calculation of relative errors in deformation, it can be observed that the calculation results with constant al and ac keep relatively consistent with the experimental results. During the cycling process, the peak pore pressure of 5 MPa during the freezing stage of concrete appears at 4.4 hours, and the peak pore suction pressure of 9 MPa during the thawing stage appears at 21.5 h. In the evolution of frost heave strain, the peak strain appears at 13 h.In the simulation of non-uniform temperature distribution, the variation depth of relative ice content gradually expands towards the temperature invariant region as the number of freeze–thaw cycles increases. The peak value also increases with the pore pressure expands towards the rear side. In addition, the peak deformation occurs alternately with the peak pore pressure. During the process of water freezing and thawing, the mechanical loads of the ice-water mixture in pores on the solid skeleton presents two stages. The first stage is dominated by fluid pressure and is computed by the principle of generalized effective stress. The second stage is ice crystallization induced expansion computed by the poroelasticity.
Conclusions
A thermo–hydro coupled model for concrete during the freeze–thaw cycles is proposed within the framework of FLN. The numerical results of relative ice content and the corresponding deformation are in good agreement with the experimental results, which verifies the accuracy of the proposed model. The consideration of the temperature-dependent mathematical pore distribution in the calculation of ice content can simulate the hysteresis of ice content during freezing and thawing. It is numerically found that the pore pressure and frost heave strain play the rotatory dominant role during the freeze–thaw cycle. One stage is dominated by fluid pressure and another stage is dominated by ice crystallization induced expansion computed. The effective stress caused by the pore fluid and the calculation of frost heave strain are comprehensively considered.