Click to view page image in pdf format.
7th Middle East Geosciences Conference and Exhibition
Manama, Bahrain
March 27-29, 2006
Modeling
of Structure, Porosity and Permeability Using a Combination of
Geology-Controlled Deterministic and Stochastic Techniques to Minimize Uncertainty: Example
from a Giant Carbonate Reservoir, Onshore Abu Dhabi, United Arab Emirates
1 UFR Team, ADCO, Abu Dhabi Company for Onshore Oil Operations, P.O. Box 270, Abu Dhabi, United Arab Emirates,
phone: +971504140995, fax: +97126667276,
[email protected]
2 Planning & Quality Department (PQD) / Petroleum Dev. Division, Abu Dhabi Company for Onshore
Oil Operations (ADCO), PO Box 270, Abu Dhabi, United Arab Emirates
3 Kelkar & Associates, Inc, 3528 East 104th Street, Tulsa, OK 74137
Applied to a giant undeveloped carbonate reservoir located onshore Abu Dhabi, a comprehensive and effective workflow
using stronger geological controls on 3D modeling
of structure, porosity and permeability is described.
Depth grids are constructed using 3D seismic and isochores. Geology honoring positive or negative depth trends are used respectively for porous and dense intervals isochores.
Porosity modeling
starts, at well locations with overburden correction of core porosity followed by a linear correction of log
porosity controlled by the former. To honor the core porosity and account for the uncertainty of measurements, a stochastic
technique using the residuals between core and log corrected porosity was developed. Then 3D porosity models are
created through stochastic
modeling
using porosity versus depth trends and a base case is selected.
Permeability modeling
uses a mixed deterministic/stochastic technique. Missing permeability intervals from wells with
production tests are edited using porosity versus permeability plots and Kh derived from test results or dynamic history
matching, minimizing the multipliers need. Geology based power lines fitted to porosity versus permeability clouds are then
used to convert the base case porosity model into a 3D synthetic permeability model lacking the well data variability. This is
re-incorporated through stochastic simulations of the residuals between the synthetic and real well values and a base case
permeability model is selected. Finally, a permeability multiplier volume is generated to match test results. Uncertainty is
accessed through creation of high and low cases.
Produced static models were successfully used in dynamic simulations with very limited adjustments.