Reservoir Simulation: History

Matching and Forecasting

Reference: Reservoir Simulation: History Matching and Forecasting; James R. Gilman, Chet Ozgen; Publisher: Society of Petroleum Engineers (SPE)
Reservoir Simulation. Simulator selection
• Problem is to be stated first which is to be solved • To get understanding of the model, sensitivity
analysis is to be done; performance data can never
• Assumptions and limitations (data, etc.) are to be fully constraint the model;
understand
• Investigate sensitivities to:
• The least sophisticated model/simulator that
• operational changes
solves particular problem is to be applied
• reservoir uncertainties
• Misuse of simulator options may cause incorrect • Fluid, fluid-rock uncertainties
results
• Approach to simulation:
• Study reservoir performance
• Make hypothesis
• Test your hypothesis on simulation by comparing model behavior
with observed data

Reference: Reservoir Simulation: History Matching and Forecasting; James R. Gilman, Chet Ozgen; Publisher: Society of Petroleum Engineers (SPE)
Reservoir Simulation. Input data
• Quality of HM is directly proportional to the • THP to BHP conversion is to be Qced
quality/accuracy and amount of available history
data • Production/Injection profile: to calibrate
permeability , rock types
• Data with errors must only be highlighted but not
discarded as still can bring qualitative info • Salinity data: to differentiate injected water from
reservoir
• GOR – PVT
• Compositions: may be CO2, H2S, to HM reservoirs
• Breakthrought time – lateral/vertical barriers; with lateral/vertical compositional gradients
SCAL, HC volume,
• Interference test: connectivity within reservoir
• PTA shut-in pressure - compare to average
pressure in the well drainage area • 4D seismic: to track contact movements
• High perm reservoirs – pressure stabilize fast; low
perm data – drainage radius is to be taken into
account
• RFT/MDT – for contacts, compartments,
initialization

Reference: Reservoir Simulation: History Matching and Forecasting; James R. Gilman, Chet Ozgen; Publisher: Society of Petroleum Engineers (SPE)
Reservoir Simulation. Calibrating the Model
• Before start history matching (HM) exercise, there is The purpose of HM
strong recommendation to use engineering tools first
such as mat bal, PTA, DCA to form opinion on aquifer • Narrow down uncertainties associated with reservoir
size, compartmentalization, recovery mechanism and characterization
reservoir drive • Early history-matching efforts are to iterate back to
• Next stage is to run simple 2D, single well, pattern static model reconstruction and refinement instead of
models are to be run to further understand reservoir simply performing fixes (poro and perm multipliers)
in short time • To calibrate the model so it can be used to reasonably
• HM exercise should be treated as extension of predict future performance of the reservoir under
reservoir characterization process various development and operating scenarios.
• Run model – analyze response (how model replicates • High dependence on static model and stochastic
historical data) – decide what and how to change nature of the static model; thus static model
simulation model to improve match modifications are to be done instead of making not
feasible alteration of other key input parameters
• Mostly “uncertain” information is to be altered during
HM process • Remember about the reliability of production / field
performance data itself
• The model may be falsely HM without fixing the real
uncertainty.
• For example in waterflood case, instead of solving stratigraphy
characterization, rel perm are modified. Thus predictive power
decrease as model will fail to predict gas injection

Reference: Reservoir Simulation: History Matching and Forecasting; James R. Gilman, Chet Ozgen; Publisher: Society of Petroleum Engineers (SPE)
Reservoir Simulation. Parameters modifications
• Can be grouped into volumetric and flow parameters Simulation grid – important items
• Volumetric parameters: • Fow parameters:
• Pore volume • Permeability distribution
• Fluid contacts • Barrieres
• Initial pressure • High perm streaks
• Drainage capillary pressure • Conductive faults
• Compressibility • Porosity distribution
• Composition distribution • Fractures
• PVT properties • SCAL endpoints
• Aquifer • Imbibition capillary pressure
• Leakage, integrity (casing leak) • Rel perm curves
• Aquifer parameters – least known parameters during • Local varioation of pore volume and permeability
reservoir characterization (lareg range of uncertainty) must be backed by geologic resoning
• Prerequisite task to HM: • Local pore volume and permeability multipliers are 2
• compartmentalization within reservoir (sealing faults, shale most common parameters to be changed by
barriers, unconformities, or combination) engineers; however instead of making local
• Aquifer volume and strength range estimation modifications, static model is to be reconstructed
based on dynamic model findings

Reference: Reservoir Simulation: History Matching and Forecasting; James R. Gilman, Chet Ozgen; Publisher: Society of Petroleum Engineers (SPE)