As the Prudhoe Bay Miscible Gas Project (PBMGP) approaches its fourteenth year of operation, the water alternating gas (WAG) flood is facing new challenges. These challenges typically involve projects that lead to small miscible injectant, or solvent, supply variations. For example, upgrades of the surface facilities can increase the solvent supply available to the PBMGP. Conversely, undertaking new EOR projects or supplying solvent to non-Prudhoe Bay projects will reduce the supply available to the PBMGP.
To accurately assess these projects, a new methodology was developed for forecasting enhanced oil recovery (EOR). This methodology provides increased ease of use and greater reliability in assessing small changes in solvent supply when compared with direct application of large-scale models.
Full-field modeling of miscible injectant (MI) allocation and enhanced oil recovery for the Prudhoe Bay WAG flood is currently performed with large scale-up models. While these models are excellent for evaluating large, full-field decisions related to miscible injection at Prudhoe Bay, they can yield ambiguous results when used to evaluate projects that cause small solvent supply perturbations. The difficulty in resolving the impacts of small incremental projects is a common weakness of large-scale reservoir models.
Through analysis of numerous Prudhoe Bay full-field simulations of large variations in the MI supply, it was found that a two-component analytical equation was excellent for fitting the character of the cumulative EOR impact. The logarithmic function represents the ultimate EOR impact of a change in MI supply, whereas the Weibull function represents the EOR oil rate acceleration (or deceleration) component. The optimum coefficients for the composite analytical equation were found through regression of large-scale simulation results.
This paper documents the methodology used to create the analytical equation as well as the equation's successful testing. The methodology could be applied to other large-scale models that suffer from an inability to accurately evaluate small incremental projects.
Introduction
The Prudhoe Bay Miscible Gas Project is the world's largest miscible hydrocarbon gas flood. The project began in 1987 with the completion of the Prudhoe Bay Central Gas Facility. Expanding considerably since then, the project area now encompasses 149 of the Prudhoe Bay waterflood patterns, with overall miscible injectant supply averaging about 525 MMscf/D. Waterflood and EOR operations at Prudhoe Bay account for approximately 250,000 stb/D in production. Figure 1 shows a map of the current EOR project area and the remaining expansion potential in the Prudhoe Bay field.
The Role of Scale-up Tools at Prudhoe Bay
COBRA is Phillips' scale-up tool for the Prudhoe Bay miscible flood.1 Because three-dimensional finite difference modeling of the miscible process at full-field scale is too CPU intensive, other techniques are required to scale-up the detailed injector-producer pair simulations to the field level. Within the framework of COBRA, each of these injector-producer pairs (or "segments") is assigned an EOR pattern type curve. The assignment of these curves is based on reservoir characteristics, zonal conformance, injection profile splits, etc.
COBRA internally ranks all of these patterns and allocates the solvent based on some key input parameters. These user inputs include the overall solvent supply, facility limits, development strategies, and field-wide solvent allocation strategy. These inputs to COBRA are pictorially represented in Figure 2.
The Role of Scale-up Tools at Prudhoe Bay
COBRA is Phillips' scale-up tool for the Prudhoe Bay miscible flood.1 Because three-dimensional finite difference modeling of the miscible process at full-field scale is too CPU intensive, other techniques are required to scale-up the detailed injector-producer pair simulations to the field level. Within the framework of COBRA, each of these injector-producer pairs (or "segments") is assigned an EOR pattern type curve. The assignment of these curves is based on reservoir characteristics, zonal conformance, injection profile splits, etc.
COBRA internally ranks all of these patterns and allocates the solvent based on some key input parameters. These user inputs include the overall solvent supply, facility limits, development strategies, and field-wide solvent allocation strategy. These inputs to COBRA are pictorially represented in Figure 2.
