Permeability and Water Saturation Characterization and Prediction in Wellington Oil Field Using Core Analysis and Seismic Inversion

Abstract

ABSTRACT: Reservoir permeability is a critical parameter for the evaluation of hydrocarbon reservoirs. It is often measured in the laboratory from reservoir core samples or evaluated from well test data. The prediction of reservoir rock permeability utilizing well log and seismic data is important because the core analysis and well test data are usually only available from a few wells in a field and are associated with high coring and laboratory analysis costs. In this work, a joint inversion strategy based on Multi Attribute Analysis (MAA) and Deep Feed-forward Neural Networks (DFNNs) algorithm was applied to predict the spatial variations of key petrophysical parameters (porosity, permeability, and saturation) for inter-well regions. In this method, acoustic impedance (AI) models are computed from post-stack seismic amplitude data by seismic inversion in the time domain with measured log density and velocity as constraints. The obtained results reveal that the proposed strategy, which combines MAA and PNNs, leads to the optimization of lateral and vertical facies heterogeneities and accurate prediction of reservoir parameter distribution. The time slice maps of inverted porosity and permeability at various time intervals indicate a reasonable calibration with the measured core and well log data. The methodology proposed in this study may be considered useful for other basins in Kansas with similar geological settings and anywhere in the world for reservoir characterization, particularly for carbonate Cambrian-Ordovician formations and variable depositional environments. 1. INTRODUCTION Recently, acoustic impedance (AI) mapping, which is derived from the inversion of poststack seismic amplitude data, has emerged as a widely used technique for predicting spatial reservoir properties. It has been proposed that the inverted models of AI, i.e., P-impedance, S-impedance, and density, can be used to estimate lithofacies, petrophysical, and elastic parameters (Singha and Chatterjee 2014; Leiphart and Hart 2001; Pramanik et al. 2004; Calderon 2007; Kumar et al. 2016; Kumar et al. 2016; Walls et al. 2002). Based on the literature, several researchers have developed geological features and petrophysical properties using particular seismic attributes. For example, Russell et al. (2004), demonstrated how seismic attributes may be used to estimate the porosity cube. They employed stepwise regression and neural network approaches to create the porosity cube. De Groot and Aminzadeh (2004) employed the seismic attribute to discover faults, cracks, gas chimneys, and salt bodies in the ground.