Second EAGE Conference on Pre-Salt Reservoir

We propose a workflow to evaluate 4D seismic signals around Water-Alternating-Gas (WAG) injection wells in a Brazilian pre-salt carbonate field. Using a carbonate rock physics modeling approach introduced previously, we forward model 4D seismic responses in the form of 4D attributes and compare these with attributes from expected 4D signals to quantify the impact of different fluids (water, oil and gas with C02 content) on the reservoir 4D seismic detectability for a high-fidelity OBN acquisition system. We explore the uncertainties in the saturation scenarios via a cascading analysis of multiple WAG injection switches through time. We illustrate some non-unique 4D responses that complicate 4D analyses and infer the minimum saturation changes required for 4D seismic detection in this carbonate reservoir using OBN data.

Automatic seismic interpretation is one of the main applications of machine learning for exploration geophysics. In recent years, an increase in the popularity of convolutional neural networks to perform interpretation related tasks has been observed. Here we develop and train a 3D Convolutional Neural Network for predicting igneous seismic facies in the Santos Basin Pre-Salt. Often one of the main challenges regarding automatic seismic interpretation using neural networks relates to the scarcity or lack of training data. To overcome this problem, we used a sampling strategy to train the network with just a few partially interpreted sections. After the training process, the network was applied in the full seismic amplitude volume, outputting the igneous facies probability for the whole area. The results show good conformity with the input manual interpretation and with well logs, which was not part of the training.

Reservoir scenarios, generated from multiscale data for Non-Matrix across a variety of static and dynamic data types in the Pre-Salt carbonates, are integrated into 3D cross-sectional reservoir model. These scenarios include different geological concepts of matrix and non-matrix types through Single Porosity/Single Permeability (SPSP) and Dual Porosity/Dual Permeability (DPDP) models. Different fluid types ranging from heavy to light hydrocarbon, column heights, and displacement functions are tested. Compositional flow simulations are run to quantify the impact of these scenarios and test the sensitivity to recovery, well rate, water and gas arrivals using primary depletion and secondary recovery of water, gas reinjection, and a combination of both. Moreover, scenarios for miscible CO2 injections are conducted to learn variable impacts on production and reservoir performance.

The study produced credible scenarios with an impact on appraisal and development decisions (Estimated Ultimate Recovery/ EUR basis) for CO2-rich gas cap or high CO2 oil outcomes. It also provides guidance, thresholds, and scenarios to support business decision-making when CO2 is encountered including Gas-cap development feasibility and gas management.

Performing velocity analysis step aiming to estimate velocity of a reflection event, by inverting travel-time curves, demands the extraction of the arrival times of a reflection event in order to obtain its travel-time curve. This extraction is usually performed manually or with techniques which are not automated. For this reason, we propose a technique able to extract travel-time curves of reflection events related to pre-salt structures by reconstructing and comparing their seismic spectra. With our technique, it is possible to find, in a reliable manner, each wavelet and each arrival time of PP and PS reflection events for a pre-salt structure and with OBN data.

Natural open fractures are present in pre-salt carbonate reservoirs of the Barra Velha Formation, Santos Basin. These fractures contribute to excess permeability, i.e. well test permeability significantly larger than core or log derived permeability. Fractures have been interpreted in acoustic borehole images in 19 wells across the Santos Basin, and the P21 fracture density, which is fracture trace length per area, has been derived. Silica content from multimin modelling appears to be a primary control on fracture occurrence, and P21 fracture density correlates with volume of silica. Fractures are also frequent in fault damage zones, above and below cavities, and they are also associated with presence of vugs. Natural open fractures are rare in high porosity rocks with a shrubby or grainy texture in the studied wells, and also in intervals with a conglomeratic appearance in the borehole image logs.

Velocity model building (VMB) in regions containing complex salt geometries and volcanic bodies has long been challenged. The increased uncertainty in surrounding sediment velocities, combined with imaging and illumination difficulties, often make salt body picking ambiguous. Here, 3D salt mesh interpretation, full wavefield inversion (FWI) and tomography, combined with map migration, are tightly integrated to overcome these challenges in velocity model building at an outboard Brazil pre-salt exploration prospect. 3D mesh interpretation enables interpreters to pick salt bodies in creative ways, leveraging 2D/3D QC views. The mesh can then be quickly and accurately modified to capture the new small-/medium-size 3D salt features revealed through the iterative FWI/tomography updates. Finally, map migration is used to re-position salt picks. This integrated approach, tested here at an outboard pre-salt prospect, facilitates rapid scenario testing of salt bodies and reduces cycle time by up to 50%. Resulting imaging uplifts derived from this workflow create business value by allowing easier seismic facies interpretation in the pre-salt and more accurate gross rock volume calculations.

