First EAGE Rock Physics Workshop in Latin America

Summary is not available.

Summary is not available.

Summary is not available.

We test the Xu-Payne (X-P) rock physics model in a carefully selected database of carbonate rocks with a wide porosity range, where velocities were measured under dry conditions in the laboratory. We assess the model by using different stiff and compliant pore aspect ratio combinations and calculating the mean percent error (MPE) in model predictions. We also compare the X-P with the Vernik-Kachanov (V-K) model and some empirical models for S-wave velocity prediction.

The model fits the P-wave velocity reasonably well, but it tends to systematically overpredict the S-wave velocity, leading to underprediction of Vp/Vs ratio. We stress the non-uniqueness in the model application, such that numerous feasible pore aspect ratio combinations and their volume fractions result in the same velocity outcomes.

The V-K model for carbonates produces an excellent fit to the data in the entire porosity range, resulting in lower MPE values than the X-P for both Vp and Vs. The V-K model, which is based on more realistic pore shapes, is also easier to calibrate. The V-K model combined with the Gassmann modelling compares favorably against the purely empirical Greenberg-Castagna model for Vs prediction, with the added advantages of accounting for realistic pore microstructure and stress sensitivity.

We study Thomsen parameters δ, ε and γ of fractured reservoirs. Using the Linear Slip model, we discover the properties of 8 and its relationship with ε and y. We derive an analytical formula for δ, from which we found out different nature of the interconnection of δ, ε and γ in the model with empty dry cracks compared to the model with the cracks completely filled either with a solid or liquid substance. We successfully validate our theoretical research on published data of physical modeling.

In a model with dry (empty) fractures, δ tends to ε, and γ turns out to be much less than ε (following our theory, γ tends to one third of ε). The data of experiments are also in excellent agreement with our theory for models with filled cracks. Two physical models with different crack filler are considered: one is the cracks saturated with a liquid fluid (honey), and the other is the cracks filled with a "weak" solid (rubber). In both of these models, γ turns out to be significantly larger than ε (following our theory y>4/3 ε), and δ turns out to be negative.

The study of the scaling of rock porosity has a great importance as (1) porosity is key in reservoir characterization of hydrocarbons, geothermal resources, and aquifers; (2) porosity measurements are usually made at scales that are different to the one(s) of interest; and (3) scaling porosity in heterogeneous media is still not fully understood. Pore space fractal geometry can be analyzed using different lab techniques and/or processing digital images. Specifically, the fractal scaling porosity seems to be one of the ways to go in this topic. The present work proposes a systematic methodology to study scaling porosity in 3D digital images, acquired with X-rays microtomography. It also applies and compares the fractal scaling porosity at three different scales in an andesite sample. We found that the scaling porosity estimates relatively well the image porosity when using the mean pore size in each image. Whereas the porosity estimation of the physical rock sample seems to be better when using the maximum pore size in each image. Consequently, the results indicate that the scaling porosity has the potential to be used for both, calculating the image porosity and the porosity of the actual rocks, if the right pore size is used.