Background & Objectives
In the United States, several RCSPs are currently in the planning stages for deployment-phase projects involving geologic storage of one or more million tons of CO2, at rates on the order of several hundred-thousand tons of CO2 per year. Among the early projects starting in 2009 are the injection at Cranfield in Mississippi (SECARB), at Decatur in Illinois (MGSC), and at the Farnham Dome in the Uinta Basin in Utah (SWP); other projects are planned at Kimberlina in the southern San Joaquin Valley in California (WESTCARB), in the Mt. Simon aquifer near Greenville, Ohio (MRSCP), at the Riley Ridge field in Wyoming (Big Sky), and at Fort Nelson, Alberta, and Williston Basin, North Dakota (PCOR). With carefully developed monitoring strategies in place, these tests will provide a wealth of data on relevant site performance measures, such as the growth and migration of the CO2 plume, local and large-scale pressure changes, injectivity, stress evolution, brine migration, and geochemical processes. These data will allow (1) evaluation of the predictive modeling efforts that each partnership will have conducted prior to and during field testing, and (2) improvement of the predictive models through comparison with observation data.
Predictive models for the RCSP field tests are being developed by several different research teams, using different codes and different modeling approaches. There is currently no process in place that (1) would objectively evaluate the respective models using defined and agreed-upon performance metrics, and (2) would provide a forum for discussion, interaction, cooperation, and learning among the various modeling groups. Sim-SEQ is a multi-year effort to initiate and manage such a process. The project goals are to demonstrate in an objective manner (1) that the observed system behavior at GCS sites can be predicted with confidence, and (2) that the remaining differences between models and measurements for different pilot tests, as well as between different models applied to one specific site, are well understood. Further goals are to ensure (3) that model uncertainties are evaluated and their impact is assessed, and (4) that lessons learned and improvements made by one specific research team are documented and available to all other research teams for use to improve future modeling efforts. All this is done in a project environment that fosters mutual respect, and multidisciplinary interactive and cooperative research.
A template for this proposed project is the highly successful DECOVALEX project on model comparison (not code verification), which was conducted by several international organizations involved in the geologic disposal of nuclear wastes. DECOVALEX (DEvelopment of COupled models and their VALidation against EXperiments) was established in 1992 to cooperate in developing and testing models capable of simulating coupled processes. Four multi-year project stages have since been completed. The general goal of the project is to encourage multidisciplinary, interactive, and cooperative research on modeling coupled processes in geologic formations in support of performance assessment for underground storage of radioactive waste. LBNL has been involved in this project for many years, both in research and modeling aspects as well as in managing the cooperative efforts.