The building, calibration and validation of a coupled simulation model treating two giant subfields simultaneously with their common water injection facilities are presented. The primary objectives of the project were to develop a tool providing more accurate forecasts by consistently allocating injected volumes to both subfields, and to help identify network upgrades necessary to accommodate the long- term development plan. Simulation results presented in this paper are nevertheless based on an eight-years (2015-2023) drilling schedule, followed by no further activity. Building of the model involved migration of separate reservoir and network models to a next generation simulator capable of treating both surface and subsurface flows fully implicitly. Coupled model operation without network (i.e., with wells constrained by guide-rates considering a common field injection target) and with network (i.e., with wells constrained by self-consistently calculated backpressures) are discussed and compared in detail. For the latter case, particular care is necessary in terms of modeling injection modules. Calibration consisted of shifting the lift performance curves of more than 200 active injection strings, in order to make up for any mismatch in the flowing wellhead pressure and help ensure rate continuity at the transition from history to forecast. Tuning of the surface network, modeled "as is" according to its physical layout, was not necessary based on the good backpressure match obtained for the more than 40 active injection towers. Validation was performed through a one-year blind test, from April 2015 (start of forecasts) to March 2016. The procedure first involved well-by-well comparison of injection rates, for which, because of allocation uncertainties, a qualitative match was accepted; it second involved a comparison of the injection split between the two subfields, for which less than a 1% mismatch was achieved.
展开▼
