No abstract

No abstract

No abstract

Calibration of reservoir models for unconventional hydrocarbon reservoirs requires permeability data as input. Accurate permeability prediction from velocity data is desirable due to the relative abundance of velocity data that is typically available during exploration and development programs (e.g., through seismic imaging and well logging). Therefore, the development of fast and inexpensive ‘screening’ techniques tha can provide reliable estimation of permeability at high-resolution (cm-scale) using velocity data could be valuable to exploration/development programs in unconventional reservoirs. A new experimental apparatus is described herein for measuring ultrasonic velocities (P- and S-wave) along the length of slabbed cores of low-permeability rocks at high-resolution (cm-scale). A statistical approach that combines along-core (profile) ultrasonic velocity testing and non-destructive experimental techniques (X-ray fluorescence, mechanical hardness, and profile permeability) is employed to develop predictive models for estimating permeability. Two slabbed cores from the Canadian Montney and Bakken formations, covering multiple geological intervals (tight siltstones/sandstones units), were analyzed for validation purposes. Reasonable agreement is found between log- and lab-derived (ultra)sonic velocity data, indicating similar trends with depth. However, the exact log- and lab-derived (ultra)sonic velocity values are different due to the differences in stress conditions between the field and laboratory measurements and the direction of wave travel. A maximum variation of ±20 m/s is observed for both P- and S-wave velocities when measurements were repeated on the same points, providing evidence of experimental repeatability and reproducibility. Relationships exist between laboratory-measured profile ultrasonic velocities (S-wave), profile permeability, mechanical hardness, and clay content (inferred from elemental composition data). The profile S-wave velocities decrease with increasing permeability (R2 = 0.6, n = 230). Advanced statistical methods (e.g., genetic algorithms) are employed to improve the velocity-permeability relationship and develop models for indirect estimation of permeability from S-wave velocities. The performance of these models is dependent upon lithology and rock fabric (e.g., silt vs. sand, degree of cementation), with a better correlation achieved for intervals with lower porosity and permeability (<±15% maximum discrepancy between measured and predicted permeability values; R2 = 0.78, n = 230). This study introduces a new experimental apparatus, and a practical ‘screening’ workflow, that can be used for permeability prediction using S-wave velocities collected on slabbed cores. This predictive model can be used to estimate permeability below the lower limit (0.001 md) of pressure-decay profile permeability measurements. The findings are beneficial to operators developing tight siltstone/sandstone resources by allowing them to characterize permeability in low-permeability (<0.001 md) intervals for applications such as optimizing stimulation design and subsurface fluid injection.

No abstract