Step Change in Reservoir Simulation Breathes Life Into a Mature Oil Field
- Dennis Denney (JPT Technology Editor)
- Document ID
- Society of Petroleum Engineers
- Journal of Petroleum Technology
- Publication Date
- January 2006
- Document Type
- Journal Paper
- 35 - 37
- 2006. Society of Petroleum Engineers
- 0 in the last 30 days
- 43 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||Free|
|SPE Non-Member Price:||USD 15.00|
This article, written by Technology Editor Dennis Denney, contains highlights of paper SPE 94940, "Step Change in Reservoir Simulation Breathes Life Into a Mature Oil Field," by M.J. Kromah, SPE, J. Liou, and D.G. MacDonald, SPE, BP, prepared for the 2005 SPE Latin American and Caribbean Petroleum Engineering Conference, Rio de Janeiro, 20-23 June.
Understanding the uncertainties and optimizing the depletion plans are key to successful reservoir management. BP Trinidad and Tobago uses reservoir simulation to history match and then predict reservoir performance from highly faulted, stacked sands in Trinidad. The full-length paper details successful use of BP’s top-down reservoir-modeling (TDRM) technology for optimizing infill-well locations in a mature oil field. The study used multiple scenarios with 16 history-matched models to broaden the scope of analysis of the reservoirs in a time-efficient manner.
TDRM is BP’s proprietary technology to per-form and interpret production-performance prediction. As Fig. 1 shows, this approach encapsulates reservoir uncertainties in the simplest appropriate model, adding only the level of detail required. With sparse data sets and the need to optimize reservoirs, these models enable understanding the sensitivity of the performance prediction.
Traditionally, reservoir engineers populate the “most-likely” geological interpretation with the “most-likely” reservoir and fluid properties to develop a base-case understanding of the reservoir. Uncertainty analysis is limited to a few sensitivities of what is thought to be the most significant and least-known properties, but there is no full under-standing of the dependency of the parameters or their effect when combined. The full range of combinations cannot be explored in a timely manner. This top-down method uses simpler models with shorter run times and advanced case management, allowing the use of more models to explore the region of the uncertainty space that could affect the business decision.
The history-matching module uses a genetic algorithm to intelligently explore the uncertainty space and match the field and well data. An assisted-history-matching process is set up in which the reservoir engineer defines the uncertainties that affect the business decision and the acceptable precision for the history match (an objective function). The algorithm minimizes the objective function within the defined uncertainty space. This process finds multiple unique solutions that are equally matched to the historical data.
The conventional approach to history matching consumes most of the project time in building one detailed deterministic model and getting a single acceptable match. There is not sufficient time for proper analysis and understanding of what the model represents. This top-down method uses massively parallel computing to build multiple history-matched models in a time-efficient manner. Model updates and changing of uncertainty space may be done more frequently.
The Teak field is 25 miles off Trinidad’s southeast coast and has been producing since 1972. Oil production is from multiple thick, stacked reservoirs that are highly faulted and blocky. The T sand, at a depth of 5,500 ft, has a 60-ft-thick oil column between a gas cap and an active aquifer. Thirteen wells have produced from the T-sand reservoir.
|File Size||467 KB||Number of Pages||3|