Fluid Pressure Modeling of Faults and Simulation Optimization of Wastewater Injection in the Raton Basin, CO-NM
Robert Hernandez
MS Candidate
Advisor: Dr. Matt Weingarten
December 11, 2020
10 am
Abstract
The Raton Basin of Colorado and New Mexico had a sudden increase in seismicity beginning in 2001 that has been associated with the onset of large-scale wastewater injection operations in the region (Rubinstein et al., 2014; Nakai et al., 2017). A key parameter in understanding the spatial and temporal relationship between injection and seismicity is injection reservoir permeability. Here, we analyze injection-recovery step rate tests of the main injection reservoirs, the Dakota Formation and the Entrada Formation, to establish calibrated permeability values for those formations. Both Theis injection and recovery analytical solutions are applied using AQTESOLV, producing a permeability range of 6.4 x 10-14 m2 to 6.8 x 10-14 m2 for the Dakota Formation and 5.8 x 10-14 m2 to 8.9 x 10-14 m2 for the Entrada Formation. Our results are consistent with those used in past hydrogeologic models of the basin. These calibrated permeabilities are then used in a new, three-dimensional pore pressure model which explicitly models injection from 1994 through 2019 on the three fault zones in which most seismicity is concentrated: the Trinidad, Vermejo Park and Tercio fault zones. Using the calibrated reservoir pore pressure model, we explore how fault zone pressure is affected by two remaining hydrogeologic uncertainties: (1) the Sangre de Cristo Mountain Thrust as a hydrologic boundary on the basin’s western edge, and (2) the permeability of the three fault zones themselves. The Sangre de Cristo Mountain Thrust had little effect on the model results, with pressure over time primarily dictated by injection rates in all cases considered. Pressure changes of up to 0.50 MPa were observed at injection reservoir depths, large enough to induce seismicity and consistent with the results from the previous pressure model. We then modeled future pore pressure changes on the three fault zones in a simulation-optimization framework. The goal of the optimization is to maximize future injection volume across the basin’s 29 existing injection wells while constraining pressure changes in each of the basin’s three fault zones to (1) the onset of seismicity, (2) the peak injection rate, or (3) the 2019 pressure change. The optimization successfully produced a solution for future injection practices to remain within the desired constraints. Future work will further constrain the optimization to limit the variation in monthly injection rates. The successful permeability calibration allowed for the reconstruction of a pore pressure model of the basin to be used in the optimization work and serve as guidance for future injection practices.