Digital Rock Physics: from 3D Image Data to Simulations for Oil & Gas Applications

Posted on 22 June 2015 by Theo Verbruggen

Learn about our image-based modelling techniques for streamlining routes from 3D scans to simulation-ready models for digital rock physics, and how they can help you with various oil & gas workflows.

Petrophysical characterisation on microCT reconstructed datasets provided by Repsol

Representing a fast growing area in oil and gas exploration, digital rock physics allows researchers to analyse the properties of rock samples, including permeability and pore characteristics. The creation of 3D models from different scan modalities (such as micro-CT and FIB-SEM), can complement experimental testing and generate valuable data for analysis of rock types. At Simpleware we’ve developed solutions for doing more with 3D image data for digital rock physics, with a particular focus on streamlining multiple stages of image visualisation, analysis and model generation for simulation.

Digital Rock Physics

Core sample visualised in Simpleware ScanIP

When working with geological samples, it is possible to create volume images that contain different material phases, enabling comprehensive analysis of grain structure, porous networks and other characteristics. Quantifying these features and using finite element simulation (FEA) and computational fluid dynamics (CFD) allows for a deeper understanding of the potential behaviour of samples; this has applications to reservoir characterisation, oil recovery and hydrocarbon production potential, among many other areas. Analysing samples at different length scales means being able to better understand flow and transport properties for even very small pore connections.

Visualising and processing image data

3D image data can be comprehensively processed

Our solution for digital rock physics starts with visualising image data, acquired as a stack of images that can be converted from 2D pixels into 3D pixels (voxels) in ScanIP. From this image data, different material phases can be segmented into regions of interest (ROIs) using semi-automatic and manual tools, for example to create solid and fluid masks. In addition, filters can be applied to reduce scan noise and better visualise image data, making it easier to view different features.

Options are also available for obtaining quantified data on the properties of samples, including porosity analysis. Mask and model statistics can be quickly obtained for the volume and surface area of data, as well as on tortuosity, connectivity and pore sizes. We have also developed image processing and measurement features that are useful for analysing multi-phase and porous materials, including watershed segmentation for analysing particles, and centrelines. With this approach, it is possible to quickly visualise, segment and obtain statistics from samples taken from 3D scans.

Multi-phase Meshing for Digital Rock Physics

Mesh generated from Berea sandstone sample (Data source: the Rock Physics Network, ETH Zurich)

We have developed image-based meshing techniques that allow robust FE and CFD models to be generated directly from image data with conforming interfaces and shared nodes; models are suitable for digital rock physics simulation. There are several meshing approaches available when working with rock physics data, including a grid-based mesher (+FE Grid), and an adaptive meshing algorithm (+FE Free). The grid-based algorithm developed by Simpleware generates mixed hexahedral and tetrahedral meshes – voxels are effectively converted directly to hex elements and then converted to tet elements at the interfaces for a smooth surface. This image-based approach takes into account partial volume effects, and preserves topology and volume from the original scan. It is also possible to use +FE Free to adaptively remesh data.

+FE Free re-meshes multi-part surfaces while preserving the original geometry generated using a grid-based approach. Mesh inspection tools are also available to ensure that the model exported to FE and CFD solvers is robust. For digital rock physics, models can be created that support multiple phases and enable FEA/CFD of stress and permeability, among other physical stimuli. Being able to mesh multiple segmented parts solves the problem of working with very complex scans when generating models.

Calculating Effective Material Properties

Visualisation of fluid velocity streamlines using +FLOW

As discussed in our post on the Simpleware Physics Modules, our software’s image-based techniques can also be used to calculate effective material properties from scanned samples. This approach is particularly useful when working with porous media or multi-phase composite. A complex heterogeneous material can be substituted with a homogeneous material, whereby effective properties are chosen so that its response to external loads resembles as closely as possible the original material. A built-in finite element solver calculates effective properties and exports results as tensors and fields. Stiffness tensor and elastic moduli can be calculated with the +SOLID module, absolute permeability with the +FLOW module, and electrical conductivity/permittivity, thermal conductivity and molecular diffusivity with the +LAPLACE module.

When using these techniques, it is possible to obtain detailed information on the material properties of rock samples. For example, calculating effective material properties can be useful when analysing scans of different rock types obtained during reservoir characterisation. The intuitive features of the Physics Modules, including convergence graphs to indicate whether obtained results are accurate, and multiple visualisation options, help with analysis of rock samples.

There are many ways in which 3D image data can be used in digital rock physics. Obtaining qualitative and quantitative information from samples provides information that can be used alongside experimental testing of rocks. By integrating multiple digital rock physics tools into a single software environment, we have aimed to make it easier to process and generate models from scan data. With demand growing for digital rock physics workflows to take into account very fine porous structures and more unusual rock types found during exploration, keeping these processes easy-to-use is crucial. The models created in our software are very robust, validated, and capable of being used for different types of rock physics simulations.

To see our software in action, see our webinars.

We are happy to hear questions, so please contact us to learn more about how we can help with your workflow. Or try our ScanIP software suite for yourself with a free trial.