Permeability Estimation Using Spectral Induced Polarization Measurements

Many different models have been developed for the hydraulic permeability of<br>sedimentary rocks and soils. Generally, the models fall into two basic categories: fluid flow<br>through equivalent tubes and fluid flow around grains. Although different in approach, both tube<br>and grain models have been observed to produce reasonable results. The major obstacle to in-situ<br>permeability prediction is accurately estimating the relevant model parameters such as porosity,<br>specific surface area, tortuosity, and grain/pore size distribution. Spectral Induced Polarization<br>(SIP) measurements are sensitive to many of the same micro-geometricalparameters which are<br>required for permeability prediction. The objective of this paper is to test a new methodology<br>for inverting broadband SIP measurements for micro-geometricalparameters and then to use the<br>derived model parameters to estimate the sample permeabilities. A series of laboratory<br>experiments were performed using five different sandstone samples, for which the hydraulic<br>permeability, porosity, specific surface area, and SIP response was measured over the frequency<br>range of 10” to lo6 Hz. The SIP data were then used to independently estimate the formation<br>factor, porosity, specific surface area, surface fractal dimension and the grain size distribution of<br>each sample. As has been previously demonstrated, we observe that the IP response at a fixed<br>frequency increases with the specific surface area of the sample. However, permeability<br>estimates based on surface area alone are inaccurate. Much more robust permeability estimates<br>are obtained when using the grain size distribution obtained in the inversion of the broadband SIP<br>data.