No abstract

Summary Recent studies have indicated that huff ‘n’ puff (HNP) gas injection has the potential to recover an additional 30 to 70% oil from multifractured horizontal wells in shale reservoirs. Nonetheless, this technique is very sensitive to production constraints and is impacted by uncertainty related to measurement quality (particularly frequency and resolution) and lack of constraining data. In this paper, a Bayesian workflow is provided to optimize the HNP process under uncertainty using a Duvernay shale well as an example. Compositional simulations are conducted that incorporate a tuned pressure/volume/temperature (PVT) model and a set of measured cyclic injection/compaction pressure-sensitive permeability data. Markov-Chain Monte Carlo (MCMC) is used to estimate the posterior distributions of the model uncertain variables by matching the primary production data. The MCMC process is accelerated by using an accurate proxy model (kriging) that is updated using a highly adaptive sampling algorithm. Gaussian processes are then used to optimize the HNP control variables by maximizing the lower confidence interval (μ-σ) of cumulative oil production (after 10 years) across a fixed ensemble of uncertain variables sampled from posterior distributions. The uncertain variable space includes several parameters representing reservoir and fracture properties. The posterior distributions for some parameters, such as primary fracture permeability and effective half-length, are narrower, whereas wider distributions are obtained for other parameters. The results indicate that the impact of uncertain variables on HNP performance is nonlinear. Some uncertain variables (such as molecular diffusion) that do not show strong sensitivity during the primary production strongly impact gas injection HNP performance. The results of optimization under uncertainty confirm that the lower confidence interval of cumulative oil production can be maximized by an injection time of approximately 1.5 months, a production time of approximately 2.5 months, and very short soaking times. In addition, a maximum injection rate and a flowing bottomhole pressure around the bubblepoint are required to ensure maximum incremental recovery. Analysis of the objective function surface highlights some other sets of production constraints with competitive results. Finally, the optimal set of production constraints, in combination with an ensemble of uncertain variables, results in a median HNP cumulative oil production that is 30% greater than that for primary production. The application of a Bayesian framework for optimizing the HNP performance in a real shale reservoir is introduced for the first time. This work provides practical guidelines for the efficient application of advanced techniques for optimization under uncertainty, resulting in better decision making.

Accurate assessment of Huff-n-Puff (HnP) performance using compositional reservoir simulation requires a representative fluid model tuned to several PVT measurements.In unconventional reservoir applications, fluid models are typically constructed using laboratory depletion tests (e.g. CCE and CVD) only. In this work, multiple depletion and gas injection tests (e.g. swelling, shrinkage, and multiple-contact tests) are integrated to construct a common Equation of State (EOS) that is used to evaluate HnP performance for a Montney light oil example. Several sets of depletion and gas injection PVT data were available for this study.However,the injection tests were conducted using oil samples taken at different production times. Further, different hydrocarbon injection gases were used to perform the experiments. Building a common EOS for this range of measurements, which were conducted on multiple samples, is not a straightforward task. Therefore, a workflow, and several computer programs, are developed to simulate all the PVT tests simultaneously and to conduct the regression process. The resulting EOS is then used to construct a representative compositional simulation model. The model is calibrated through history-matching and employed to design an optimal HnP process for the studied Montney well. The results are then compared with a case where no injection tests were used to develop the fluid model. The results indicate that it is particularly challenging for the regression process to maintain a balance between the quality of the match for the depletion and the injection tests.This process required some unique global optimization methods to build a reliable EOS that matched all the measured data. For this study, the importance of the injection PVT tests is mainly reflected in tuning the interfacial tension, and secondarily the viscosity and phase density values. However, in this case study, it appears that the importance of the injection tests for tuning the EOS is marginal. In other words, depletion tests were sufficient to calibrate an EOS that resulted in an acceptable match to many measured data points obtained from multi-contact and swelling tests. This finding is mainly related to the fact that all the injected gases are hydrocarbon gases with a composition consistent with the solution gas in the oil samples. Therefore, the PVT model could also be used for injection simulations, even though the EOS was calibrated to the depletion tests only. However, it is expected that this is not the case for other non-hydrocarbon gas injection tests (e.g. using CO2 or N2) where the depletion tests cannot easily constrain the properties of the injectants during the depletion process. The constructed PVT models are used as input to dual-porosity dual-permeability (DP-DK) models, which are calibrated using multi-phase production data. The results further indicate that the two EOSs could predict an optimal HnP process with a minimal recovery difference. A new fluid modelling workflow is introduced for the first time to evaluate the importance of various gas injection PVT experiments on HnP performance prediction. This new method is tested against a field example with several measurements from a multi-fractured horizontal well (MFHW) in the Montney Formation in Canada.

Recent studies have indicated that Huff-n-Puff (HNP) gas injection has the potential to recover an additional 30-70% oil from multi-fractured horizontal wells in shale reservoirs. Nonetheless, this technique is very sensitive to production constraints and is impacted by uncertainty related to measurement quality (particularly frequency and resolution), and lack of constraining data. In this paper, a Bayesian workflow is provided to optimize the HNP process under uncertainty using a Duvernay shale well as an example. Compositional simulations are conducted which incorporate a tuned PVT model and a set of measured cyclic injection/compaction pressure-sensitive permeability data. Markov chain Monte Carlo (McMC) is used to estimate the posterior distributions of the model uncertain variables by matching the primary production data. The McMC process is accelerated by employing an accurate proxy model (kriging) which is updated using a highly adaptive sampling algorithm. Gaussian Processes are then used to optimize the HNP control variables by maximizing the lower confidence interval (μ-σ) of cumulative oil production (after 10 years) across a fixed ensemble of uncertain variables sampled from posterior distributions. The uncertain variable space includes several parameters representing reservoir and fracture properties. The posterior distributions for some parameters, such as primary fracture permeability and effective half-length, are narrower, while wider distributions are obtained for other parameters. The results indicate that the impact of uncertain variables on HNP performance is nonlinear. Some uncertain variables (such as molecular diffusion) that do not show strong sensitivity during the primary production strongly impact gas injection HNP performance. The results of optimization under uncertainty confirm that the lower confidence interval of cumulative oil production can be maximized by an injection time of around 1.5 months, a production time of around 2.5 months, and very short soaking times. In addition, a maximum injection rate and a flowing bottomhole pressure around the bubble point are required to ensure maximum incremental recovery. Analysis of the objective function surface highlights some other sets of production constraints with competitive results. Finally, the optimal set of production constraints, in combination with an ensemble of uncertain variables, results in a median HNP cumulative oil production that is 30% greater than that for primary production. The application of a Bayesian framework for optimizing the HNP performance in a real shale reservoir is introduced for the first time. This work provides practical guidelines for the efficient application of advanced machine learning techniques for optimization under uncertainty, resulting in better decision making.

No abstract