Integrated Full-Field Review of Betty Field for Building Predictive Models To Reduce Uncertainty in Reserves and Identifying Future Potential

摘要

A multi-discipline integrated full field review (FFR) was conducted for Betty field, offshore Malaysia, in order to build a set of 3D predictive numerical models from multi-scale geological, seismic, petrophysical, reservoir and production engineering data. This study reassessed field stock tank oil initially in-place (STOIIP) volumes and remaining reserves, determined infill drilling potential and identified opportunities to improve both short-term and long-term field performance. Betty field comprises of multiple stacked, laterally continuous, vertically heterogeneous reservoirs. Some reservoirs have experienced relatively high recovery factor (RF) to date (I.e., >69%), while others have underperformed (RF<15%). One of the priorities of this study was to resolve these anomalies. Detailed evaluation of the core identified five predominant lithofacies in each reservoir. Horizons, which were interpreted from 3D seismic and tied to the well logs, formed the framework for the static model structure. The neural networkbased lithofacies analysis for all the wells enabled distribution of the lithofacies data in a 3D geocellular model, which significantly improved the accuracy of the rock property distribution in the reservoirs. Vertical trend functions based on electrofacies logs were input to control the facies content of each layer, and the resultant facies model was used to control the porosity distribution using Sequential Gaussian Simulation (SGS) throughout a fine grid (11 million cells, 1-ft. layers). SGS was also used as co-simulation for permeability, coupled with vertical and horizontal variograms honoring the appropriate facies proportion in each layer. The static model was upscaled (to 600,000 cells, 6-ft layers) after correlating the lithofacies from well to well with over 2800 geological markers to preserve vertical heterogeneity. The initial saturation distribution determined from gravity-capillary equilibrium and a single J-function anchored at the base of the water-oil transition zone, was essentially corroborated by petrophysical analyses. Representative drainage/imbibition relative permeability curves were established from the available data and deployed for dynamic modeling. As a testament to the integrity of the data, the technical interpretation and overall approach that was used, it was found that almost 75% of the 66 historical completions were essentially matched for 27 years of history after making trial runs of the dynamic model without any adjustments to the static reservoir description.