Abstract: Fracture Prediction From 3D Strain Variations Using New Kinematic Flow Algorithms for Hanging-wall Deformation
GIBBS, A., P GRIFFITHS, T MURRAY, R OSBORN and N SALTER Midland Valley Exploration Ltd., 14 Park Circus, Glasgow G3 6AX, UK
Structural modeling, when carried out concurrently with interpretation, provides significant benefits to the interpreter. Alternate picking strategies can be assessed and the interpretation validated. With prepared workflow strategies (figure 1) this process can be carded out within the time frame scheduled for the picking and results in a model with significantly reduced error. Post-interpretation specialist studies are improved by using an inherently better model.
Until recently the deformation algorithms used in such in analytical strategies in 3D have been relatively simple extensions of either an incline shear or a flexural slip approach. These techniques were designed for 2D use as essentially constructional rather than kinematic tools and as such are difficult to validate against each other or against real rock behavior. Ultimately this leads to requiring specialist knowledge from the interpreter or over generalized assumptions about the process.
In this paper we describe a new class of algorithms which are process based. In end member cases, solutions are included which correspond either to flexural slip or incline shear depending on the boundary conditions set during the analysis. The boundary conditions can be varied as a continuum and this allows the interpreter to rapidly assess and understand the significant deformation processes applicable to the study area.
The new approach can be validated in a number of ways. Firstly, sensitivity can be compared directly with that of older constructional techniques. Secondly, it is possible to set up the analytical process in such a way as to minimize misfit of the restored components or hanging-wall strains predicted from the analysis. The modeling can be validated relative to geo-histories assumed from analysis of other data. Furthermore, the authors propose that the new algorithms allow a much more rigorous validation procedure where the prediction of hanging-wall properties. These can be compared directly with independent observations of well or seismic attributes.
The models are constructed using a triangular mesh for surfaces or a tetrahedral mesh for volumes. Monitor distortion of these meshes during deformation allows strain to be computed interactively. Strain can be displayed in a number of ways for example as magnitude, dilation, principal strain or strain ratio as required. The selected strain parameter is then saved as a model attribute and color mapped to the display (figure 2). Direct comparison of the predicted strain can then be calibrated with well or other observations. With the strain model interactive in the display flow and shear parameters in the deformation algorithm can be “tuned” to provide minimum strain solutions and for sensitivity analysis.
We describe how this analytical approach can be used to predict poorly imaged components of the structural model and to model either fault geometry or horizon geometry as a guide to interpretational picking strategy. The computed variation in deformation attributes, through the 3D volume, provides a tool for predicting and directly imaging rock properties which are a consequence of the structural history. This technique allows a direct prediction of possible fracture or rock damage zones throughout the reservoir A case history is illustrated where detailed structural modeling using this approach was integrated with well and core studies to predict rock properties throughout the volume away from well control.
AAPG Search and Discovery Article #90937©1998 AAPG Annual Convention and Exhibition, Salt Lake City, Utah