DMX-FMX: Bridging the Gaps Between Data Analyses and Combined Data–Model Driven 3D Fault and Fracture Network Modeling - A Novel Approach

Abstract

Abstract In the present paper new original approach is presented that constitutes the link between the analyses techniques of fracture and fault (Discontinuity) networks in surface and subsurface such as obtained from borehole imaging, conventional observations and ctscan on cores, laser scan and drone observations on outcrops, and 3d seismic attribute volumes, and the novel DMX Protocol, a numerical code developed over the last 15 years to converge towards a more realistic Discontinuity Network representation in space, on the other. This new methodology uses the following flow chart: The FMX protocol analyses datasets for mathematical characterizations of distributions of parameters such as size, spacing, orientation and truncations (distribution models such as Normal, Power Law, LogNormal, etc.). These data analyses tools are linked to different type of Data Driven Modeling technologies, already presented during the last 25 years, both in 2d and 3d (FPDM statistical mapping of orientation and density, 3d deterministic surface fitting) in order to obtain realistic 2d and 3d fracture and fault networks directly based and linked to the observations. On the other hand, the DMX 3d modeling protocol (launched in 2019), includes new features such as Fracture interaction, and IXTV truncation, termination and cross cutting in 3d space, all based on newly designed collision algorithms and fracture propagation principles. The DMX protocol furthermore permits modelling at any scale range of unlimited basic 3d fracture shapes, specific 3d fracture morphology, and 3d fracture aperture types. A complete integration between classical geological/geomechanical drivers such as stress ellipse, fault zones with 3d slip vectors, and different fold models (axial plane, fold axis and bedding orientation conditioning), geological assembly modelling such as joint spacing and set dependency, offset/faulting, and probabilistic conditioning of any of the parameters and drivers. The newly presented work flow permits to integrate these technologies so that both the analyzed statistical distributions and the data driven 3d models can be successfully introduced inside the DMX 3d modeling, using its capabilities directly on the data driven models, and extending them into space towards unknown areas (e.g. inter well); I.e. the observed features are the germs of the 3d model itself. In this way the DMX protocol which has the capability of modeling an unlimited amount of combinations of geological parameters in 3d space, is now successfully integrated in the 3d modelling of the data directly observed on outcrops, seismic images and borehole Logs. Case studies of the application of the workflow will be presented of 3d models built using different sets of observations of fracture and fault networks on various scales, as examples of the application of new technologies in Artificial Intelligence to complex 3d modeling.