The effect of scale on the petrophysical estimation of intergranular permeability

An examination of empirical relationships for predicting permeability from porosity has revealed a significant scale dependence that has to be accommodated if integrated reservoir studies are to secure maximum benefit. If this is not done, the predicted permeability can be seriously in error. for example, case histories show that predicted permeability can be less than 50% of the benchmark value if a porosity-permeability relationship established at the core scale is indiscriminately applied to well logs. The errors are governed by the reservoir character, e.g. cyclically or monotonically distributed reservoir properties: they depend to a lesser extent on the way in which log-response information is used in defining the upscaling procedures. The errors increase where a permeability algorithm established at the core scale is indiscriminately applied at the grid-cell and reservoir zonal scales. A matrix of normalized cross-scale errors, generated by benchmarking against known permeability values, has allowed the effects of indiscriminate scale transgression to be quantified and compared for different situations. The resulting errors in permeability, as predicted from porosity, range from about -80% for indiscriminate transgressions upscale to around 400% for transgressions downscale. These observations have stimulated the development of procedures for using groundtruthing core data to estimate permeability at the well-log, grid-cell and zonal scales. Predictive relationships established at each of these scales can be markedly different and therefore they must be applied in a manner that is fit-for-purpose. This means that different predictive algorithms can legitimately co-exist. Where this is done, there is a marked reduction in the number of iterations required to generate an internally self-consistent reservoir model. In other words, a fit-for-purpose application of petrophysics leads to a greater synergy between the static and dynamic components of the reservoir model.