Correlations between porosity,Φ,and permeability,k,are often tested for sedimentary rocks in relation to petroleum geology and reservoir characterization.A general trend of increase in permeability with porosity can be expected.However,the effects of grain size,packing,compaction,and solutionldissolution processes related to development,preservation or loss of primary and secondary porosity can lead to a wide variety of relationships between permeability and various fonns of porosity.In the present work,general relationsllips have been tested for a very large data set(578 samples).Tile correlation coefficient for the logarithm of permeability arid porosity was R=0.42.In addition to permeability and porosity,the information for tlb data set included mercury injection capillary pressure,PC,curves.The Thomeer model was used to fit the PC curves.Tlle fitted curves were defined by the two parameters of the model: the pore structure factor,G,and the threshold pressure,PT,given by the model at 100 % saturation.This procedure avoids the problem of defining the entry pressure from PC data especially if there is a knee associated with small or vuggy samples.P,provides an estimate of the largest connected pore throat size in the high saturation range and G represents the distribution of pore throat sizes with respect to their control of mercury invasion.As G increases,the fraction of the volume of the pore network that is dominant in determining permeability decreases.Of the 578 samples,PC data points for 463 samples could be fitted by the Thomeer model with R≥0.85.Distributions of k,Φ,PT,and G are presented for these data sets.The pore structure dependence of k-Φ correlations is illustrated explicitly in tenns of G and PT.Permeability correlated strongly with PT(R=0.82).This correlation was Improved by inchdmg Φ(R= 0.87),and further improved by including G(R=0.93).The Swanson parameter,Φ(SN/Pc)_(MAX),which iucludes Φ,PT,and G implicitly and is obtained duectly from the measured PC curves,gave the best correlation(Re0.95).Application of the curve fitting parameters demonstrates the advantages and disadvantages of fitting experimental data to mathematicd models for the purpose of developing generalized correlations of petrophysical data.
展开▼
