A Novel Approach to Initialize a 3D Dynamic Model Using Saturation Height Modelling

Abstract Saturation Height Model (SHM) definition is a primary activity for estimating volume of hydrocarbon initially in place. SHM describes the variation of the water saturation (SW) as a function of the main rock properties (porosity and permeability), the fluid properties (density and interfacial tension), the rock-fluid interaction (wettability) and the distance from free water level. Algorithms defining these models can be directly implemented into 3D reservoir modelling for continuous SW profile calculation, once the input properties have been propagated. The computed SW honors the saturation profiles at wells and its variation in the inter-wells space is driven by the distribution of input properties. Conversely, most used commercial platforms for dynamic simulation do not accept equations for saturation calculation. We propose a method to initialize a 3D dynamic model in a way consistent with the 3D static one. Consistency of the defined reservoir model, and relevant estimate of volume in place, can be verified after field production start-up through the behavior of production wells. A workflow was implemented to define a discrete number of reservoir SW classes (SatClass) with homogeneous behavior as a reference for volume calculations. The workflow consists of the following steps: 1) Definition of a number of SatClasses corresponding to the most frequent permeability values (modes) in a permeability distribution. Each SatClass represents a reservoir facies described by a range of Rock Quality Index (RQI) values. 2) Characterization of SatClasses with permeability and porosity values, from relevant modes, and a synthetic capillary pressure (PC) curve, calculated for the previous properties using the SHM algorithm. 3) Calculation of SatClass-based permeability and porosity discrete profiles at wells. 4) Calculation of SatClass-based Saturation Height Discrete Model at wells (SHDM). 5) Satclass- based comparison of SHDM and SHM at wells. If the single value from SHDM and the most frequent value (the mode) from SHM distribution are close, the comparison can be considered acceptable and the SatClass correctly characterized; if not, properties are revised and verified iteratively. Defined SatClasses are used to initialize the 3D dynamic model through the PC curves most representative of the reservoir behavior. Main achievements of proposed method are: 1) Defining an initial dynamic hydrocarbon volume consistent with the one from 3D static model. 2) Validating the reservoir properties distribution in inter-wells space through the consistency between the dynamic model simulation and the observed dynamic response (history match). To assess the workflow, it was applied to an oil field in the Norwegian offshore. SHM was built on core PC data from appraisal wells and propagated to horizontal producer wells targeting heterogeneous fluvial deposits. Approaching 3D dynamic model initialization with the proposed method makes more robust the choice of PC behaviors as input for reservoir characterization.