Reservoir Digital Twins
A reservoir digital twin combines a calibrated reservoir simulation model with real-time production and injection data to create a continuously updated virtual representation of the subsurface. This enables near-real-time forecasting, optimisation, and scenario testing that traditional annual model updates cannot provide.
Components of a Reservoir Twin
Static Model
The geological framework - structure, stratigraphy, porosity, permeability, saturation - built from seismic, well logs, and core data. Updated when new wells are drilled or new seismic is acquired.
Dynamic Model (Simulator)
The flow simulation engine (Eclipse, CMG, tNavigator) that calculates fluid movement, pressure changes, and production forecasts based on the static model and well controls.
Real-Time Data Feed
Continuous production rates, injection rates, pressures, and well status data from the historian/SCADA - used to keep the twin synchronised with reality.
Automated History Matching
AI-assisted algorithms continuously adjust model parameters (permeability multipliers, aquifer strength, well skin) so the model's predictions match observed production data.
Use Cases
Infill Well Placement
Test hundreds of proposed well locations in the twin to find the optimal placement that maximises incremental recovery.
Waterflood Pattern Optimisation
Simulate different injection rate allocations and patterns to identify the strategy that maximises sweep efficiency and oil recovery.
Reserves Estimation
Run probabilistic forecasts (P10/P50/P90) using the calibrated twin to support annual reserves booking and field development planning.
EOR Screening
Evaluate enhanced oil recovery methods (CO₂ injection, polymer flooding) in the twin before committing to costly field pilots.
Example: A Middle East operator uses a reservoir digital twin covering 1,200 wells and 15 reservoirs. The twin auto-updates weekly with production data and reruns forecasts. When a new infill campaign was proposed, the twin screened 300 candidate locations in 48 hours - a process that previously took the subsurface team 6 months using manual workflows. The optimised selection increased expected recovery by 12 MMbbl over the original plan.