Predicting the future performance of oil and gas wells is one of the major tasks reservoir engineers must periodically perform to estimate remaining reserves. This is an important job that should not be relegated to a mere exercise in curve fitting. This is because an engineer needs to select the appropriate method from the suite of available reserve estimation techniques such as material balance, decline curve analysis, etc. Within the decline curve analysis (DCA) technique, there are available models that are suitable for different reservoir types and for different points in the history of the reservoir. Even for the Arps’ model, different reservoir drive mechanisms would require the use of different model types depending on the b factor.

In this presentation, we discuss the use of the Log(WOR) vs Cum Oil plot as an alternative DCA technique for strong water drive reservoirs. The technique is applicable to waterfloods or reservoirs with high water cuts in general. We discuss the theoretical bases for the technique, as well as that of Arps’ method. Both models are used to forecast the future performance of 5 high water-cut wells. The results are evaluated and compared. From our findings, we make recommendations regarding the suitability of each method for use in modeling the cases at hand and why one method is preferred over the other from a technical standpoint.

This discussion highlights why production forecasting is not just a curve-fitting exercise and why the engineer is, and continues to be, an important piece in the reserve evaluation process. Good production forecasting needs an engineer with the education or appropriate training qualifications to make good engineering judgments based on that training and experience. The engineer also has a good understanding of applicable engineering principles to be able to screen available DCA techniques for use in specific applications. You can’t accomplish these with an autofit.