Waterflood Forecasting for More Accurate EOR Performance Evaluation

Abstract

This paper presents modeling CO2 enhanced oil recovery (EOR) flood performance through the application of dimensionless scaling for both forecasting and surveillance purposes. While the methodology has been used successfully for West Texas CO2 floods for more than two decades, a recent modification in the process enhances the certainty of forecasted tertiary response based on simulation and analog results. The primary focus of this paper is on how this new approach improves the use of analog or observed production history to develop more reliable forecasts for EOR processes. Business units favor analog methods since they are fast, adaptable and explicit. Analog tertiary production response is the incremental oil production over an estimated base waterflood oil recovery. The original formulation, published in a different paper (Simmons and Falls 2005), for the underlying base waterflood was modeled using an exponential decline throughput-based regression fit of historical pattern based performance, but in effective waterfloods, many times the oil production approaches a harmonic decline. In this paper, the impact of waterflood maturity level on analog dimensionless analysis is demonstrated by both simulation and multiple historical waterflood scenarios across the Permian Basin. The authors offer an improved approach to predict base waterflood and consequently tertiary oil recovery response. This method integrates multiple waterflood forecast methods, e.g., hyperbolic and dimensionless. The new approach results in a reduction in difference between simulation and analog forecasts. Also, the estimated final tertiary response using this method converges closer for various San Andres CO2 floods started at different times in West Texas. Finally, the modified analog response is compared against simulation.