Rock physics based velocity-porosity correlations developed for estimation of the elastic properties of the Bakken formations of the Williston Basin, North Dakota

Abstract

Rock physics models are predictive tools used to estimate the velocity, or elastic properties of formations, based on strong theoretical foundations, as opposed to some of the simple empirical correlations that have been developed for specific regional formations. While the basic rock physics models use simplified assumptions and involve less input, more complex models, such as inclusion-based models, estimate velocity-porosity relation as function of different parameters including pore's fluid saturation and type, cementation, confining pressure and diagenesis. The downside is that these models require performing tedious calculations and in some cases solving complex differential equations. These models may also differ depending on the type of formation, for example, separate models may be used for shales compared to the siliciclastic or carbonate formations.In this paper, we present newly developed correlations that can substitute the rock physics modeling. The developed velocity-porosity correlationsare based on the mineralogical data from the Bakken formation in the Williston Basin, North Dakota, however, they can be used with good accuracy for other formations too. These models were developed based on large volumes of data from simulation of many cases using the differential effective medium (DEM) theory, a commonly used rock physics model. DEM models were developed for composite rock, but assuming single mineral type constituting the rock. The inclusion phase was incrementally added to the matrix phase. The pores were then modeled with three phase fluid (water, oil, and gas) at different saturation levels and finally, different pore's aspects ratios (AR) were assumed to simulate pores heterogeneity (i.e., crack, intraparticle, intergranular and moldic type pore geometries). The correlations constants were extracted for different key minerals in the Bakken, including quartz, calcite, dolomite, anhydrite, illite and kerogen. Having the volume fraction of minerals from lab-based XRD or Elemental Capture Spectroscopy (ECS) logs, linear averaging was applied to estimate the equivalent velocity and elastic properties of the formations.The correlations were applied to several case studies in the Bakken and also compared with the existing lab data, which showed a good agreement with the DEM model. The simplicity of using the correlations consists of simply entering the input values in an excel template that accounts for the constant parameters, using a single approach for different type of formations. This offers a great advantage for their applications. The workflow of developing the correlations and its applications in one case stduy using data from a Well in the Mountrail County will be presented.