From Bureau of Economic Geology, The
University of Texas at Austin (www.beg.utexas.edu).
For more information, please contact the author.
Bureau Seminar, May 11, 2010
Reservoir Characterization and Connectivity Analysis of a Multi-TCF Gas Field:
Quantifying Risk and Uncertainty Through Multi-Disciplinary Analysis
Dr. Lawrence D. Meckel III
Woodside Energy (USA) Ltd
The Sunrise gas field, Timor Sea, covers an area >900 km2. The field consists of two main reservoirs: an upper, laterally extensive shoreface unit and a lower unit comprising incised fluvial valleys, separated by a laterally extensive layer of low-permeability sands, silts, and shales. The trap is distinctive in a regional context in that it is heavily faulted, bounded by faults that reach the seabed, and is only recently formed, constraining hydrocarbon charge to the last 500K years.
Reservoir connectivity is a key development uncertainty, as data from seven widely-spaced appraisal wells present contradictory evidence against and for compartmentalisation. For example, the gas column appears to be in pressure equilibrium and there is extraordinarily good mixing of light end hydrocarbons over distances as large as 10 km laterally, as well as vertically between the major reservoir units. However, there are lateral variations in condensate-gas ratio, gas-water contact, and aquifer gradients, suggesting the potential for compartmentalisation.
The paper presents an interdisciplinary investigation into geological- & production-timescale compartmentalisation at Sunrise, based on stratigraphic, structural, and geochemical analyses as well as numerical models that represent mixing variables over geological time. Multidisciplinary study is essential to being able to understand the potential for compartmentalisation across such a large, dynamic field, sampled by a limited data set.
The study concludes that the field is extraordinarily well-connected over very large distances. Convection is demonstrated to be the most reasonable explanation for the excellent mixing observed at Sunrise. More specifically, rapid mixing (consistent with active charge and migration) requires a “sandwich” model, in which two or more continuous sand layers are vertically connected by conductive fault and fracture networks with a minimum spacing on the order of a few kilometres. Large attic zones would cause mixing to be too slow to fit the field data.
The mixing models confirm that neither reservoir can reasonably contain laterally impermeable barriers, which is contrary to earlier stratigraphic models that did not consider mixing models. These insights were used to justify a vertical interference test on a subsequent appraisal well, the first dynamic data collected in the field. Predictions were later verified by independent stratigraphic work and the vertical interference test.