How does reservoir simulation work in oil and gas production?

I. Purpose and Value-Chain Placement

Reservoir simulation is the physics-based, numerical prediction of subsurface fluid flow used to design field development and optimize production over the asset life.

• I.I Sits between subsurface characterization and production operations; converts geological and petrophysical understanding into actionable well, facility, and injection plans.
• I.II Guides well count/placement, completion strategy, injection schemes (water, gas, chemical/thermal), production ramp, and abandonment timing.
• I.III Integrates data across the value chain: exploration (structure/contacts), drilling (pressures, mud losses), petrophysics/core (porosity, perms, SCAL), PVT labs (EOS/black-oil tables), and production allocation (rates, pressures).

II. Step-by-Step Workflow

1. 1. Data assembly and QC

   • 1.1 Structural and stratigraphic framework from seismic interpretation.
   • 1.2 Petrophysics: porosity, permeability, water saturation, net-to-gross; core SCAL (relative permeability, capillary pressure).
   • 1.3 PVT: black-oil tables (Rs, Bo, µo, µg, µw) or EOS compositional model tuned to lab data.
   • 1.4 Production/pressure history and well tests for history matching; allocation QC.

2. 2. Static model build (geomodel)

   • 2.1 Corner-point or unstructured grid honoring faults, horizons, and facies.
   • 2.2 Property modeling (?, k, Swirr, Sor) with geostatistics; upscaling fine-scale models to simulator grid.

3. 3. Discretization and well representation

   • 3.1 Grid selection: structured (corner-point), local grid refinement (LGR), or unstructured (Pebi/polyhedral) for complex wells/fractures.
   • 3.2 Well trajectories, completions, and controls (BHP, oil/water/gas rate targets, WOR/GOR limits); near-wellbore skin and multiphase productivity index.

4. 4. Fluids and rock–fluid interactions

   • 4.1 Black-oil or compositional PVT; water salinity where relevant.
   • 4.2 Relative permeability and capillary pressure tables per rock type/facies; hysteresis where needed.

5. 5. Initialization and boundaries

   • 5.1 Pressure–depth functions, fluid contacts, temperature gradient; aquifer models (lateral or bottom water).
   • 5.2 Boundary conditions: no-flow, constant pressure, or analytic aquifer (Carter–Tracy/Fetkovich).

6. 6. History matching (calibration)

   • 6.1 Adjust uncertain parameters (fault transmissibility multipliers, relperm endpoints, kv/kh, aquifer strength) within geologic plausibility.
   • 6.2 Objective is to reduce mismatch in rates, pressures, GOR/WOR, and 4D seismic without overfitting.

7. 7. Forecasting and optimization

   • 7.1 Run scenarios: well count/spacing, injector–producer patterns, lift strategies, EOR screening, facility constraints.
   • 7.2 Production optimization under constraints (BHP, facility capacity, water handling, gas export).

8. 8. Uncertainty and decision support

   • 8.1 Ensemble models spanning subsurface and PVT uncertainties; probabilistic forecasts (P10–P90).
   • 8.2 Decision metrics: expected value, downside protection, option value of phased developments.

Core physics and numerical solution

• Governing equations
  • Darcy’s law for each phase i:

    \( \mathbf{v}_i = - \dfrac{k\,k_{ri}}{\mu_i} \left(\nabla p_i - \rho_i \,\mathbf{g}\right) \)

  • Mass conservation per component/phase:

    \( \dfrac{\partial}{\partial t}\left(\phi\,\rho_i\,S_i\right) + \nabla \cdot \left(\rho_i\,\mathbf{v}_i\right) = q_i \)

  • Black-oil relations (examples):

    \( R_s = f(p), \quad B_o = f(p), \quad B_g = f(p) \)

  • Capillary pressure:

    \( p_o - p_w = p_{c,ow}(S_w), \quad p_g - p_o = p_{c,go}(S_g) \)

• Well model (Peaceman approximation)

  \( q = WI \,\lambda \,(p_{bh} - p_{cell}) \), with \( \lambda = \dfrac{k_{r,eff}}{\mu} \). For a horizontal well segment in a Cartesian grid:

  \( WI = \dfrac{2\pi k_h \,\Delta z}{\ln\left(\dfrac{r_e}{r_w}\right)+s} \), where \( r_e \approx 0.14\sqrt{\Delta x^2 + \Delta y^2} \) (estimated), s = skin.

