THE MISSING LINK BETWEEN PORE-SCALE ANCHORING AND PORE-SCALE PREDICTION

An earlier paper by the authors (SCA 2001-15) discussed the predictive capability of pore-scale network models by using real experimental data as lithological "anchors". The ultimate goal of this approach was to produce an anchored model capable of relative permeability prediction. An additional advantage of having such a calibrated model would be that a wide range of rock/fluid sensitivities could be examined without recourse to additional experiments. Paper SCA 2001-15 presented a preliminary methodology — utilising mercury injection capillary pressure (MICP) data — that could permit the matching of existing experimental gas/oil relative permeability curves. Results demonstrated that a basic fourparameter model was sufficient to reproduce the vast majority of experimental drainage relative permeability curves that were examined. The constrained set of adjustable parameters in the macropore network model comprised: coordination number (z), pore size distribution exponent (n), pore volume exponent (ν) and pore conductivity exponent (λ). Each of these quantities has a clear physical interpretation. However, we also showed that anchoring network models to mercury intrusion data alone was insufficient for predicting relative permeabilities a prior. There was an interdependence of parameters and, consequently, an infinite set of parameter combinations could be chosen that matched the MICP data but gave very different relative permeability predictions. It was concluded that future analysis of MICP data should be performed in conjunction with the analysis of some other independent experiment — an experiment that would give the additional data that could form the “missing link” between anchoring and prediction. We have developed two approaches to evaluating this missing link. In the present paper, we present one potential approach and show how a unique parameter combination can be derived using only MICP data and routinely-measured residual saturations (Srw and Srnw). The methodology incorporates some simple ideas from percolation theory and is first verified using synthetic MICP data and residual saturations obtained from anonymous network simulations. The anonymous data is analysed using our new approach and we find that we can successfully retrieve the unique parameter combination used to generate the synthetic data. Having obtained the relevant parameter combination, we then go on to predict the corresponding relative permeability curves. Predictions are shown to be in excellent agreement with the anonymous data and we then go on to examine laboratory core data using the same approach.