Pore Pressure Effects on Net Fracture Pressure and Hydraulic Fracture Containment

Abstract

In hydraulic fracture modelling, fracture geometry along with the net pressure distribution in the fracture are the parameters the values of which are sought to be predicted away from the wellbore. In current industry standard fracture modelling packages, fracture dimensions are obtained from equilibrium height growth simulation models that are based on linear elasticity equations. Due to simplifying assumptions made in these pseudo 3D lumped models, there is considerable discrepancy between the model predicted fracture geometry and the ground truth as revealed from tiltmeter and microseismic fracture mapping techniques. Not only is the geometry different, but the observed fracturing net pressure is also higher. From numerous mapped field cases it has been established that the fractures are more contained than what is modeled.The lumped models that are commonly used for real time pressure matching and diagnostics have the advantages of faster computation and can also be trained to adhere to mapped microseismic results using field specific coefficients. Coupled simulation models that use discretized grids are computationally intensive but can give more accurate results since they incorporate more of the relevant physics that affects fracture propagation. The use of effective stress as the propagation criteria rather than the stress intensity factor approach from Linear Elastic Fracture Mechanics introduces the pore pressure effect into the simulation. The pore pressure effect on effective stress and its relationship with fracture net pressure has been qualitatively observed from both field and experimental observations. The pore pressure effect is thus surmised to be an important factor that affects fracture growth and this study seeks to investigate that hypothesis. An analysis of more than 400 datafrac injections performed in clastic reservoirs and spanning a wide range of geological ages and depositional environments were studied to establish a relationship between observed net pressure and the effective stress in the reservoir. Synthetic geomechanical cases that represent underpressured, hydrostatic and overpressured reservoirs were built using the range of fracture design parameters from the database. An additional high effective stress case was built to appraise the upper limits of effective stress as observed from the treatments. A planar explicitly coupled 3D fracture simulator (Frac3D) based on an effective stress criterion and which uses cohesive element nodes near fracture tip was used to simulate fracture growth for the synthetic cases. The numerical results obtained from this FEM – FDM coupled simulator established that the pore pressure effect is indeed a significant parameter. The fracture dimensions were seen to be more radial with depletion and longer when overpressured. The net pressure increased with depletion and fracture width was highest in this case. The high pressure high effective stress case resulted in higher net pressure than the depleted case and this was indicative of the dominance of a high closure stress over the pore pressure effect. In direct comparisons with the lumped model the Frac3D simulator predicts higher net pressures and more contained fractures whereas the lumped model tended to understimate net pressure and overestimate fracture dimensions. Based on comparison of the Frac3D results and the field fracture treatment analysis, this thesis establishes a empirical correlation between fracture net pressure and effective stress for datafrac injections. The comparison between the synthetic simulations and the Frac3D simulations presents conclusive evidence that lumped models based on linear elasticity are inadequate in representing the pore pressure effect. The obtained correlation may be applied as a pressure matching parameter for datafrac injections prior to using other parameters in the general fracture diagnostics workflow. The pore pressure effect also serves as an explanation for fracture containment in the absence of vertical stress contrast.