Middle-Jurassic Carbonate Velocity Models in Saudi Arabia: Development and Validation

Abstract

Abstract Carbonate reservoirs are complex and require care in choosing an appropriate velocity model (Rafavich et al, 1984; Wang et al, 1991). Saudi Aramco has recently conducted a study to understand the effect of saturation and pore structure on the velocity of a particular Middle Jurassic carbonate reservoir in Saudi Arabia (Ayadiuno et al, 2015). As part of the study, several velocity models were empirically developed as a function of pore fluid type, density, and porosity. The study results suggested that at low velocity values, the obtained brine saturated core model correlated with Pickett while oil saturated core model correlated with Castagna (Ayadiuno et al, 2015). However, at high velocity values, the obtained models didn’t correlate with the published models. In addition, the study suggests that rock’s compressibility is less dependent on pore fluid type when the aspect ratio of the pores is high (Ayadiuno et al, 2015). In this paper, three core-derived velocity models, developed as part of the earlier study (Ayadiuno et al., 2015), at 100% water saturation were extended and validated on the Middle Jurassic carbonate reservoir across four fields in Saudi Arabia. These models are: (1) compressional velocity as a function of total porosity (Model A), (2) compressional velocity as a function of bulk density (Model B), and (3) shear velocity as a function of compressional velocity (Model C). A total of 11 wells from four fields were selected to validate the velocity models. Fluid substitution was performed using log data and water as in-situ fluid to reduce prediction uncertainty related to the differences in saturation fluid type. In addition, we addressed the difference in scale between data used to develop the models (core plugs) and data used to validate the models (well logs). Our findings are summarized as follows: (1) Model A, velocity as a function of porosity, is less sensitive to changes in rock stiffness and thus, is more reliable than Model B, (2) Model C, which predicts shear velocity, shows an overall better estimation of shear velocity than publicly available carbonate models from Castagna et al, (1985) and Pickett (1963). Thus, velocity models obtained from the study are more applicable to the four fields than published (Castagna and Pickett) models. The evaluation of these internally-developed carbonate velocity models across different fields shows a promising result in decreasing the uncertainty of velocity estimation in the Middle-Jurassic carbonate reservoir in Saudi Arabia. This work emphasizes the importance of assessing factors that affect acoustic properties, as well as the importance of multi-scale integration between core-plug measurements and logs.