Context
Originally Geosiris started to study the development of Geomodelling functions “in premise” with the objective to be used by Geomodelling workflows to build Earth Models from seismic and drilling surveys data in tectonically complex areas.
These functions are embedded in several explorational prototypes : into two releases of the GeoTopoModeler, GTM and PALEO-GTM in collaboration with the University of Strasbourg, into an automatic 3D Grid Builder Jerboa in collaboration with the University of Poitiers, and with an experimental development using Mathlab in collaboration with the University of Aix-Marseille.
These Experimental prototypes are working “on premise” , with a Classical UI . The user defines the RESQML Entities used as Input, specify some parameters and trigger the calculation. The function itself “knows” which type of RESQML Entities should be generated as output. At the and the result is a newly created RESQML Entity. This is the same behavior than the execution of RESQML Activities.
Our goal is to extract these functions dedicated to Earth Modelling and define from this base Key Geomodelling Services which can be used by RESQML Activities in an event base Workflow on the cloud (see the Cloud-native workflow platform Solution)
GEOTOPOMODELER (GTM)
In GTM, the topological relationships between the various horizons and between horizons and faults are defined by using rules, which take into account their geology and their respective ages (Perrin & Rainaud, 2013).
The GTM software was developed through collaborations with the XLIM laboratory (University of Poitiers), which produces tools for formalizing topological relationships and with the ICube laboratory (University of Strasbourg), which provided the CGoGN library and the Schnapps application for software development. The GTM software was tested on several case studies provided by the Total company. Other case studies were provided by Equinor, Petrobras and some other companies so that we could fully check the operability.
PALEO GTM
The Paleo-GTM tool provides a reconstruction of the deposition spaces in which the various stratigraphic units were deposited through time. For each period of deposition, the reconstruction takes into account the offsets generated by the faults that were already present.
Paleo GTM se décline en deux prototypes respectivement en charge de la modélisation 2D et 3D. Ces prototypes réalisent des activités métier indépendantes dont les entrées/sorties sont formatées au standard RESQML 2.2. Les applications multi-échelle fournies par l’Université de Strasbourg permettent de préciser et de visualiser les états du système à différents niveaux de granularité.
The Paleo-GTM tool consists in two prototypes that are respectively used for 2D and 3D modeling. They perform independent business activities whose inputs and outputs are RESQML entities. The CgoGN et Schnapps multi-scale applications allow the end users to display the states of the system at different levels of detail.
Paleo-GTM : view of a meshed horizon and of the associated fault network (Volve prospect )
Implicit Fault Representation
We developed a methodology for modeling faults and zones of seismic noise around them by means of implicit functions with the hypothesis that fault shapes can be mostly represented by cylindrical surfaces. This methodology consists in modeling cylindrical fault by a set of planar strips approximately parallel to the cylinder generatrix and adjusted by least square to N data points. Local and global user defined orientation criteria are used for possibly merging neighbor strips. The implicit function is obtained by smoothing the strips with a CSRBF Wendland function satisfying a partition of unity.
Resulting fault representations can be mutually intersected for constituting a fault network. A complementary methodology is defined for intersecting faults with horizons and determining corresponding fault throw profiles. This intersection methodology applies to all types of faults including those terminating inside a geological block and does not require a division of the horizon into separate patches.
This methodology has been applied on 59 faults belonging to 4 prospects with excellent results. It provides representations faithful to data and facilitates fault networks construction and fault throw profiles determination. The methodology also facilitates fault geometry analysis. Faults having variable orientations can be individualized and divided into homogeneous portions.
Modeling of a fault network using discrete planes (Prospect Volve)
Key Geo-modeling Functions
GEOSIRIS presently studies the possible building of interactive workflows operating atomic geo-modeling functions.
In RESQML, the user of a geo-modeling function, inserts RESQML standardized input data, specifies the values of the function parameters and triggers the computation. The function is itself a RESQML activity that generates an output in the RESQML format.
The GTM and Paleo-GTM prototypes operate functions similar to RESQML activities. For inserting a GTM function in a RESQML workflow, one has to :
- define the function semantics with a name that characterizes the task that the function must perform,
- specify the RESQML entities that the function takes in input and those that it delivers in output,
- extract and adapt the algorithms that should be used by the computation functions or build new ones if necessary.
The information that a function has to use must be defined in the ETP v1.2 format in order to be read by the GABBRO data server.
The services that we plan to build in priority, will operate the following functions :
- Fault surface modeling using MBA triangulated surfaces
- Horizon modeling using MBA triangulated surfaces
- Quality control of the resulting surfaces with respect to the input data,
- Implicit representation of fault volumes
- Quality control of the resulting volume representations with respect to the input data,
- Intersection between implicit fault volumes.
The computation protocols operated in the Geosiris prototypes will be reused and complemented in order to be operated by geo-modeling functions that are presently developed.