Reservoir Management
Reservoir management is the discipline of maximising the economic recovery of hydrocarbons from a reservoir over its full lifecycle. Digital tools have transformed this from a periodic, model-heavy exercise into a continuous, data-driven process that integrates real-time production data with subsurface models.
Digital Reservoir Management Workflow
Integrated Reservoir Modelling
Static geological models are combined with dynamic simulation models to predict how fluids move through the reservoir. These models are continuously updated as new well and production data becomes available.
Example: A reservoir engineer updates a simulation model with 6 months of new production data, revealing that the water flood front is advancing faster than predicted on the eastern flank.
History Matching
Simulation models are calibrated by adjusting parameters (permeability, porosity, fault transmissibility) until the model reproduces actual historical production data. AI-assisted history matching automates this labour-intensive process.
Example: An ML-assisted history match runs 10,000 parameter combinations overnight, identifying the 50 best-fit models - a task that would take a reservoir engineer months manually.
Injection Optimisation
For waterflooded or gas-injected reservoirs, digital tools optimise injection rates and patterns to maximise sweep efficiency and delay water/gas breakthrough at producing wells.
Example: An optimisation algorithm redistributes water injection across 15 injectors, improving oil recovery by 2% and delaying water breakthrough by 18 months in 3 key producers.
Surveillance & Monitoring
Permanent downhole gauges, 4D seismic surveys, and tracer tests provide ongoing data about reservoir behaviour. This data feeds back into models for continuous improvement.
Example: A 4D seismic survey shows that injected water has bypassed a low-permeability zone, prompting a sidetrack well to access unswept reserves.
Use Case: Chevron's Reservoir Digital Twin
Chevron has deployed digital twin technology for reservoir management in its Permian Basin assets. By integrating real-time well data with reservoir simulation models, Chevron's reservoir engineers can test "what-if" scenarios - such as changing injection rates or drilling new infill wells - on the digital twin before committing to costly field operations. This approach has improved recovery forecasts and reduced the time to make reservoir management decisions from weeks to hours.