Abstract
The 1984 National Petroleum Council (NPC) enhanced oil recovery (EOR) study used five computerized EOR process predictive models to estimate the potential for EOR. These models were reviewed by industry representatives and determined to be accurate when used with representative reservoir data. These models have continued to be a primary tool to assist the Department of Energy (DOE) and other agencies of the federal government to assess policy alternatives for tax reform, environmental regulations, petroleum research, and others areas. It is believed the estimates made by the models are representative for the United States petroleum resource although they may not be accurate for any one specific project or reservoir. Until recently, there has not been a data source that can statistically verify that belief.
The DOE encouraged EOR projects by starting the Tertiary Incentive Program (TIP) in 1978. Details of proposed projects were submitted, and the projects were certified as qualifying projects. Production and fluid injection data from those projects have been voluntarily submitted to the DOE since 1982. This program has produced historical data on more than 100 EOR projects with sufficient data for comparison with simulations using the predictive models. ՛
In this work, the predictive models were run using the average wellhead crude oil price for each year from 1980 to 1991 to estimate the potential production of the projects. The calculated production from the models was compared with the reported actual production until the project was terminated or no more production was reported. The results showed that steam production estimates were reasonable with a 50% confidence level for individual projects, whereas aggregate oil production for a group of projects compared more favorably. A rough estimate of the accuracy of the total aggregate oil production for a group of projects was ±20% at the 75% confidence level. Availability of data for other processes were insufficient for complete statistical analysis. Results for miscible gas projects showed that the carbon dioxide predictive model makes conservative estimates of tertiary oil production but optimistic predictions when the total oil target is considered. The surfactant option of the chemical flood model and the in-situ combustion model made reasonable oil production predictions. The polymer model made oil production estimates much higher than that reported for polymer augmented waterfloods.
The measured accuracy is for the combination of the models and the data. The results indicated that first upgrading the data and then the models would be desirable. Although this study indicates reasonable results for a group of reservoirs or projects, there was low confidence in predicted oil production for individual projects.
