ENHANCED FORMATION EVALUATION BASED ON FORWARD-MODELLING OF RESISTIVITY AND POROSITY LOGS FROM FLOW-UNIT ARCHITECTURE SIMULATIONS OF CARBONATE AND CLASTIC RESERVOIRS

摘要

Traditional log analysis is an inverse reasoning process where petrophysical equations are linked with pattern recognition methods to establish the reservoir structure that caused the petrophysical log responses. As an alternative, log responses of hydrocarbon reservoirs can be forward-modelled from a spatial architecture of flow units. Within these units, stochastic variability of porosity and permeability is applied to the computation of a composite fluid saturation profile and a forwardmodel of petrophysical logs. The procedure was implemented in EXCEL, so that multiple simulations could be run rapidly in a transparent computational medium that is accessible to the user. A discrete Markov chain generator creates a petrofacies sequence where transition probabilities can either be hypothetical or drawn from empirical measurements. Each petrofacies is matched by a distinctive flow unit type with descriptors of porosity and permeability. Where capillary pressure measurement data are not available, pore-throat descriptors can be generated either by the Pittman equations for clastics or the Lucia equations for carbonates. When coupled with a free-water level, the interaction of buoyancy pressure with the petrofacies realization generates a fluid saturation profile. Finally, petrophysical logs are forward-modelled from the simulation. A primary purpose for the simulation program is to train neophyte log analysts in the explicit relationships between reservoir architecture and resistivity-porosity logs both as depth-scaled curves and crossplots. The log patterns generated by the interaction of capillary pressures with a variety of hypothetical simple and complex reservoir architectures can be shown on multiple Pickett plots. In more advanced applications, the modelling is used to reconcile logging measurements with interpretations of specific reservoirs that have complex flow-unit architecture.