The discovery of giant hydrocarbon accumulations in the pre-salt play, offshore south-eastern Brazil led to large-scale data acquisition, with dense 2D seismic grid and broad 3D seismic data coverage, with more data being acquired currently. More than 100 exploratory wells had penetrated the pre-salt, with most of the acquired data already in the public domain, including logs, cores, and well reports. This rich geologic dataset is being interpreted in the ATLANTIS Project to build a regional and integrated paleogeographic reconstruction of the Brazilian Southeastern Margin. The goal is to de-risk play-elements presence and quality in the pre-salt section, as well as to build predictive models for CO2 distribution in the region.

This project also has the potential to evaluate the remaining potential for economic hydrocarbon accumulations in this important play. Preliminary results indicate that the main, regional controls on accommodation creation and the petroleum systems elements distribution from continental rifting to early drift phases in the region are:

1. inherited Neoproterozoic structural lineaments and Mesozoic continental weaknesses zones,
2. subaerial volcanism pre- and post-continental breakup controlling the development of early seaways and sub-basins, and
3. transcurrent faults zones and consequent transtension and transpression.

Forward stratigraphic models such as Geological Process Modeling (GPM) are rules-based numerical approximations of physical processes simulated across geological timescales. Key business questions and scenario selection can be directly tested by the generation of GPM digital concept models that are tied back to fundamental and consistent processes, with the advantage of rapid transformation to reservoir models to evaluate the dynamic in addition to static impact of alternative scenarios.

This study aims to present a workflow to augment deterministic carbonate modelling by using forward stratigraphic modeling to compare scenarios and generate heat maps of probabilistic simulation convergence and divergence. Through this technology integration, geoscientists and geomodelers can move towards a new way of working that is complementary to traditional methods, with a systematic approach to both filter depositional scenarios by direct testing of geologic feasibility, and to introduce process-based inputs for observed spatial heterogeneity of facies distribution and reservoir quality drivers into an integrated geomodel.

Several oil fields on the pre-salt have high heterogeneity in terms of pore types and scale, with the occurrence of non-matrix pore systems displaying important impact on productivity. Not surprisingly, discrepancies were verified between matrix permeability in 1D petrophysical models and the permeability scale estimated by DST measurements, which are generally understood as related to the presence of dissolution structures formed by karstification processes.

Intending to represent the non-matrix permeability scale in our petrophysical models, Petrobras’s researchers began the development of techniques for permeability estimation, where it is possible to separate the matrix pore systems from the non-matrix and to assign different parameters to each distinct medium. This technique is known as K-image, where attributes extracted from image logs are combined with porosity logs in algorithms that calibrate the models, both with the matrix absolute permeability and DST estimated permeability, and then they are used in a permeability equation which has terms representing matrix with different flow capacities and a term that represents the extra-matrix pore systems.

The application of this technique brought a significant gain in the ability to characterize permeability of non-matrix pore systems on pre-salt oil fields where the dynamic impact of these systems has been observed.

The giant pre-salt Bacalhau field in the Santos basin comprises two distinct Aptian Barra Velha Formation reservoir intervals (BVE200 and BVE100) developed on an isolated lacustrine platform. The reservoirs were deposited onto a complex paleotopography comprising both rifted BVE300 Paraty volcanics and younger volcanic pedestal highs. Bacalhau is a high-pressure field, with a 500 m hydrocarbon column. Hydrocarbons occur in rocks interpreted as being from both shallow water platform top and slope environments. Here we discuss how comprehensive new data from the 2018/2019 appraisal campaign, integrated with existing well data, narrow-azimuth 3D seismic, together with stratigraphic- and seismic-forward modeling provides new insights as to the controls on: 1) rifted and volcanic paleotopography, 2) relative lake-level changes, 3) dominant wind directions, and 4) slope re-deposition on lacustrine carbonate platform, margin, slope development and so reservoir heterogeneity.