Petroelastic Model PEM for a Highly Heterogeneous Cretaceous Reservoir in Middle East

Abstract

The objective of this project is to simulate elastic logs (sonic P, sonic S and density) through a Petroelastic Model (PEM) for a complex lithology reservoir in the Middle East, that later will be used as input for a new 4D seismic feasibility study. A log conditioning (despike, depth shift, hydrocarbon correction and normalization) and comprehensive petrophysical analysis was first performed, to obtain lithology volumetric, porosity and saturation, that later were used as input for the PEM. Some wells with recorded P and S sonic log were used to conduct different cross plots of elastic properties (e.g. Vp/Vs vs. Acoustic Impedance) in order to understand how lithology, porosity and saturation affect the elastic parameters of the reservoir. After understanding and assessing the elastic behavior with the reservoir properties, three approaches to construct a PEM were tested on this reservoir. The first approach used to construct PEM applying Hashin Shtrikman (H-S) mixt, considering the solid part as a mixture of dolomite and limestone and pore space filled with a mix of oil and water. This model is limited because assumes a homogenous geometry of the pores. To address the pore geometry a Kuster Toksoz (K-T) approach was subsequently tested but the challenge was that there was no clear organization of the aspect ratio (either by lithofacies or petrophysical groups) so the original logs were used to control of the aspect ratio trough a fit function. The third approach was to use a function that models the incompressibility model of the frame (Kdry) with porosity. The result of H-S was a good agreement in the low porosity areas but in the porous intervals, it is observed that the velocities were quite high due the effect of the pore geometry that was not properly assessed by H-S. Despite reasonable reconstructions, K-T was limited by the impossibility to apply it to the wells without sonic P and S (uncalibrated aspect ratio) or a fortiori to a 3D grid. For the Kdry vs. Porosity function the result was very successful since the function is not dependent on the pore geometry, and addresses the ratio issue between solid and pore space. Then with the help of the Gassman Equation, the final Incompressibility Mix Module (Kmix) was calculated and a reconstructed sonic P and S were available for all the wells. The PEM was coded in order to deploy over a 3D property model hence a volumetric elastic model was available to assess the feasibility for new seismic acquisition.