83rd EAGE Annual Conference & Exhibition

Despite different concerns, fossil fuel continues to be the world’s most significant source of energy. To meet the increasing demand, the petroleum industry has turned its focus towards enhanced oil recovery techniques, which guarantee 30 to 60% of oil recovery after primary and secondary productions. One of these techniques is foam injection. Foams are preferred injection fluids over water or gas because of their low sensitivity to gravity and permeability heterogeneities that increase sweep efficiency. Due to the thermodynamic instability of foams, this recovery strategy is not extensively used. The purpose of this study is to develop long-lasting nanoparticle-stabilized foams using current breakthroughs in nanoparticle engineering. In this study, we investigated foam stability of ionic surfactants with and without silica nanoparticles at ambient and increased temperatures, as well as bulk foam experiments with air, nitrogen, and CO2. We devised a novel reverse method for calculating foam quality at foam’s half-life. Surfactant screening revealed that surfactant solutions may produce foam in a variety of reservoir conditions, and nanoparticles can improve foam stability when used at the optimal concentration. Core flooding experiments and numerical modeling in CMG were used to validate the results of the screening and bulk stability tests.

The conventional joint migration inversion method uses a velocity model and an angle-independent reflectivity model to parameterize its solution space, and it faces the amplitude-versus-offset (AVO) challenge. We propose a new joint migration inversion method that uses the two-way wave equation and a wavefield separation technology for wavefield simulation, and our new method overcomes the AVO challenge. Different from the conventional joint migration inversion method, our new method parameterizes the solution space via a velocity model and a density model, and it considers all orders of multiples in one go. Our wavefield separation technology is FK-filtering based, and it separates forward and backward propagating wavefields along any user-defined direction. Gradients of velocity and density are evaluated using these separated wavefields over the whole solution space. The models are updated via these gradients along with an adaptive step-length estimation. We use a synthetic example to demonstrate the success of our new method.

In this paper, we introduce a novel upscaling method in Digital Rock Physics for porosity and permeability properties. The upscaling method is based on a machine learning method characterizing quantitatively image textures from 3D X-ray Micro-Computed Tomography (MCT) images. The procedure starts by imaging a core plug sample at a coarse scale, corresponding to a resolution of around 20 μm, to visualize texture heterogeneity. The machine learning method identifies representative texture spatial locations by classification. We extract physically subsets representing each texture and image the subset at a resolution of 1 μm. Then, we run pore scale simulations to obtain porosity and permeability properties for each representative texture. Finally, we upscale rock properties using classification result to populate coarse scale model from fine scale results. We illustrate our workflow results for a carbonate rock sample from an oilfield reservoir in Abu Dhabi.

In Vibroseis acquisition, improving S/N ratio of the seismic vibrator in a wide frequency range as well as increasing acquisition productivity becomes extremely challenging. This paper attempts to use field data to show that a new generation low frequency vibrator (TITAN) is capable of producing seismic data with a broad bandwidth that is as low as 1 Hz. Besides an improved seismic data quality achieved with the new generation low frequency vibrator, a series of simulations demonstrate that this new generation low frequency vibrator potentially opens a door to dramatically improve simultaneous Vibroseis acquisition.

Carbonate Stimulation is a practical approach to increase well productivity. The use of reactive fluids in the treatment helps open a conductive path (wormhole) for hydrocarbon flow. Reactive transport modeling is used to understand this process and optimize the treatment. This process occurs at different length scales such as Darcy’s and pore scale. Thus, the parameters used in the simulation model are subjected to uncertainty, which affects the model’s predictability and stimulation optimization. In this work, we adopted an integrated uncertainty framework that combines a two-scale continuum model, design of experiments (DoE), and Sobol sensitivity method. This approach helps study wormhole propagation in carbonate rocks during acidizing treatment and investigate the impact of input parameters uncertainty on the selected model response such as pore volume to breakthrough and the increase in the effective permeability.

unconventionnal reservoir characterization, cenomanian Turonian bounbary, TOC, carbonate, pull apart, depocenter

A comparison between FreeCable and OBS was recently presented based on modeled data (Manin and Haumonté, 2018): the polarization analysis demonstrated that FreeCable raw data give easy means to naturally characterize seismic events and possibly apprehend the subsurface geology, while the same analysis with OBS indicated that propagation angle and motion angle are contaminated by a variety of noise and distortion effects, making the polarization study unworkable except to detect S-waves. Here the same type of analysis is performed with real data in complex sea and subsurface conditions. The results confirm the modelling outcome and show that this method enables to interpret qualitatively the wave types i.e. its propagation direction and particle motion direction. This study proves the design quality of the acquisition system and its value to perform high-fidelity (4C) interpretation. Theory, workflow, and examples are presented.

We introduce a simple approach for computing the electrical properties of natural rock at full and partial brine saturation using 3D pore-scale digital images. Specifically, we assume that conductive brine is a wetting fluid and uniformly coats the grain surface with the thickness of this coating increasing with increasing brine saturation. The remaining (inner) part of the pores is occupied by a non-conductive hydrocarbon. Examples using Fontainebleau sandstone indicate that resistivity computations produce realistic cementation exponent m values for the original small porosity volume, as well as for altered higher-porosity volumes. At the same time, only higher-porosity volumes yield realistic values for the saturation exponent n due to the presence of thin conduits that connect larger pores in a low-porosity rock. Such fully saturated conduits remain even at low water saturations. We suggest ways of overcoming this limitation. Our results open a way for digitally obtaining Archie’s constants without simulating multiphase fluid flow to ascertain fluid phase distribution inside the pore space prior to resistivity computation.

Seismic-while-drilling (SWD) provides a cost-effective method to acquire borehole geophysical data without interfering with the drilling operations. SWD data can assist in building accurate shallow subsurface models that can be used for geological interpretation. Here, we analyze SWD data acquired in a desert environment using novel wireless surface receivers placed in an areal 3D geometry around the well. Since the drillbit signal is unknown, top-drive and downhole sensors were used to record the drillbit signature. We utilize the pilot from the downhole sensor that recorded a more accurate version of the drillbit signature to deconvolve drillbit gathers. This approach enhanced the data quality compared to a previous study that employed the top-drive pilot. An advanced workflow that involves vertical stacking of drillbit gathers and nonlinear beamforming was applied to further improve the data quality. The first break picks from geophones within 180–500 m from the well were vertically projected, averaged, and smoothed to construct a robust zero-offset 1D velocity profile. To generate a velocity model that considers offset-dependent velocity information, a 3D traveltime inversion was implemented for depth range 190–800 m which yielded an accurate near-surface velocity model.

High-resolution seismic data are required to describe subsurface reservoirs in detail. However, seismic waves propagating through the subsurface suffer from amplitude attenuation and shape distortion of the waveform. The quality factor (Q) is directly related to this phenomenon (i.e., attenuation or absorption). Meanwhile, Q is also an important hydrocarbon indicator for reservoir characterization. Thus, both Q estimation and seismic data attenuation compensation are very important. However, most of the existing methods are based on a two-step strategy, in which the Q-value is estimated first and then attenuation compensation is performed, and there are problems such as the low resolution of the estimated Q-value is difficult to use for reservoir characterization and the attenuation compensation is greatly affected by noise. Inspired by the strong nonlinear mapping capability of deep learning, we propose a single-step simultaneous implementation of interval-Q estimation and seismic data attenuation compensation based on fusion networks. A series of numerical examples demonstrate that the proposed method can obtain Q values that meet the requirements of reservoir characterization and also obtain reliable attenuation compensation results at low signal-to-noise ratios.