Title: Generation of a Reference Petrophysical Seismic Data Set the Stanford V reservoir
1Generation of a Reference Petrophysical /
Seismic Data Set - the Stanford V reservoir -
- Shuguang Mao and André G. Journel
- Thanks to Tapan Mukerji, Geophysics Dept.
- Stanford Center for Reservoir Forecasting
- Stanford University
- May 4-5, 2000
2Outline
- Objectives
- Five steps to generate a reference data set
- Simulate layer geometry top / bottom depths
- Simulate lithofacies types
- Obtain petrophysical properties porosity,
permeability, density, velocity, impedance - Obtain approximate seismic survey result
seismic measured impedance, travel time - Extract well paths and obtain well log data
- Summary
3Objectives
- To allow extensive testing of any proposed
reservoir characterization algorithm on 3D
multi-layered fluvial channel reservoir - Layer surfaces top and bottom for all layers
- Lithofacies types facies indiators
- Petrophysical parameters ?, k, ?, v
- Seismic data travel time, impedance
- Extract any number of well log data along any
deviated well path
4Methodology
Layer geometry top and bottom depths
Petrophysical properties facies, ?, k, ?, v,
imp
Vertical strat. grid
z
high resolution
Forward modeling
Well planning
Seismic impedance
t
Seismic travel time
bilinear interpolation
low resolution
5Step 1 simulate layer geometry
- Bottom surface top surface thickness
- Topmost surface
- Thicknesses
bottom surface of layer k
top surface of layer k
thickness of layer k
Non-conditional Gaussian simulation (sgsim)
with small nugget ( 1) and relative large range.
Non-conditional Gaussian simulation (sgsim)
with small nugget ( 5), various ranges and
anisotropies.
6Simulation results layer geometry
7Step 2 simulate lithofacies types
- A three-layered reservoir, deposited in a fluvial
channel environment - Use object-based simulation algorithm, code
fluvsim, see Deutsch Wang, 1996 - Non-conditional facies simulation
- Different channel parameters for each layer,
e.g., orientation, width, channel number, etc.
8Simulation results lithofacies
Channel Sand
Channel Sand
Crevasse
Crevasse
Mudstone
Layer 1
Layer 2
Layer 3
9Step 3 simulate rock petrophysics
Permeability
3 facies indicators from 3 layers
Porosity distribution for different facies
Rock Physics
Porosity
Density
Non-conditional SGSIM
Velocity
Regression (channel, crevasse)
Regression (mudstone)
Porosity simulation results in 3 facies
Impedance
Re-assemble
10Step 3a simulate porosity values
- Non-conditional sequential Gaussian simulation
(sgsim) for each lithofacies, conditional to
reference distribution of porosity - Merge the above simulated results according to
simulated lithofacies type
11Simulation results porosity in 3 facies
Channel sand
m 0.29 cv 0.11
Crevasse
m 0.13 cv 0.37
Mudstone
m 0.08 cv 0.55
12Simulation results porosity
Porosity
13Step 3b simulate permeability
- Sequential Gaussian simulation of permeability
conditional to collocated porosity - Separately, (1) mudstone facies (2)
non-mudstone facies - Back transform simulation results to lognormal
distribution for each facies - Merge the above two permeability fields according
to facies type
14Simulation results permeability
Permeability
15Step 3c obtain density and velocity
- Rock mineral matrix fluid in void space
- Use empirical regression models
16Simulation results density
Density
17Simulation results velocity, impedance
18Step 4 Approximate seismic response
- Recall impedance velocity x density
- Apply a frequency-domain Born filter to mimic a
realistic impedance as obtained from a seismic
inversion - A local Backus average with vertical sliding
window to smooth out thin layers - Apply normal incidence ray theory to get
- seismic velocity
- seismic two-way travel time
19Approximate seismic responses
20Seismic response travel time
true
available
x-cross sec.
y-cross sec.
21Step 5 Extract well log data
- Decide number of wells, locations and types of
well (vertical, deviated, horizontal) - Design well paths in 3D
- Extract well logs by interpolation
- find layer k to which location u belongs
- compute vertical stratigraphic coordinate
- find the nearest 8 grid blocks from data set
cubes, then perform bilinear interpolation
22Results well paths in 3D
- Inspired by real well setting of North Sea
reservoir - 4 platforms
- 2 vertical wells
- 2 deviated wells
- 8 horizontal wells
3D view of well paths
23Sample well log
- Well data include
- horizontal location
- true vertical depth
- porosity
- permeability
- density
- sonic log
- lithology
- layer indicator
- Well markers true vertical depth when well
intercepts a layer surface
24Summary
- A reference data set is generated to mimic a
three layers fluvial channel reservoir, which - has realistic channel number, size, shape,
porosity, density and velocity distributions - provides approximate seismic data, including
seismic measured impedance, seismic two-way
travel time to each layer surface - can provide well log data according to any user-
designed well path - Future work
- Design OWC then change brine in channel sand
facies which are above OWC to oil. - Perform forward modeling for accurate seismic.