Empirical Models for Oil Rate and CO2 Retention in CO2-EOR projects

Abstract:

CO₂ utilization for EOR and associated storage in depleted oil fields is emerging as an attractive business proposition because of new provisions in the US tax code, in addition to its promise as a bridge technology for large-scale deployment of CO₂ sequestration to expedite emissions reduction. Many medium and small operators own sizable assets that could be attractive candidates for CO₂-EOR, especially in the US Midwest. However, many of these entities lack the data and/or technical resources to carry out detailed model-based assessments of key performance metrics (e.g., incremental oil recovery, CO₂ retention in the subsurface). Traditional reservoir engineering screening models such as the one proposed by Claridge (1968) are restricted to standard patterns such as the 5-spot. More recently, Azzolina et al. (2015) have proposed a set of statistical models for these metrics based on data from a number of field projects.

The goal of this paper is to propose two new empirical models – one for the incremental oil rate following CO₂ injection, and the second for the fraction of CO₂ retained in the reservoir. To normalize the approach, both are formulated as a function of the hydrocarbon pore volumes of CO₂ injected. The oil rate model is expressed in terms of four parameters: PV0 (injected pore volumes at which oil is first produced), PV1 (injected pore volumes at which oil rate reaches a peak), alpha (slope of linear growth of oil rate between PV0 and PV1) and beta (exponential decline of oil rate from the peak value at PV1). The CO₂ retention model is expressed in terms of two parameters: F_max (maximum retention factor), and gamma (coefficient of Michaelis-Menten type decay in fraction of CO₂ retained).

The presentation will discuss how the models are applied to field data from three representative CO₂-EOR projects in small compartmentalized units with a few wells each that are located within the Northern Michigan Pinnacle Reef Trend. Despite considerable scatter in the field data resulting from significant variations in injection rate, both models are successful in capturing the trends in oil production and CO₂ retention. Of particular importance is the physical significance of these fitting parameters and how that can facilitate their application in a predictive mode beyond simple calibration to historical data.