Recognition Criteria, Significance to Field Performance, and Reservoir Modeling of Sand Injections in the Balder Field, North Sea

Abstract The Balder field comprises seven structurally and stratigraphically trapped oil accumulations, in three separate stratigraphic intervals of Paleocene to early Eocene age. The reservoirs are deep-water gravity-flow-deposited sands, draped and sealed by hemipelagic mud and volcanic tuff. Strong initial reservoir compartmentalization has been significantly modified by postdepositional sand remobilization and sand injection. Sand injections were first recognized in 1969, in core from one of the early exploration wells. At the time, they were not thought to be significant. More than 150 sand injections have subsequently been identified in cores from the exploration, appraisal, and development drilling programs. This led to the realization that they were not just isolated occurrences but a common phenomenon that could impact production. In addition to the cores, sand injections are now recognized and mapped using seismic and well-log correlations, aided by principles derived from outcrop observation. Initial geologic and reservoir modeling of the Balder field consisted of separate models for each of the seven main accumulations. This was partially caused by the complexity of sand distribution and an incomplete understanding of how individual sand bodies might be interconnected. It was also caused by limitations in available software for building and simulating models of high complexity. The field went on production in September 1999. In early 2000, when the first phase of infill drilling was considered and more efficient tools for geologic and reservoir modeling were available, the decision was taken to build a full-field model. To better understand how to incorporate the sand injections, outcrop analogs were studied to develop models for injection dimensions, orientations, and distribution patterns. Following the outcrop work, a detailed reexamination of the Balder core, log, and seismic data was undertaken, cataloging all evidence for sand injections in the field. Strict criteria were developed for recognizing sand injections of all scales. Full-field geologic and reservoir models have now been built for the field, incorporating both the depositional and injected sand bodies. Because of resolution limitations, particularly with the smaller and more steeply inclined sand injections, parts of the model are nonunique. Ongoing production history matching has led to model modifications, which continue to adhere to the core, well-log, and seismic constraints, but now yield improved matches to individual well performances and gas-water breakthrough histories. This has been an iterative process, which is expected to continue throughout the life of the field. The full-field model is now being used to predict future production milestones and help identify optimum locations for potential infill drilling and well workovers.