Injection Pattern Design to Maximize the Efficiency of Carbon Dioxide Injection for Sequestration Purposes in Brine Formations

Sun, Qian; Ertekin, Turgay

Carbon dioxide injection into subsurface brine formations is one of the more promising carbon storage technologies. Horizontal injectors have been suggested in previous research studies due to higher injectivities and lower bottomhole injection pressures encountered as compared to conventional vertical injectors. Simultaneous production of brine from the formation as carbon dioxide is injected does not only decrease the formation pressure but also increases the total carbon dioxide intake capacity of the formation. However, dealing with the produced brine could be expensive and cumbersome. Accordingly, the goal should be producing as small amount of brine while injecting as much CO2 as possible. An injection efficiency term which is calculated as the ratio of cumulative carbon dioxide injection to the cumulative brine production was defined to evaluate the enhancement effectiveness of a brine production pattern under a carbon dioxide injection scenario. Another significant parameter in evaluating the efficacy of CO2 injection into brine formations is the value of the stabilized reservoir pressure of the CO2- brine system. A dimensionless pressure depletion ratio term which is calculated as the ratio of the stabilized injector block pressure to the initial reservoir pressure introduced to monitor the pressure constraint in the CO2-brine system after stabilization. The pressure depletion ratio has to be less than one to ensure that the injection process is conducted under the consideration of a safety factor. The injection efficiency and the pressure depletion ratio is determined by the following four parameters: size of the pattern, injection rate, horizontal extension of the horizontal well, amount of brine produced from the formation. In order to reduce the uncertainties arising from various reservoir properties, an analysis protocol is structured via Monte Carlo simulation. A commercial numerical model with a compositional formulation capable of incorporating the thermodynamic behavior of CO2-brine system under operating reservoir conditions is used. Running the high-fidelity model for numerous times in Monte Carlo Simulation runs can be extremely time consuming. The development of an artificial neural network could be an effective method to circumvent this issue. In this study, a forward looking network and an inverse network with specified bottomhole pressure and constant production rate of the brine producer have been designed.

AAPG Search and Discovery Article #90163©2013AAPG 2013 Annual Convention and Exhibition, Pittsburgh, Pennsylvania, May 19-22, 2013