Seismic attribute classification using Rock Physics

1
Seismic attribute classification using Rock
Physics

• T.Tonsing
• SAGA Western Cape Branch

2
Objective

• Discussion on alternative classification scheme
• Not a case study but presentation of workflow
• Discussion on differences

3
The bigger picture
Structural frame
Stratigraphic frame
Gridding
Property modeling
Upscaling and construction of dynamic model
Exploration
Appraisal
4
The Rock Physics workflow
Calibrating rock phys engine
Geological Model
Log Analysis

• Expected variation of physical properties
• Porosity
• Pore geometry
• Cementation
• Pressure
• Clay
• Fluids

Geo model with Vp, Vs and ? distributions
Rock Physics engine
Synthetic attributes simulation
Wavelet
Seismic data
Attribute extraction
Calibration / upscaling
Classification
5
Seismic analysis and geo model
6
Log analysis
7
Velocity Porosity Models
8
Xu WhiteInvestigation into acoustic wave
propagation in sedimentary rocks

• The Kuster-Toksoz model

• Porosity
• Pore geometry
• Cementation
• Clays
• Pressure

9
The Xu-White Model

• Kuster Toksoz Model
• Kuster's model assumes ?/? ltlt 1
• ie porosity much less that aspect ratio
• This implies that there is no interaction between
  pores

10
The Xu-White Model

• Realised Kuster limitation
• Uses Kuster on dry frame (sandstone, clay, ?, ?)
• Aspect ratios from best fit to sonic logs
• Uses Gassmann to model fluid relaxation (pore
  interaction )

11
The Xu-White Model
12
The Xu-White Model
North sea example improved results through
calibration
DT
?
ILD
GR
DT
?
GR
13
The Xu-White Model
Plots of predicted vs recorded p-wave velocities
Xu-White with best-fit aspect ratios
Wyllies time-average
14
The Xu-White Model
Modeling the effects of clay on VP
15
The Rock Physics workflow
Calibrating rock phys engine
Geological Model
Log Analysis

• Expected variation of physical properties
• Porosity
• Pore geometry
• Cementation
• Pressure
• Clay
• Fluids

Geo model with Vp, Vs and ? distributions
Rock Physics engine
Synthetic attributes simulation
Wavelet
Seismic data
Attribute extraction
Calibration / upscaling
Classification
16
Per Avseth, Norsk Hydro 
17
Rock Phys vs Cluster Analysis

• Rock Physics
• Not reconnaissance tool
• Forces Interpreter to understand both log and
  seismic response
• Requires good data
• Leads into a seismically constrained reservoir
  model!!
• The calibration step.

• Cluster Analysis
• Fast analysis of large amounts of data
• Incorporates more (independent?) attributes than
  intercept and gradient
• Could be combined with rock physics?

Thank you