Optimization of Primary and Enhanced Shale Oil Recovery Using Statistical Design of Experiments and Data Analytics

The application of statistical analysis and data analytics is not widespread in the petroleum industry, but is gaining recognition and will be applied more in the future. Applying these tools could improve the accuracy and precision of results, which could have a positive effect on the decision-making process. Therefore, both statistical analysis and data analytics constituted the core of this research to generate an efficient systematic workflow that was applied to study one of the major problems that characterize the complex shale reservoirs, namely how to improve their low recovery factor (RF). A simulation-based design-of-experiments (DoE) workflow was developed to optimize the design of four recovery schemes to maximize the low RF in Eagle Ford shale. These schemes are primary production (PP), waterflooding, continuous miscible gas flooding, and miscible gas huff ’n’ puff. The workflow was used to pinpoint the optimum spots in the multidimensional variable space. Using an innovative injection pattern that relies on alternating injection and producing fractures along the same lateral, continuous miscible gas flooding was found to have the highest potential to maximize RF. Developing this injection pattern might be the next breakthrough to boost the low shale RF. Adequate representation of the complex and challenging shale reservoirs using numerical simulation necessitates making many uncertain assumptions, which could affect the reliability of its results. Leveraging less presumptive techniques like data analytics (DA) is required to validate modeling results. Both DA and DoE were applied on a Bakken shale PP case study. Results showed that RF has a physical limit that cannot be exceeded by merely optimizing PP, which can only accelerate oil production and cash flow. Metamodeling was used for optimization and quantifying the effects of uncertainties in reservoir characteristics and design variables. The main purpose of this research is to illustrate the use of DoE in the oil and gas industry. DoE provides a systematic research framework that can be leveraged for an efficient exploration of the multidimensional variable space to pinpoint the optimum spots. This framework produces statistically-based conclusions and data-driven facts, which could improve the objectivity of the decision-making process. We hope that this work could help encourage petroleum researchers and engineers to incorporate statistics at the heart of industrial and academic research and problem-solving tools, and to learn and apply DoE regardless of the type of experiments that they conduct; physical experiments, simulation runs, or field trials