Summary Recent advances in software and hardware technology have made possible the development of field-scale, fully compositional CO2 flood simulations capable of capturing areal variations in performance on an individual well basis. This paper first describes the general methodology developed, including key elements used to construct such large models, followed by the validation of this approach. Two modeling studies conducted on the Wasson Denver Unit are then presented to highlight several of the significant outcomes realized so far: high-grading the profitability of new CO2 projects, pinpointing best well candidates to return to production, identifying infill and horizontal drilling locations, and identifying and quantifying injectant losses. Introduction The Denver Unit is located in Yoakum and Gaines County, Texas. The on-going CO2 project in this Unit is one of the world's largest. Each day, more than 500 million SCF of CO2 are injected into, while 39,000 barrels of oil are produced from the 1500 wells within the 21,000-acre project area. Since mid 1984, incremental oil recovery attributable to this tertiary process has exceeded 53 million barrels. Well-planned field operations and careful surveillance have both contributed to the success of this project. As of January 1994, CO2 developments of the Oil Column, the historical producing interval of the San Andres formation. have been relatively complete; the largest untapped reserves within the Unit are the 650 million barrels of paleo residual oil accumulated below the Oil Column. To profitably operate the on-going floods and to continue expansions of new CO2 developments under current oil prices, a much greater engineering and surveillance effort than ever before is required. As part of this effort, field-scale simulations have been employed to provide well-specific forecasts - a key piece of information which can help uncover flood improvement opportunities and high-grade new CO2 projects. An equation-of-state, compositional simulator has been chosen in the study because it can provide more reliable predictions of CO2-oil phase behavior than a black-oil-type model. Recent advances in simulation technology have made this effort possible. A study by Hill et al. has indicated that the simulator employed in the present study can provide reasonable predictions of the production response of a CO2 pilot in the South Welch Unit. In this paper, basic building elements of the modeling methodology such as reservoir characterization, phase behavior, relative permeability, and the process of validating the models are described first. Subsequently, key results of two field-scale simulation studies are summarized to highlight the impact of simulations on the field performance. The first case illustrates how we utilized a "TZ model" to design and high-grade CO2 flood expansions to recover residual oil in the Transition Zone. The second case demonstrates the use of a "Battery 1 model" to improve the on-going CO2 developments.

No abstract

The profitable conversion of off pattern water injectors to production wells is possible in CO2 floods. An initial water production period often precedes the breakthrough of oil into such injector-to-producer (ITP) conversions. At the Wasson (San Andres) Denver Unit in West Texas, time to economic oil production and initial economic oil production rates may be estimated by simulation or analog methods. Performance data from more than 25 wells is discussed in this paper. Introduction An irregular injection pattern, established during waterflooding of the Denver Unit, is inappropriate for a CO2 flood where the high cost of the EOR injectant requires its effective use in maximizing recovery. In developing regular inverted nine-spot CO2 flood patterns prior to CO2 startup, many waterflood injectors were temporarily abandoned because they were situated in pattern producer locations. Regularization of the CO2 flood patterns by converting former water injectors to oil producers was desired to enhance the sweep efficiency. When CO2 is injected into the reservoir at miscible pressures, it mobilizes waterflood residual oil trapped by oil-water relative permeability effects. Oil can then be transported through water swept regions into former water injectors. Thus ITP conversions are technically feasible. Although oil almost certainly will flow into the ITP wells, the operator must answer the economic questions of when to convert the wells and how much oil it will produce. At the Denver Unit both simulation and analog methods have been employed with success to answer these questions. Background The Denver (San Andres) Unit is one of the world's largest CO2 enhanced oil recovery (EOR) projects. P. 469^