USE OF THE ALLAN VARIANCE FOR CHARACTERIZINGRESERVOIR HETEROGENEITY

摘要

Increased interest in enhanced oil recovery applicationshas placed a greater need to broadly characterizereservoir heterogeneity in a way that can be related tosweep efficiencies and ultimate recoveries.Traditionally, characterization of vertical heterogeneityin static models is carried out using a number of indicessuch as the Dykstra Parsons (DP) coefficient or theLorentz coefficient. However, the DP coefficient inparticular does not give heterogeneity information overa range of scales or subsurface thicknesses for a givenlog or set of core data. Proposed in this work is the useof the Allan variance (AV) to characterizeheterogeneity field-wide from open hole or cased holelogs. A key advantage of the AV over other availableindices is that the AV is presented as a function ofreservoir thickness and converges for a number ofsignal distributions, including random walk, flicker, andnormal, all of which can be used to characterizereservoir heterogeneity. In addition, the AV can be usedto determine the limits of signal averaging, which hasapplications in upscaling. A number of example wellslogs from various fields will be presented along withcorresponding AV plots.