Developing a Surrogate Reservoir Model (SRM) using CMG and IBM - PowerPoint PPT Presentation

Developing a Surrogate Reservoir Model (SRM) using CMG and IBM Analytics to Optimize Waterflooding By: Trixie Anne Roque Mentors: Muhammad Zulqarnain, Esmail Ansari, and Mayank Tyagi Louisiana State University Center for ComputaDon and Technology July 27, 2016

Oil Reserves are Formed from Dead Organic Materials Under Extreme Heat and Pressure Source: h*p://www.nzpam.govt.nz/

Oil is Extracted Using a Three-Phase Process • Natural drive – uses the naturally exis<ng pressure in the reserve to extract the oil • Water flooding – uses pressurized water and gas to Source: chinaoilfieldtech.com push the oil to the surface • Enhanced oil recovery (EOR) – uses chemicals to change the proper<es of the oil Source: rigzone.com

Oil is Extracted Using a Three-Phase Process • Natural drive – uses the naturally exis<ng pressure in the reserve to extract the oil • Water flooding – uses pressurized water and gas to Source: chinaoilfieldtech.com push the oil to the surface • Enhanced oil recovery (EOR) – uses chemicals to change the proper<es of the oil Source: rigzone.com

Why Use Water Flooding? • Natural drive only recovers 5% - 10% of the original oil in place (OOIP) from the reservoir. • Water flooding recovers 25% - 30% more of the OOIP. • Without water flooding, we are only recovering a small percentage of oil.

Process Flow Diagram

First Step Involved Data Generation Using CMG - IMEX Reservoir Simulator

Time Series Data on Field Production Rates, Cumulative Oil, and Water Injection Rates were Collected

Second Step Involved IBM SPSS Modeler to Create the Surrogate Reservoir Model

Goodness-of-fit Values were Used to Determine the Best Model

History Matching and Forecasting • Generated data were divided into “inputs” and “targets”. • Inputs were used for history matching. • Once matched, the model was used to forecast future reservoir behavior.

THANK YOU FOR LISTENING!