Prudhoe Bay: Rebuilding a Giant Oil and Gas Full Field Model

Abstract

The Prudhoe Bay field is the largest accumulation of oil and gas in North America. Because of the size, and the fact that it produces from multiple mechanisms including lean gas cycling, gravity drainage, pattern waterfloods, pattern MI / WAG injection and gas cap water injection for pressure maintenance, a full-field model (FFM) with a rigorous surface pipeline network and facilities model is necessary to answer many depletion planning questions and to evaluate the benefits of largescale or field-wide projects. With additional reservoir description data and production history, the opportunity existed to use current software and hardware to build an improved model. In 2005 an effort began to build a new FFM, including both geocellular and simulation models. This paper discusses the issues addressed at the start of the rebuild in preparation for the history match, the history matching effort, and the transition to predictive runs. Preparation for the history match included setting the objectives for the model, the grid design, generating pseudo relative permeability curves, and implementing parallel processing. The key parameters and data used to obtain a history match are discussed, as are the issues and methodology utilized in constructing predictive cases. Some of the key interactions with the parallel geologic model construction are also discussed. It was demonstrated that a compositional model with approximately 1 million active cells and over 650,000 non-neighbor connections associated with over 1000 structural faults could be run and history matched in a commercial parallel reservoir simulator in a reasonable time. With approximately 25 BSTB of oil, 46 TSCF of gas in place, and over 2500 historical wells, the challenge of building a fullfield model for a field the size of Prudhoe Bay was daunting. A project of this magnitude required excellent up-front preparation and cross-discipline coordination. Introduction The current full-field simulation model of the Prudhoe Bay Field was built in 1995. Due to the size of the field, as well as hardware and software limitations at that time, the simulation grid was very coarse with 60-acre (1617ft X 1617ft) grid blocks, making areal discretization of faults, wells and waterflood/MI injection patterns problematic. As the oil column thins and remaining targets shrink, a model with greater resolution for future project evaluations was required. The decision was taken by the major partner companies to build a new simulation model based upon a new 3D geological and petrophysical model, to take advantage of better hardware, software and workflows. The rebuild effort started in January, 2005. Initially the focus of the work was on the geocellular model build. However, many aspects of the simulation model had to be set before final delivery of the static description to avoid delay of the history matching process. This paper discusses the work that was done in preparation for the history match and the history matching process and results. There is also discussion of the transition and calibration from history to predictive mode.

Compositional Surveillance Program for Assessing Prudhoe Bay Vaporization Process

“…The project benefits were calculated by using a full field, fully compositional reservoir simulation model that captured black oil production mechanisms in the gravity drainage and waterflood regions in addition to compositional vaporization mechanisms in the gas cap and expanded gas cap regions (see Ding et al 2009 for a rebuilt version of the Prudhoe Bay Full Field Model). Accurate prediction of vaporization mechanism required development of an equation of state that was tuned to the laboratory phase behavior experiments in which original Prudhoe Bay oil was subjected to multiple contacts by lean injected gas.…”