Machine-Learned Surrogate Models for Efficient Oil Well Placement Under Operational Reservoir Constraints
Machine learning has been shown to have a promising role in oil and gas explorations in recent years. Among the applications, determining a proper location for injection and production wells along with their optimal operating conditions is a complex problem.
Accurately predicting reservoir characteristics is essential for maximizing hydrocarbon production and increasing profitability. Determining the optimal number, location, and operating conditions of wells is particularly important because of the complex nature of influential parameters such as reservoir heterogeneity.
These parameters significantly determine the ultimate recovery factor and should be accurately identified during the planning process for new green or explored brown hydrocarbon fields. To accurately predict these parameters, algorithms that can handle the nonlinearity of objective functions in oil and gas reservoir production characteristics are required.
The optimization of this issue is intricate and requires substantial computer resources during both reservoir simulations and post-processing phases.
The presence of geological complexity, characterized by heterogeneity, may result in significant variations in overall production or recovery via minor modifications in operating circumstances or well placements. The growing traction of artificial intelligence, especially supervised machine-learning (ML) techniques, in the oil and gas industry can be credited to its efficiency in handling large-scale data, cutting down computational expenses, and time-saving capabilities.
Among the applications of ML in the oil and gas industry, determining a proper location for injection and production wells along with their optimal operating conditions is a complex problem. This research aims to develop a unified process using surrogate proxy models to address this issue.
Five robust ML models—extreme gradient boosting, light gradient boosting machine, gradient boosting with categorical features support, support vector regression, and multilayer perceptron (MLP)—are implemented to create surrogate proxy models for estimating the net present value (NPV) of an oil reservoir. A systematic approach is used to find the best-fit hyperparameter inputs for these models.
The objective of this method was to refine a broad set of hyperparameters through a random cross-validation search technique. This grid cross-validation method investigates the space narrowed in more accurate intervals. The following four reservoir scenarios are considered:
- Production from a single well in a homogeneous reservoir
- Production from a single well in a heterogeneous channelized reservoir
- Production from multiple wells in a heterogeneous reservoir
- Waterflooding into a heterogeneous reservoir
A reservoir simulator is implemented to create a data set of reservoir realizations with various input parameters (i.e., well location, number of wells‚ production-injection well distance, and interwell angles) in a broad range of operating conditions. The prediction of gradient boosting and MLP models showed a better fit to the simulated data with an R-squared (R2) above 95% in the first three scenarios and 75% in the fourth scenario.
The results indicate that the implemented proxies are promising approaches to efficiently estimate the NPV of the reservoir models both during primary and secondary recovery scenarios.