Heat transfer in a 3D rough rock fracture with heterogeneous apertures

Abstract

A clear understanding of the convective heat transfer characteristics in the three-dimensional (3D) rough rock fracture is important for evaluating heat recovery in fractured reservoirs. A granite sample containing a Brazilian-induced artificial fracture, which was tested at increasing normal compressive stresses, was adopted to build 3D numerical models for rough rock fractures. At each stress level, flow-through simulation tests with different injection velocities were performed to examine the effect of fracture geometrical alterations induced by changing stresses on the heat transfer processes within hot fractured rock samples. Meanwhile, a hypothetical rough fracture and a parallel plate fracture, with an equivalent mechanical aperture to the artificial fracture, were constructed for the same heat transfer simulations as references. The results show that the flow and heat transfer behaviors in three types of fractures are significantly different, and the difference becomes more obvious as the normal stress and flow velocity increase. The alterations in asperity contacts and void spaces due to stress changes increase the heterogeneities of the distributions of streamlines and water temperatures in the artificial fracture. The tortuosity induced by the global fluctuation of the surface roughness in the hypothetical fracture cannot describe the flow and heat transport in the artificial fracture. The heat transfer coefficient in three types of fractures decreases non-linearly as the hydraulic aperture increases. Under the same hydraulic aperture, the heat transfer coefficient is larger in the hypothetical fracture than in the other two fractures. The channeling flow within the artificial fracture under high normal stresses weakens the heat transfer, resulting in the smallest heat transfer coefficient compared to the other two fractures. An empirical model was developed to describe the relationship between the normalized heat transfer coefficient and the hydraulic aperture, and this model was validated well using the published heated flow-through experimental results.