University of CalgaryUniversity of CalgaryUniversity of Calgary
H. Hamdi;C.R. Clarkson;A. Ghanizadeh
Hydraulic fracturing in tight and shale reservoirs is a revolutionary technology enabling economically-viable production. Nevertheless, oil recovery factors achieved from primary production are still very low, typically 5-10% of original oil in place (OOIP). Recent laboratory and certain pilot results have demonstrated the technical success of enhanced oil recovery (EOR) techniques to increase oil recovery from unconventional reservoirs. In this paper, the potential of cyclic gas injection (i.e. huff-n-puff; HNP) in the Duvernay shale in Alberta, Canada is evaluated. A compositional numerical model was used to simulate the multi-contact extraction HNP process using lean and rich gas injection. A fluid model was constructed using collected fluid samples, and was tuned to several laboratory experiments. Pressure-dependent permeability (PDP) data, used to constrain the simulation, were compiled from a series of laboratory experiments. From the laboratory experiments, the impact of confining stress, during loading and unloading cycles, on the permeability of intact and artificially-fractured (propped and unpropped) core plug samples was quantified. The fluid model, laboratory measurements, and additional petrophysical data, were used as input to the simulation model which was calibrated (history-matched) against historical production data (primary recovery). The calibrated model was subsequently used to optimize the operational conditions for HNP. To assess the value of the measured PDP data, a history-matching trial, with PDP curves allowed to be a free-floating (adjustable) parameter, was performed. Importantly, the PDP curves providing the best history-match closely resembled the experimentally-measured depletion PDP curves. This step provided important validation of the laboratory-derived data, and the confidence to use PDP curves generated for the injection case. Optimized HNP simulation results, obtained using the calibrated numerical model, indicate that a 1.5-2 times increase in recovery can be obtained in a 20-year time span. The highest recoveries for the studied well are the result of higher injection rates, shorter injection, soaking, and production times, and higher flowing pressures during the production cycle. This study provides a comprehensive workflow with practical guidelines for integrating laboratory measurements and field observations for the successful calibration of a reservoir model used to history-match a Duvernay shale well. The calibrated model was employed to investigate the feasibility of HNP operations in the Duvernay shale, with a variety of injection fluids. The history-matching and HNP optimization processes were conducted using novel algorithms to minimize the number of simulation runs and achieve a balance between computation time and the quality of the history-match.