Case Study Two

Case Study 2: Multiphase Waterflood Simulator & Production Forecasting Platform
A fully implicit 3D black-oil reservoir simulator built from first principles in Python to evaluate waterflood performance, implementing coupled oil-water flow with Newton-Raphson iterations and scenario-based injection optimization.
geo2
Client & context
Developed internally at GSERC as a production forecasting and waterflood optimization tool for African operators. The simulator enables scenario-based decision-making for injection strategies, recovery optimization, and reserves estimation — all without the cost barrier of commercial simulators.
Problem statement
 
Waterflood operations in African fields face critical challenges:
• Early water breakthrough and rising water cut during long-term injection
• Pressure maintenance requiring optimized injection strategies
• Balancing oil recovery gains against increasing water production costs
• Lack of affordable simulation tools for scenario analysis and reserves forecasting
Methodology
 

Step 1 — 3D grid & physics. Designed a cell-centered 3D finite-difference reservoir model using structured Cartesian grids in x, y, and z directions. Implemented coupled oil-water flow equations with pressure-dependent fluid properties, formation volume factors, viscosities, fluid densities, and rock compressibility effects.

Step 2 — Multiphase characterization. Incorporated relative permeability and capillary pressure functions to model saturation-dependent phase mobility and displacement efficiency. Integrated fluid PVT models including gas solubility relationships.

Step 3 — Nonlinear solver. Developed a fully implicit Newton-Raphson solver with analytical Jacobian matrix construction for robust convergence under strong pressure and saturation nonlinearities. Solved large systems via SciPy sparse linear algebra.

Step 4 — Well modeling. Integrated injection and production well models to calculate dynamic bottom-hole pressure, oil production rate, water production rate, water cut, and water-oil ratio throughout the simulation period.

Step 5 — Scenario analysis. Evaluated five waterflood development scenarios (Cases 1-5) to quantify the impact of injection rates on reservoir pressure support and recovery performance.

Step 6 — Validation. Performed numerical validation through material-balance error tracking and production history consistency checks.

Key parameters
 
Grid: 3D Cartesian (ngx × ngy × ngz)
Phases: Oil + Water (black-oil)
Formulation: Fully implicit
Solver: Newton-Raphson
Forecast: 400+ days
Scenarios: 5 cases
Validation metrics
 
Method: Material-balance error
Convergence: Nonlinear stability
Consistency: Production history
Scenarios: 5 injection cases
Duration: >400 days
Status: Validated

[Figure 4: 3D saturation distribution — waterflood front propagation]

[Figure 5: Oil production rate vs. time — 5 scenario comparison]

[Figure 6: Water cut evolution and WOR trends across cases]

Results & deliverables
 
• Five waterflood scenarios quantifying injection rate impact on recovery
• Oil production rate forecasts for each scenario over 400+ days
• Water cut evolution and breakthrough timing per case
• Water-oil ratio (WOR) trends for economic evaluation
• Cumulative oil recovery (Np) and cumulative water production
• Injector and producer bottom-hole pressures for facilities planning
• Material-balance error reports confirming numerical integrity
Impact
The waterflood simulator enables African operators to evaluate injection strategies, predict water breakthrough, and optimize recovery — all within an open Python environment. For trainees, it provides end-to-end exposure to multiphase reservoir simulation: from PVT and relative permeability theory to Jacobian assembly and production forecasting.