This paper presents a numerical method to predict the temperature weakening effects on granite rock. Thermally induced cracking is modelled in the continuum sense by using a damage-viscoplasticity model based on the rounded Rankine surface. The governing thermo-mechanical problem is solved with an explicit staggered method. Rock heterogeneity is described as random clusters of finite elements assigned with the constituent mineral, here Quartz, Feldspar, and Biotite, material properties. The temperature dependence of the minerals is accounted for up to 800 °C, i.e. well beyond the Curie point (573 °C) of Quartz. The simulations demonstrate that the present approach can accurately predict the experimental weakening effects on the rock strength and stiffness as well as the macroscopic failure modes in tension. Moreover, it does so in a noncircular way, i.e. not using the laboratory data on rock strength as an input data in the constitutive description.
High temperature has a detrimental effect on rock strength and stiffness (e.g. Wang & Konietzky 2019; and Toifl et al. 2017). The mechanism behind the thermal weakening under slow uniform heating, i.e., with negligible thermal gradients, can be traced to thermal cracking due to rock heterogeneity. More specifically, the mismatch of the elastic constants and thermal expansion coefficients of different mineral phases induces thermal stresses, which in turn cause cracking. Rocks with Quartz are especially prone to thermal cracking due to its highly nonlinear behavior upon approaching the α-β-transition at 573 °C.
Numerical prediction of thermal effects in rocks is an important topic in rock engineering. Saksala (2022) modelled the thermal weakening effects in granite rock under uniaxial compression and tension by assuming that only the thermal expansion of Quartz phase is (linearly) temperature dependent. This simplified approach successfully predicted the granite strength and stiffness degradation, as well as the 3D failure modes, up to 500 °C. However, deviations from the experimental data occurred at 700 °C, i.e., beyond the α-β-transition. This was clearly due to the simplifying assumption of linear temperature dependence, which is not valid for Quartz mineral. The purpose of the present paper is to mend this shortcoming by properly accounting for the nonlinear temperature dependence of Quartz thermal and elasticity properties. However, only tension tests are considered in the present study.
