Scaling Up a 900-Million-Cell Static Model to a Dynamic Model

Abstract

Listen

Translate article

This article, written by Senior Technology Editor Dennis Denney, contains highlights of paper SPE 164187, ’Answering the Challenge of Scaling Up a 900-Million-Cell Static Model to a Dynamic Model - Greater Burgan Field, Kuwait,’ by Eddie Ma, SPE, Kuwait Oil Company; Sergey Ryzhov, SPE, Schlumberger; Yuandong Wang, SPE, Petrobras; Reham Al-Houti, SPE, Laila Dashti, SPE, and Farida Ali, Kuwait Oil Company; and Muhammad Ibrahim, SPE, Schlumberger, prepared for the 2013 SPE Middle East Oil and Gas Show and Exhibition, Manama, Bahrain, 10-13 March. The paper has not been peer reviewed. The Greater Burgan field began producing in 1946. It remains under primary depletion with natural waterdrive. Subsurface modeling is an integral part of reservoir management. In 2001, the first comprehensive full-field geological model was built with 65 million cells, encompassing all of the major reservoirs. A reservoir-simulation study (1.6-million-cell dynamic model) was conducted in 2003 by use of parallel-simulation technology. During the last decade, active field-development plans have resulted in major surface-facility upgrades and the drilling of +more than 300 new wells. This paper discusses scaling up the high-resolution geological model and specific problems encountered by the study team. Introduction The Greater Burgan field is in southeastern Kuwait, with an area of 320 sq miles. Fig. 1 shows the five main reservoir units of the Greater Burgan field complex: the Wara sand, Mauddud, Burgan Upper sand, Burgan Middle sand, and the Burgan Lower sand. Areally, the field is separated into three producing areas, Burgan, Magwa, and Ahmadi. Mauddud, the only carbonate reservoir in the sequence, is relatively tight and, together with the extensive Wara shale, acts as a barrier separating the Wara sand from the massive sands of the underlying Burgan formation. However, extensive faulting prevents communication between the Wara and Burgan sands. The Burgan oil field was discovered in 1938, with first oil production in 1946. Burgan has good energy support from strong natural aquifers, and after 66 years of production, most of the field is still under primary depletion. More than 1,200 wells have been drilled across the field. While drilling spacing is dense in the crestal area, well control in the flanks is relatively weak, giving rise to more geostatistical uncertainty. As the field has matured, the reservoir pressure has declined, reducing productivity. Scaling-Up Approach Because the field comprises significantly different reservoir units, a detailed geological description is required. A very-fine-scale geological model with 900 million cells was required. This geological model was constructed as a master model integrating all types of data available, and will be maintained and updated over time with new data from drilling, production, and other activities. For dynamic modeling, the very-fine-scale geological model is too big and must be scaled up. A multiscale approach is needed for the Greater Burgan field. Several studies, including reservoir management and optimization, facilities design and planning, recovery optimization, and enhanced oil recovery, are envisaged for the field. These studies are expected to require different resolutions and different levels of reservoir-description detail. Use of a single model with the highest possible resolution for all of the studies is impractical for a field this size. Thus, each study had to formulate model-resolution requirements in accordance with the phenomena to be studied, which resulted in different scaling of the models.