Mechanical properties of the Bakken Formation

Abstract

The Bakken Formation is located in the Williston Basin in North Dakota, Montana, and up into southern Saskatchewan, Canada. The Bakken Formation lies unconformably over the Upper Devonian Three Forks Formation and is conformably overlain by the Lodgepole Formation. Production in the Bakken depends on horizontal wells with multistage fracture stimulations. The effective minimum horizontal stress is a primary controller of fracture growth. Knowledge of the elastic properties and Biot's poroelastic coefficients is required to accurately determine the effective minimum horizontal stress. In this study I have measured dry rock velocities for four geologic facies from the middle Bakken interval and one from the Lodgepole Formation. Mineralogy data was also obtained for many samples including the rocks measured in the laboratory. This data along with available literature measurements in the Bakken Shales allows for estimation of dry rock elastic constants and Biot's coefficients in-situ. The dry rock stiffness tensor was determined by treating the dipole shear log as a dry rock measurement. Empirical equations were then derived from laboratory data and applied to the shear waves to predict the remaining components of the stiffness tensor. Fluid substitution was performed with Gassmann's equation to acquire the saturated stiffness tensor. Biot's coefficients were calculated using the dry rock stiffness tensor and an estimate of the mineral bulk modulus by assuming a Voigt-Reuss-Hill effective medium for the pure grain moduli. Biot's coefficients describe the ability of the pore pressure to counteract the outward stresses on the rock and will be between zero and one. The values for all formations were well below one and ranged from 0.15-0.75 over the unit of interest. The saturated stiffness tensor and Biot's coefficients were input into the effective minimum horizontal stress equation assuming uniaxial strain. The stress profile showed no major contrast over the area of interest. A slight decrease in horizontal stress was observed in the common reservoir facies of the middle Bakken, but the remaining units all had similar horizontal stress. Mini-Frac tests were performed in the Upper Bakken Shale and the Scallion member of the Lodgepole Formation. The tests provide estimates of reservoir pressure, total minimum horizontal stress, tensile strength, and total maximum horizontal stress. The interpreted stress from the Mini-Frac tests matched well with the modeled results, and showed low stress contrast between the Upper Bakken Shale and Scallion member. A total minimum horizontal stress profile was provided by Schlumberger for the same well along with transversely isotropic elastic properties. The stress profile was only able to predict the Mini-Frac test in the Upper Bakken Shale, and a calculation of Thomsen anisotropy parameters (Thomsen, 1986) showed the delta parameter in the Bakken Shales ranged from 0.5-1.2. These delta values are much higher than any existing anisotropy measurements in the Bakken Shales (Vernik and Nur, 1992). The likely cause for the high values was an attempt by Schlumberger to match the Mini-Frac tests by adjusting the anisotropy and disregarding the possibility of Biot's coefficients less than unity. This lead to a massive stress contrast in the Bakken Shales that would ultimately be interpreted as a strong fracture barrier. The in-situ pressure testing and modeled results show low stress contrast throughout the unit of interest. The analysis by Schlumberger predicted a contrasting stress profile, but the input parameters were unrealistic and the profile did not match in-situ pressure testing. This demonstrates the importance of estimating accurate Biot's coefficients and realistic anisotropy parameters. A poor interpretation will impact completion strategies and potentially damage resource recovery.