IOR 1997 - 9th European Symposium on Improved Oil Recovery

The technique of fracturing the wells is frequently used in petroleum engineering to improve the oil production. This technique is applied not only to vertical wells, but also to horizontal wells or multi-lateral wells. However, there are no efficient numerical methods to simulate these kinds of wells, especially for the fractured horizontal or multi-lateral wells. The handling of such a well by a reservoir simulator is particularly discussed in this paper.

Horizontal wells have seen a dramatic rise in the range of applications and the number of wells completed during the past decade. However, horizontal wells have been primarily drilled as producers maximising the benefit from the large reservoir contact which such boreholes allow. To date there have been only a handful of reported applications of horizontal wells as injectors. The potential benefit of horizontal injectors could lie in improving the sweep efficiency and enhancing the degree of pressure maintenance characteristics of reservoirs as they enter an advanced stage of depletion in redevelopment projects.

This paper presents novel applications of the Bayesian inversion approach and the simulated annealing method (SAM) for identifying and reducing uncertainties involved in automatic history matching and forecasting processes which are of paramount importance for optimisation of hydrocarbon recovery. The Bayesian inversion approach enables the priori knowledge about the inversion parameters to be incorporated into the objective function to form a posteriori, in conjunction with the likelihood function reflecting the mismatch between the history data and the data predicted from the numerical model. The simulated annealing method (SAM) bas been applied to escape local optima which may still exist in the posteriori objective function, in order to further reduce the nonuniqueness of the inversion solutions.

Subsurface appraisal is a key activity in exploration and production. Large sums of money, an estimated US$ 1 billion a year at least, are being spent by the industry to appraise new discoveries as well as fields that are candidates tor development.

We reconstruct 3-D sandstone models that give a realistic description of the complex microstructure observed in actual sandstones. Pore networks replicating the microstructure of the reconstructed samples are constructed and used as input to a two-phase network model. The network model simulates primary drainage and water injection for both water wet and mixed wet systems. Predicted transport properties for different reconstructed sandstones are found to be in good agreement with available experimental data. The agreement provides a strong hint that it is possible to a priori predict average transport properties from the associated pore-scale parameters.

An approach have been tested to obtain a reservoir estimate of residual oil saturation after gas injection. The methodology is similar to the weIl-known Single-Well Tracer Test that has been used to measure water flood residual oil saturation in-sits in more than 200 field trials since 1968. The gas tracer test was performed at the Oseberg Field, well 30/9-B-03. The actual formation had been gas flooded and was assumed to be at residual oil saturation.

Laboratory experiments have shown that wettability affects fluid distribution at the pore scale and therefore has an important effect on the end-point saturations and the shapes of capillary pressure (Pc) and relative permeability (Kr) curves.

The experimental results of three long core, low rate, reservoir condition gas condensate depletion tests are presented in this paper. The three cases exhibited very different PVT and rock properties. Early production of the condensed phase was observed, with CCS in the range of 0 to 11 %. Condensed phase recoveries varied from a few percent to 25% during these experiments.

Some IOR techniques create regions of three-phase flow. However, three-phase oil relative permeabilities are not generally well characterized and so there may be considerable uncertainty in the estimation of incremental recoveries. None of the many three-phase oil relative permeability formulations appear to fit the data outstandingly over the whole range [1]. Agreement is unfortunately worst in the low oil saturation regime, which is the most important for IOR. In the absence of a good physical model, and as a preliminary step, key characteristics of threephase oil relative permeabilty formulations which interpolate between given two-phase data are considered. In this way we hope to encourage best practice by avoiding those with undesirable or unlikely characteristics. In order to illustrate these conclusions a numerical sensitivity study of the effect of different formulations on the incremental oil recovery from immiscible WAG is presented.

One of the most important issues in petroleum engineering concerns the prediction of gas production during reservoir depletion - either following conventional waterflooding operations or, increasingly, in the early stages of hydrocarbon production.