si-yong lee

36
Si-Yong Lee Model development & Aneth site example

Upload: javier

Post on 23-Feb-2016

119 views

Category:

Documents


0 download

DESCRIPTION

Model development & Aneth site example . Si-Yong Lee. What is a model? . A model is a simplified representation of reality or any device that represents a system. Why model?. Predictive application (predicting the consequences of a proposed action) Interpretive application - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Si-Yong Lee

Si-Yong Lee

Model development & Aneth site example

Page 2: Si-Yong Lee

What is a model?

A model is a simplified representation of realityor any device that represents a system.

Page 3: Si-Yong Lee

Why model?

- Predictive application (predicting the consequences of a proposed action)

- Interpretive application (understanding system dynamics)

- Generic application (analyzing processes in generic/hypothetical settings)

Page 4: Si-Yong Lee

What types of models?

Conceptual Model: Qualitative description of system

Mathematical Model: Mathematical description of system- Analytical solution- Numerical solution

Physical Model: e.g. core flooding experiment

Page 5: Si-Yong Lee

Modeling Protocol

Define Problem

Conceptual model

Mathematical model

Computation

Comparison with field data

Results

Model Calibration

Model Redesign

Page 6: Si-Yong Lee

Define the problems/objectives

• Site selection

- storage capacity

- Injectivity

- Plume distribution (AOR)

• Monitoring design

• Uncertainty/Risk assessment

Page 7: Si-Yong Lee

Data Collection

• Hydrologic data (local & regional)• Geologic data (e.g., stratigraphy, formation tops, faults/fractures, tectonic

information, and seismic events)• Geophysical data (e.g., well logs, seismic survey) • Rock properties (por, perm, relative perm, Pc, bulk density, Young’s

modulus, Poisson’s ratio, mineralogy, etc)• Fluid properties (salinity, pH, density, viscosity, mutual solubility, brine

chemistry, isotope, etc)• Well information (location, vertical/horizontal, perforation interval,

injection/production history, bottom hole pressure, etc)

Page 8: Si-Yong Lee

Conceptual Model

Cross-bedded aeolian Navajo Ss(outcrop in Devil’s canyon, UT)

Conceptual modelof the cross-bedded bedform

3D cross-bedded bedform

Grain flow (dune)Wind ripple (interdune)

Page 9: Si-Yong Lee

Grid building

An optimally-sized model domain should :

- Encompass all the major flow units (formations of interest – injection zone, overlying and underlying formations)

- Include the injection, monitoring, and any production wells

- Lie within the extent of pressure response area

- Be tractable computationally

Page 10: Si-Yong Lee

Grid resolution (dx, dy, dz)

Grid resolution vs. computational efficiency

Should include heterogeneity, well configuration, and sufficient accuracy in the changes of results (pressure & saturation).

Coarsening of model grid further from the injection well (no more than 1.5 times the previous nodal spacing).

Grid coarsening could create numerical dispersion.

Page 11: Si-Yong Lee

Assigning property parameters

- Single value in a cell (REV, scale issue)

- Sparse data in space (especially horizontal direction)

- Heterogeneity

- Property upscaling

Page 12: Si-Yong Lee

Heterogeneity and Aniostropy

Heterogeneity : Variations through space

Aniosotropy : Variations with the direction of measurement at any given point

Page 13: Si-Yong Lee

(

Heterogeneity and Aniostropy

(x1,z1)

(x2,z2)

kx

kz

Homogeneous, Isotropic Homogeneous, Anisotropic

Heterogeneous, Isotropic Heterogeneous, Anisotropic

Page 14: Si-Yong Lee

Approaches to generate heterogeneity

Deterministic approach: parameter values are known with certainty (single solution)

Stochastic approach: uncertainty in parameter values (ranges in solution)

Actual Geology

Layer Cake Model

Stochastic/Geostat. Model

Page 15: Si-Yong Lee

Stochastic Approaches

• Continuous HeterogeneityGaussian model (mean, variance, and variogram)Fractal model