• Numerics
  • Discretization: finite-volume/finite-difference on grid cells; fluxes via transmissibilities.
  • Time integration: IMPES (explicit saturation, implicit pressure), fully implicit (FIM), or sequential implicit.
  • Nonlinear solve: Newton–Raphson on residual vector R(U) for unknowns U = [p, S_w, S_g, …]; linear systems solved with Krylov solvers and preconditioners (e.g., CPR-type).
  • Stability: time step ?t managed by Courant criterion and Newton convergence:

    \( \text{CFL} = \max\limits_{cells} \dfrac{|v|\,\Delta t}{\phi\,\Delta x} \le C_{max} \) (estimated)

III. Major Components and Their Functions

• III.I Static geomodeler: builds structural grid and property distributions (?, k, NTG, facies).
• III.II Reservoir simulator engines:
  • Black-oil (three-phase, solution gas, limited compositional effects) for most waterfloods.
  • Compositional (EOS) for volatile oils, gas condensates, miscible gas, WAG.
  • Thermal/chemical modules for steam, polymer/ASP, surfactant flooding; geomechanics coupling if needed.
• III.III Pre/post-processors: upscaling, PVT regression, SCAL normalization, relative perm hysteresis handling, map/cross-plot tools, 4D seismic integration.
• III.IV Well and network models: wellbore hydraulics, lift performance, facility constraints; optional coupling with surface network simulators.
• III.V Optimization and uncertainty toolkits: adjoint gradients, ensemble Kalman filter, genetic algorithms, design of experiments, proxy models.
• III.VI Compute infrastructure: multi-core workstations and HPC clusters; parallel domain decomposition and scalable I/O.

IV. Key Performance Drivers

• IV.I Accuracy vs. resolution
  • Grid quality and alignment to flow; LGR around wells/fractures to capture coning and sweep fronts.
  • SCAL representativeness; relperm endpoints, Corey exponents, hysteresis parameters.
  • PVT tuning quality; EOS regression to separator tests, CCE/CVD/DT data.
• IV.II Numerical robustness and speed
  • Time-step control via convergence measures; avoid excessive cutbacks.
  • Linear solver performance (preconditioning, partitioning); parallel scaling efficiency.
  • Minimize numerical dispersion and grid orientation effects.
• IV.III Match quality and predictive value
  • Use multiple data types: rates, pressures, tracer/PLT, 4D seismic, RFT/MDT.
  • Objective function example (rates/pressures weighted RMSE):

    \( \text{RMSE} = \sqrt{\dfrac{1}{N}\sum_{j=1}^{N} w_j \left(m_j^{sim} - m_j^{obs}\right)^2} \)

  • Avoid over-parameterization; maintain geologic plausibility.
• IV.IV Operational realism
  • Well constraints: BHP limits, lift curves, sand control, water/gas handling capacity.
  • Facility backpressure and network capacity; downtime modeling.
• IV.V Decision relevance
  • Link forecasts to economics; optimize under uncertainty.
  • Use ensembles and value-of-information to target data acquisition.

V. Typical Challenges and Mitigation

• V.I Non-uniqueness in history match
  • Mitigate with multi-objective calibration using orthogonal data (PLT, 4D seismic); constrain parameters with priors and regularization.
  • Use ensemble methods (EnKF/ES-MDA) to maintain uncertainty while improving fit.
• V.II Poor data quality or gaps
  • Allocation audits, test separator campaigns, targeted pressure surveys; tune measurement error models.
  • Design acquisition to reduce key uncertainties (SCAL, PVT, interference tests).
• V.III Heterogeneity and scale disparity
  • Flow-based upscaling; transmissibility multipliers across faults; LGR around critical features.
  • Unstructured grids for complex geology and multilateral/fractured wells.
• V.IV Complex physics (compositional/EOR/thermal)
  • Appropriate EOS and phase-behavior validation (MMP checks, slim-tube); thermal properties for steam/solvent.
  • Specialized relperm and hysteresis for miscible/immiscible floods; polymer adsorption/viscosity models.
• V.V Numerical issues
  • Convergence failures: improve scaling, timestep strategy, and Jacobian preconditioning; adjust capillary numbers carefully.
  • Grid orientation/numerical dispersion: use flux limiters, refine grid, or rotate grid to main flow direction.
• V.VI Coupling with surface system
  • Iterative coupling to surface network to honor backpressure/capacity limits; align constraints and downtime calendars.

VI. Economic and Operational Impact

• VI.I Development design
  • Optimizes well count, spacing, and patterns to maximize recovery and minimize water/gas production.
  • Phased drilling scheduling to align with facility debottlenecking and export constraints.
• VI.II EOR screening and timing
  • Quantifies incremental recovery and breakthrough timing for waterflood, WAG, polymer/ASP, steam/solvent.
  • Selects slugs, injection rates, and surveillance to protect sweep and manage conformance.
• VI.III Value linkage
  • Typical objective function for optimization (illustrative):

    \( J = \sum_{t=1}^{T} \left[p_o\,q_o(t) + p_g\,q_g(t) - c_w\,q_w(t) - c_{op}(t)\right] e^{-r\,t} \)

    where p’s are prices, q’s are simulated rates, c’s are costs, r is discount rate. Maximizing J under constraints yields economically optimal strategies.

  • Ensemble-based decision metrics (P10–P90) inform risk-weighted capex and surveillance plans.
• VI.IV HSE and emissions
  • Improved water/gas management reduces flaring/venting and water handling footprint.
  • Better sweep reduces energy per barrel produced (lower lift/compression duty).