| Popis: |
Gas injection as an EOR method is being considered for the giant Al Shaheen field, located offshore Qatar. The field is characterized by large lateral variations in fluid properties; the oil gravity ranges from 16 to 38 °API and significant variations in initial GOR and saturation pressure are observed. An extensive experimental program aimed at establishing the miscibility behaviour for this complex fluid system has been performed. Experiments included a range of gas injection tests such as swelling and multi-contact miscibility tests, enrichment studies and slimtube measurements. The miscibility behaviour across the range of oil gravities has been very well captured with a single equation of state (EOS) model, as described by Lindeloff et al. (2008). Slimtube measurements are the preferred method for establishing minimum miscibility pressures experimentally as condensing/vaporizing effects can be captured in this setup. The physical dispersion of the slimtube was characterized by a specially designed experiment using first-contact miscible (FCM) fluids. The results of the FCM experiments were used to determine the degree of numerical dispersion required in a 1D slimtube simulation model to match the measured data. An extrapolation towards an infinite number of cells was then performed to estimate the dispersion-free minimum miscibility pressure (MMP) from the slimtube experiments conducted on live reservoir oil using carbon dioxide as injection gas. In addition to the slimtube simulations, several algorithms for estimating the minimum miscibility pressure have been compared; some are empirical correlations, others are based on analytical gas injection theory using the method of characteristics (MOC) limiting tie-line method and the last one relies on a mixing-cell approach. The advantages and drawbacks of each approach are discussed, and the results are compared to laboratory data. In general, there is a very large spread in predicted MMP using carbon dioxide as injection gas. Empirical correlations generally overpredict the MMP for light oils and underestimate the MMP for heavy oils. Various key tie-line methods using the same tuned EOS model provide very different estimates of the MMP and some of them exhibit convergence problems. The results of the present work suggest that the mixing-cell model provides the most robust estimate of MMP over a large range of oil compositions, subject to EOS description. In essence, there is currently no prediction method that can replace the slimtube experiments. |
