Predicting brittle zones in the Bakken Formation using well logs and seismic data.

摘要

The oil-in-place estimate for the Bakken Formation has varied from 10 billion barrels in 1974 to 503 billion barrels in 1999. However, only a small fraction of this estimate is recoverable due to the formation having very low porosity and permeability. Implementation of hydraulic fracture stages along horizontal wells in the Bakken has been productive. Recently, identification of zones where the formation is brittle has been used to improve hydraulic fracture stimulation efficiency in an effort to improve production.;The first goal for this thesis is to identify a correlation between brittleness and production data by using elastic moduli and normalized production values. The hypothesis for this study is that rock with a low Poisson's ratio and high Young's modulus will be more brittle and will ultimately produce a higher amount of oil than more ductile rock. The next goal was to create and test a method to identify brittle zones with high normalized production in a 3D seismic data set without well control using producing wells from outside the survey with dipole sonic logs from the Bakken Formation.;Correlations between normalized production values and elastic moduli were subsequently identified. Cumulative first-four-months' production was found to have the best correlation to the elastic moduli. Correlations of normalized production values and Poisson's ratio showed that sections of the middle Bakken with low Poisson's ratio yield higher normalized production values. Correlations of Young's modulus and normalized production showed that middle Bakken zones with low Young's modulus have higher normalized production values. However, when using additional wells that were not used for well-to-3D seismic correlations, the correlation shows that higher Young's modulus yield higher normalized production. The correlation with additional wells best represented the data and agrees with the initial hypothesis. Brittle zones were mapped in a 3D seismic data set by correlating well synthetics to the seismic traces and importing the data created from the correlation into a self-organizing map (SOM). SOM analysis identified small brittle zones with high production in the northwestern part of the survey.