Metamodeling of gas flooding EOR in Pattani Basin, Thailand

• Published in: Journal of the Japanese Association for Petroleum Technology Vol. 78; no. 6; pp. 508 - 519
• Main Authors: Sato, Kozo, Vilcáez, Javier, Attavitkamthorn, Vitsarut
• Format: Journal Article
• Language: English; Japanese
• Published: The Japanese Association for Petroleum Technology 2013
• Subjects: Gas flooding EOR; Metamodeling; Pattani Basin; WAG
• ISSN: 0370-9868; 1881-4131
• DOI: 10.3720/japt.78.508

Metamodeling is a powerful tool used to create analytical relationships between input parameters and reservoir simulation responses. Once created, these analytical relationships called metamodels can be used instead of time consuming and costly reservoir simulations. This research explores the feasibility of using metamodeling for the assessment of gas flooding Enhanced Oil Recovery (EOR) in Pattani basin where performing individual simulations for thousands of reservoirs is practically impossible. Two types of metamodels were created, a parametric and a nonparametric metamodel. The parametric metamodel was created using design of experiment techniques and regression analysis, whereas the non-parametric metamodel was created using alternating conditional expectations and local linear smoother fit with varying window width. Monte Carlo simulations using actual field data were performed to verify both, the created metamodels and the strategy adopted in this study to complete the fluid and petrophysical properties data required to perform reservoir simulations. Based on the results obtained using the created metamodels, it is concluded that Double Cycle Water Alternating Gas (Double Cycle WAG) injection is the best gas flooding EOR method when CH4 is used. This method provides a 4.4% of Original Oil in Place (OOIP) incremental oil recovery over waterflooding. Whereas, when CO2 is used instead of CH4, Simultaneous Water Alternating Gas (Simultaneous WAG) injection is the best gas flooding EOR method. This method provides a 7.8% of OOIP incremental oil recovery over waterflooding. Our results prove the feasibility of using metamodeling to deal with thousands of reservoirs under lack of detailed data on fluid and petrophysical properties of reservoirs of Pattani Basin.