ABSTRACT: Gas injection in subsurface oil reservoirs is of significant interest to the petroleum industry for EOR and pressure maintenance objectives. Geological uncertainty may impact decision making due to the limited measurements in the subsurface. Optimization under uncertainty is, therefore, required to make more robust operational decisions to achieve maximum recovery while minimizing the risk of early gas breakthrough. This work introduces an integrated machine learning-assisted workflow for the optimization under uncertainty in subsurface reservoirs. The proposed workflow includes three steps. In Step 1, training sample generation are performed, where the uncertain parameters which affect the objective of interests are identified. We then generate the input designs using Latin Hypercube Sampling (LHS) based on the identified uncertain parameters. High-fidelity simulations based on a reservoir simulator (MRST) are conducted for the input designs to obtain the objective of interests as outputs. In Step 2, surrogate model development are created. A data-driven surrogate model is then built to model the nonlinear mapping between the input and output results from Step 1. Herein, Bayesian optimization technique is implemented to obtain the surrogate model. In Step 3, optimization under uncertainty is applied. In this step, we first conduct blind test on the proposed surrogate model with high-fidelity simulations. Followed by Monte Carlo to perform the uncertainty quantifications and Genetic Algorithm to conduct the optimization. This work introduces an efficient, robust and accurate machine-learning-assisted workflow for gas injection optimization under uncertainty in subsurface reservoirs. To our best knowledge, this approach is applied for the first time.

This content is only available via PDF.
You can access this article if you purchase or spend a download.