Architectural patterns and characterization methods of fault-controlled karst reservoirs in deep carbonates

Fault-controlled karst reservoirs， recognized as important deep carbonate reservoirs in the Tarim Basin， exhibit diverse morphologies， complex architectures， and extremely strong heterogeneity under the influence of strike-slip faults and karstification. The study investigates the Ordovician fault-controlled karst reservoirs in typical units including S80， S65， T705， and S99 within the main production area of the Tahe oilfield， Tarim Basin， and examine the characteristics， primary controlling factors， architectural types， and spatial distribution of these reservoirs， using data from outcrop and core observations， drilling， logging， and seismic surveys， and guided by theories and methodologies including karstology， structural geology， geophysics， geostatistics， and production dynamic analysis. Accordingly， we develop methods for the seismic prediction and geological modeling of these reservoirs. The results indicate that the Ordovician fault-controlled karst reservoirs are fracture-cave systems with varying spatial architectures that are formed by the dissolution of fracture zones within and surrounding the zones of fracturing and faulting. The scales， morphologies， and architectures of these reservoirs are governed by the properties， styles， grades， and dissolution intensity of faults. An analysis of the geometry， mechanical properties， and associated structures of strike-slip faults reveals that the strike-slip faults within the Ordovician fault-controlled karst reservoirs primarily exhibit three morphologic types on profiles （i.e.， vertical， positive flower， and negative flower structures） and six planar architectural patterns （i.e.， linear， en echelon， X-shaped， oblique， horsetail， and parallel patterns）. The reservoirs are better developed at the intersections， inflection points， overlapping parts， and ends of faults， with the karst scale positively correlated with the fault grade. Along the same strike-slip fault， fault-controlled karst reservoirs exhibit varying characteristics in fault segments with different properties. By employing attributes from spectral decomposition and seismic inversion combined with the developmental patterns of the fault-controlled karst reservoirs， we develop a comprehensive constrained probability volume （CPV）， and training images of these reservoirs are generated through simulations with a target-based approach. In addition， a geological modeling method integrating multi-point statistics and multiple constraints is established， allowing for the fine-scale characterization of the morphologies， scales， and architectural characteristics of fault-controlled karst reservoirs. Geological modeling results effectively reveal that the reservoirs mainly occur along faults and are better developed in parts closer to faults. Validation against the reservoir thickness encountered by sparse wells indicates that the model established using the proposed method increases the drilling coincidence rate from 71.6% to 85.8%.