Bulk Density Prediction through a Combination of Probabilistic and Deterministic Calculations for a Better Description of Bachaquero-02 Reservoir, Western Venezuela

Diaz, Arnoldo (PDVSA) | Chourio, Geragg (PDVSA)


In order to get a full petrophysical evaluation from log-based traditional techniques in every location, the formation density is needed in wire-line log measurements; otherwise, with a limited amount of information in terms of porosity values, the reservoir characterization has more uncertainty. That is, the case study of the giant Bachaquero-02 reservoir, there is a lack of Rhob data in the spatial data sets that prevent a good assessment of the storage capacity in the petrophysical model and thus wrong original oil in place estimation. This paper, therefore, presents a solution to this problem; this work develops a methodology for predicting formation density values which establish a link between probabilistic interpretations from multi-mineral solution and deterministic predictions from multiple linear regression with the main objective of seeking a mathematical expression which describes the best fit for the Bachaquero Member and Laguna Member in each location. The manner of estimating formation density can vary according to the available data in well logs, as a first step, this technique uses classic lithology indicators from well logging such as gamma ray, spontaneous potential and resistivity index to calculate the most probable minerals in the rock with the purpose of assessing a probabilistic approach, the second stage is to create a prediction model with surrounding wells, the input data, which is the probabilistic outcome and measured logs, it is trained using a'least squares' regression routine that will find the best fit in the data for bulk density reckoning. A reliable formation density profile according to the lithology of the reservoir was obtained for each well. The model shows more than 0.9 of correlation coefficients between the density measured by wire-line services and the new bulk density reproduced in this method. Particularly, the Bachaquero-02 reservoir has a notorious heterogeneity along the stratigraphic column; the Bachaquero Member has different depositional environment and rock properties in comparison with Laguna Member which has poor quality reservoir rock. This workflow has the ability to incorporate reservoir heterogeneities in the probabilistic module without a problem. 2 SPE-191163-MS