Using hydraulic fracturing laboratory experiments, we previously showed that different fluid injection methods for pressurizing a borehole lead to significant variations in specimen breakdown pressure. In this paper, we present a numerical analysis of these experiments to investigate the mechanism behind such variations.

A multiphase, poro-elastic numerical simulator is developed and utilized in this work. The equations for multiphase, multi-component fluid flow in a deforming porous medium are described and the solution algorithm used in the numerical simulator is explained. The simulator is verified against a Buckley-Leverett solution for multiphase flow and against Mandel's compaction problem for poro-elasticity. We constructed a two-dimensional (2D) numerical model with the same dimensions and properties of the laboratory specimen. The specimen is a square sheet with a circular borehole in the middle, and anisotropic far-field stresses are applied to the specimen. A total of five fluid injection cases were simulated. In the first four cases, glycerin was injected into initially dry specimens (air-saturated at atmospheric pressure) using a) fast injection rate, b) slow injection rate, c) cyclic pressure ramp, and d) constant pressure injection. Glycerin was injected in the fifth case into an initially glycerin-saturated specimen using the same fast injection rate in the first case.

As the fluid is injected into the specimen, we demonstrate the advancement of fluid saturation front and the development of pore pressure and effective hoop stress around the borehole for the five fluid injection cases. At the end of injection, it is clearly shown that the extent of glycerin infiltration or leakoff into the specimen is smallest for the fast injection, and becomes respectively larger for the slow injection, cyclic pressure ramp, and constant pressure injection. The higher fluid leakoff among the different cases also corresponds to higher pore pressure distributions and induced tensile stresses around the borehole that reduce the required breakdown pressure. The induced hoop stresses are most tensile for the constant pressure injection and least tensile for the fast fluid injection. We show that the least tensile stress using the fast fluid injection becomes the most tensile among all five cases when the specimen is fully saturated with glycerin initially.

This study draws a novel linkage between four different topics that are investigated separately in the literature: 1) the effect of fluid injection rate on breakdown pressure, 2) the effect of cyclic borehole pressurization on breakdown pressure, 3) the effect of constant pressure injection on breakdown pressure (static fatigue behavior), and 4) the effect of rock saturation on breakdown pressure. In terms of field applications, the fundamental findings from this study are utilized to develop new methods for pre-frac injection tests to reduce breakdown pressure in highly-stressed formations.