OPTIMIZATION OF NATURAL GAS FIELD DEVELOPMENT USING ARTIFICIAL NEURAL NETWORKS
Open Access
- Author:
- Olatunji, Adewale
- Graduate Program:
- Petroleum and Mineral Engineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- None
- Committee Members:
- Luis Ayala, Thesis Advisor/Co-Advisor
Luis F Ayala H, Thesis Advisor/Co-Advisor - Keywords:
- Natural Gas Field Development
Artificial Neural Networks - Abstract:
- Field development of natural gas reservoirs is one of the main aspects of exploration and production of natural gas for oil and gas operators. After a natural gas field is deemed economically viable for development, and the reservoir properties have been determined, a field development plan will normally be put together as a blueprint for producing the field. However, since the main objective of natural gas field operators is to maximize profits, it is imperative to understand how to optimize recovery from the field. In this study, a model that uses reservoir engineering concepts to determine the optimum hydrocarbon that can be produced per dollar spent has been developed. The model adopts an optimization-based systems approach to field development, which begins with predicting reservoir performance, and subsequently incorporating economic parameters to determine the number of wells that will yield the maximum monetary value of the field. Artificial neural network (ANN) technology as a tool is increasingly becoming popular for use in reservoir engineering applications such as reservoir characterization and prediction of enhanced oil recovery performance due to its relatively fast, computationally cost-effective, and reliable delivery compared to other tools such as reservoir simulators. In this study, ANN technology is applied to the field development optimization model in order to reliably predict the optimum number of wells for any gas field development project within the specified reservoir and economic parameters. In this regard, an ANN expert system has been developed and several data sets containing relevant parameters have been used to train and test the developed ANN system. At the end of the study, the ANN system showed considerable effectiveness and robustness in being able to predict the optimized development pattern of a natural gas field.