Fault network modelling of complex faulted structures, those containing hundreds or even thousands of faults, can be an extremely difficult and time-consuming process. Although techniques for mapping and modelling faulted structures have been in existence for nearly forty years, asset teams still struggle to create correct portrayals of such complex faulted reservoirs due to the limitations of the commonly used techniques. We have developed a new approach to fault network modelling, using a new concept of 'fused' fault blocks. The identification of fault-fault intersections is based not on a manually drawn fault network or table of relationships, but rather is derived from the fault surfaces themselves. The calculated intersection lines are then used to truncate faults against each other. Because the truncation information can be stored with the fault model, this process yields a repeatable and easily updatable fault model. The name of our technique 'fused fault blocks' refers to the fact that when a section of a fault is removed, the two fault blocks that had been created by the fault are then fused together, forming a single fault block. The resultant fault model can then be used to create a 3D reservoir grid, one in which the fault geometry has not been compromised, and therefore better reflects the actual structure. The speed of the fault-building process 'seconds or minutes, even for models with hundreds of faults' also allows multiple interpretations, placing the emphasis of the fault network building on the evaluation of the interpretation and the effects of compartmentalization, and not on the manipulation of software.One of the primary goals of computer mapping and modelling techniques is to produce a model that is structurally possible, internally consistent, and a viable representation of the interpretation. Twodimensional mapping techniques first became available in the late 1960s and early 1970s; commonly these mapping algorithms did not use standard contouring rules to create the 2D grids and therefore could produce impossible or unreasonable structures. With the advent of 3D computer graphics in the 1980s, mapping and modelling expanded into the 3D world as well. Sophisticated algorithms for distributing petrophysical properties or facies in 3D quickly became indispensable, as these techniques provide better information for reserve calculation and well planning than simple 2D maps. However, the modelling of complex structures continues to be a problem. Several methods have been developed for fault surface and fault network modelling, each of which has advantages and disadvantages. Most methods have practical, if not absolute, limitations to the number of faults that can be incorporated into a model simply due to the size of the resultant model or the complexity of building the network. Many also have limitations as to the types of fault intersections that can be modelled.Exploration and development continues to expand into increasingly complex and risky areas, where the accuracy of the models becom...

Rigorous, internally consistent three-dimensional subsurface models are extremely useful in interpretation, mapping, well planning, and simulation pre-processing. The geospatial technique to create these models has been in use for several years, and complicated, highly faulted structures (including overthrusts and other multi-valued surfaces) have been modeled quite successfully. Often, however, the gridding process used to create the horizon surfaces required additional control points, and the shape of the overall structure was not necessarily continued from one fault block to another. A new algorithm has now been developed that uses a three-dimensional model of the faulting process itself to restore data to a pre-faulted condition. Displacement on a given fault surface can vary laterally as well as in depth, and faults which terminate within the model volume are of course accommodated. All horizons are used simultaneously in the process of creating the fault displacement model, which eliminates problems with sparse control or narrow fault blocks. The structural surfaces are then calculated in unfaulted space, and the faulting model is used to transform the resulting surfaces back to the proper structural position. Not only is this algorithm significantly faster, but it also mimics the post-depositional faulting process and produces a geologically consistent model. This consistency and integrity mean that greater confidence can be placed in the model, improving volume calculations and allowing placement of wells with greater precision. The reduced cycle time allows a greater range of scenarios to be modeled and evaluated, thus enabling better risk assessment in complexly faulted fields.

No abstract

No abstract

No abstract