Multi-dimensional Constrained Geo-Modeling of Carbonate Reservoirs Based on Seismic Inversion and Facies-Controlled Collaboration

To address the challenges of strong heterogeneity and low well control in carbonate reservoir modeling, this study proposes a seismic inversion-constrained facies-controlled geological modeling framework using the JM Oilfield as a case study. The methodology integrates the high vertical resolution of 1D well data, lateral continuity of 2D seismic inversion, and depositional facies distribution patterns to achieve multi-source data synergy through three core innovations. First, Lithofacies probability maps are constructed based on depositional patterns and seismic inversion attributes, quantifying lateral constraints for facies distribution and reducing model randomness in well-sparse areas. Second, a well-seismic joint variogram model is established, where lateral ranges derived from seismic inversion attributes characterize planar continuity, while vertical ranges calibrated by high-resolution well data capture subtle vertical heterogeneity, enabling precise reservoir heterogeneity characterization. Third, a hierarchical multi-scale modeling workflow is implemented, integrating lithofacies, porosity, and permeability models through sequential constraints, seismic inversion volumes, and multidimensional variogram models. Blind well validation demonstrates a 90% match rate between model predictions and measured porosity logs, significantly enhancing reservoir prediction accuracy. Grounded in geostatistics and driven by geological-geophysical collaborative constraints, this framework transcends single-data limitations and excels in multi-scale data fusion, particularly in well-sparse regions. It provides robust technical support for high-precision modeling of strongly heterogeneous carbonate reservoirs and serves as a reference for analogous hydrocarbon reservoir characterization.