History Matching

How to check initialization of reservoir model

History matching is a process of comparing historical field production and pressures to calculated values.
History matching is used for model verification. It significantly increases the validity of a reservoir simulation model, however it does not guarantee a good model.

Common history matching process

Material balance matching
Match the pressures (field average then well)
Match the GOR/water cut
Match the BHP or THP (PI)

Reservoir Simulation: History Matching and Forecasting. Summary

From https://store.spe.org/Reservoir-Simulation-History-Matching-and-Forecasting-Digital-Edition-P841.aspx

Development Plan Predictions

HM models should be used with caution to predict future field performance with new operational conditions
Industry trend: models provide more favorable outcomes compared to actual performance of the fields
The objective of HM is to provide the model that can reasonably predict future performance of Field under various operational scenarios

Wells stimulation / workover

New well drilling

Abandonment of existing

Injection of different fluids

Surface facility modifications

Facility upgrade planning

Facility and pipeline debottlenecking

Injection rates optimization

EOR planning

The actions intend to increase or accelerate HC production and require significant capital investment. Economic model is to be established to quantify and compare different scenarios
Thus development planning is the process of optimizing economic metrics (NPV, ultimate recovery, etc.) honoring external constraints (regulations, policies, etc.)
The nature of objective function must be clearly defined before starting optimization
Two sets of parameters to be considered:

Uncertainty on which can be reduced during HM

Uncertainty on which can’t be tested and reduced during HM (analog field parameters range are to be used)

Even if reservoir model calibrated (HM), there will be many parameters which not even tested
For example with short production history (depletion), model may be not calibrated to simulate any secondary recovery processes; thus SCAL are not testes and calibrate to field level
Finally model is calibrated to a limited set of conditions based on field production and recovery history
Thus uncertainty should be integrated into Field performance prediction even for brownfields
Complete Risk and Uncertainty Analysis is to be done for any development plan
Identify uncertain parameters that may have significant impact on predicted results
Identify source and quantify the uncertainty of input parameters
Distribution is to be established
Even the best calibrated models have multiple equally valid solutions that need to be considered during uncertainty assessment
For Risk and Uncertainty analysis and probability assessment, proxy models can be used

Wells Productivity & Model Constraints

Wells productivity must be calibrated to take into account discretization and drainage volume effects (may be different from well test results)
Thus Wells productivity should be adjusted taking into account model near well pressure drop (discretization effects)
Loss of permeability as a result of compaction may appear at later stages of development
Well productivity multipliers should be used for all same type wells based on calibrated to well test tested wells
If natural depletion are used, vertical flow correlations are to be designed and calibrated to wells flow tests data
Best correlation that is applicable for the reservoir fluids, tubing parameters, wells deviation, and expected wells regimes
Vertical flow correlations may be be grouped together to minimize the number based on tubing size, MD
Most fields have facility constraints that limit maximum flow rates (pipeline/facility/compression capacity)

Facility constraint impact is to be analyzed during optimization process

Thus facility optimization process is to be iterative

Water / gas production handling constraints
Thus field / group constraints are integrated into forecast
Dynamic change of skin factor, workovers, conversion producer to injector, automatic pattern balancing (WAG, etc.), wells retesting, and drilling of new well
Workover can be performed based on certain well criterias (water cut, GOR)
Thus workover plan can be designed
Workover candidates

Development Optimization

Evaluate optimal well type for the field development

Vertical, deviated well may penetrate multiple reservoirs and cheaper

Horizontal well may be applied for low drainage zones, oil rims, etc.

Number of wells are to be optimized, drilling order

Offsohre platforms have limited number of manifold slots which constraint number of wells producing/injecting at one time

Incremental benefit analysis is to be done to determine optimal number of wells
Green field key uncertainty:

Reservoir size

Possible discontinuities and compartmentalization

Fluid contact

Aquifer size and strength

Sorg, Sgc (reservoirs with gas cap)

Industry trend: consistently overestimated parameter: reservoir connectivity

Discovery of more faults

Discovery of reduced sand channel size

Thus connectivity uncertainty is one of the main uncertainty parameters

External links

History matching of the Norne field by Eirik Morell, September 2010, Department for Petroleum Engineering and Applied Geophysics at NTNU, Norway Preview | Link

History Matching and Uncertainty Assessment of Norne Field E-Segment using Petrel Preview | Link

Errors in History Matching by Z. Tavassoli, Jonathan N. Carter, and Peter R. King, Imperial College, London Preview | Link

A Simple Method for History Matching by SONG Kao-Ping , LI Li-Li , ZHAN Fei. Preview | Link

History Matching/Prediction Preview | Link