• Discrete HeterogeneityFacies model with indicator geostatisticsDepositional simulation

Process imitation (mathematically-based equations)Structure imitation (probabilistically-based)

• Mixed Heterogeneity (continuous + discrete)

Page 16: Si-Yong Lee

x y

z

Core description (LLNL site)

Page 17: Si-Yong Lee

(TPROGS1)

TProGS Realization

x y

z

Page 18: Si-Yong Lee

(TPROGS1)

TProGS Realization(largest connected channel body)

x y

z

Page 19: Si-Yong Lee

Spatial Covariance of LnK

Lag(m)

Covariance of LnK

0. 20. 40. 60. 80. 100. 120 .0 .0

5.0

10.0

15.0

20.0

25.0

Covariance of Ln K

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.000.0

5.0

10.0

15.0

20.0

25.0

Lag(m)

Page 20: Si-Yong Lee

SGS Realization

(GAUSS1)

x y

z

Page 21: Si-Yong Lee

SGS Realization(largest connected body)

(GAUSS1)

x y

z

Page 22: Si-Yong Lee

TProGS vs. SGS

TProGS SGS

RF Discrete(e.g. facies unit) Continuous

SpatialProcess Markovian Gaussian

VariabilityMeasure Transition Probability Covariance

Advantage- Asymmetry- Juxtapositional tendency- Sharp contact

- Easy application- Simple and fast algorithm

Disadvantage- Relatively more uncertain in x, y than z direction

- Poor Connectivity of extreme values (Maximum entropy)

Page 23: Si-Yong Lee

Geologic Model Development in Aneth site

- Data Acquisition

- Petrophysical Properties Estimation

Estimation of porosity

Porosity & Permeability Relationship

- Geologic Model Development

Page 24: Si-Yong Lee

Data Acquisition

- Core plug analyses

(porosity, density, and permeability)

- Geophysical well log images

- Stratigraphic formation tops data

- Well information

- Injection/production history

Page 25: Si-Yong Lee
Page 26: Si-Yong Lee
Page 27: Si-Yong Lee
Page 28: Si-Yong Lee

NavajoKayentaWingateChinleDechellyOrgan RockHermosaIsmayGothicDesert Creek

Entrada

Page 29: Si-Yong Lee

Petrophysical Properties Estimation

Formation No. of Samples

Porosity () Permeability (mD)

Mean Median Std. Dev. Mean Median Std. Dev.

Ismay 10 0.05 0.02 0.06 0.47 0.04 0.78

Gothic Shale 1 0.009 0.009 0 0.012 0.012 0

Desert Creek 81 0.09 0.1 0.07 5.12 0.31 18.92

Page 30: Si-Yong Lee

Texaco Aneth H-117

5380

5400

5420

5440

5460

5480

5500

5520

5540

0 5 10 15 20 25 30

Porosity (%)

Dept

h (ft

)

Ambient Porosity vs. Neutron-Density Porosity

Page 31: Si-Yong Lee

Upscaled Porosity Logs

Page 32: Si-Yong Lee

Porosity Field(n=9,170,238; dx=dy=100m, dz=1m; nz=1,644)

Page 33: Si-Yong Lee

Upscaled Porosity Field(n=227,950; dx=dy=100m; nz=41)

Page 34: Si-Yong Lee

y = 0.0253e0.2824x

R2 = 0.6626

y = 0.0504e0.1655x

R2 = 0.3485

1.E-02

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

0 5 10 15 20 25 30 35

porosity (%)

k (m

D)

Desert CreekGothic

IsmayExpon. (Desert Creek)

Expon. (Ismay)

Porosity vs. Permeability

Page 35: Si-Yong Lee

Permeability Field(n=227,950; dx=dy=100m; nz=41)

Page 36: Si-Yong Lee

Questions ?