Experimental and Numerical Modelling of Nonlinear Flow Behavior in Single Fractured Granite

Abstract EN

In this study, the hydromechanical behavior was experimentally investigated by conducting fracture geometry evolution-based gas flow tests on single fractured granite.

The traditional fracture roughness value (Rp) of a single surface cannot adequately explain the intrinsic permeability in multiple samples.

Instead, the permeability is better explained by aperture distribution, and it is positively correlated with the mean aperture size.

The permeability is also affected by a combination of contact area and void space under changing confining stress.

The average embedded depth increases rapidly under low-loading stress (less than 8 MPa) and slowly under high-loading stress (higher than 8 MPa).

A simplified Hertz contact model was used to fit the experimental data.

The model uses the reciprocal of the standard deviation of the aperture as the average curvature radius of the fracture.

Good agreement was found between the experimental and theoretical results.

A modified smooth parallel-plate model to describe flow friction in fractures was developed, which takes fracture contact and void space into account.

A new friction model was also developed that takes both the nonlinear effect and the influence of fracture relative roughness into account by multiplying the ideal model by two power-law functions.

This new model is effective in predicting the friction factor.

Finally, taking into consideration the built permeability, mechanical, and friction models, a nonlinear flow model was proposed.

The flow rate can be estimated based on the 3D fracture topography data, the applied loading stress, and the pressure gradient.

The influence of four types of aperture distribution on the friction factor was investigated.

The results showed that a high mean aperture size would result in a low friction factor.

In addition, the influence of the mean aperture size on the friction factor is obvious, even under a very high-loading stress while the effect of the standard deviation is negligible when the loading stress is very high.