13.1 Seismic inversion

Seismic inversion transforms seismic data into rock property descriptions, revealing subsurface details crucial for oil and gas exploration. It uses the link between seismic reflections and acoustic impedance to estimate properties like porosity and fluid content.

Advanced techniques, including machine learning, are pushing the boundaries of seismic inversion. These methods can handle complex geology, extract more information from data, and provide insights into reservoir characteristics that were previously hard to obtain.

Seismic Inversion for Subsurface Characterization

Fundamentals of Seismic Inversion

Top images from around the web for Fundamentals of Seismic Inversion

• 

  SE - Pre-inversion normal fault geometry controls inversion style and magnitude, Farsund Basin ... View original

  Is this image relevant?

• 

  SE - Pre-inversion normal fault geometry controls inversion style and magnitude, Farsund Basin ... View original

  Is this image relevant?

1 of 1

Top images from around the web for Fundamentals of Seismic Inversion

• 

  SE - Pre-inversion normal fault geometry controls inversion style and magnitude, Farsund Basin ... View original

  Is this image relevant?

• 

  SE - Pre-inversion normal fault geometry controls inversion style and magnitude, Farsund Basin ... View original

  Is this image relevant?

1 of 1

• Seismic inversion transforms seismic reflection data into quantitative rock-property descriptions of reservoirs
• Fundamental principle relies on the relationship between seismic reflections and acoustic impedance contrasts in the subsurface
• Post-stack inversion methods (sparse-spike inversion, model-based inversion) estimate acoustic impedance using stacked seismic data
• Pre-stack inversion techniques (AVO inversion) utilize angle-dependent reflectivity to estimate multiple elastic properties simultaneously
• Deterministic inversion approaches assume a unique solution
• Stochastic inversion methods incorporate uncertainty and generate multiple realizations

Advanced Techniques and Algorithms

• Machine learning and artificial intelligence techniques apply to seismic inversion problems
  • Offer new possibilities for pattern recognition
  • Explore non-linear relationships between seismic data and rock properties
• Choice of inversion algorithm depends on:
  • Available data (post-stack vs. pre-stack)
  • Geological complexity (simple layered structures vs. complex fault systems)
  • Desired output parameters (acoustic impedance, elastic properties, reservoir characteristics)
• Examples of machine learning techniques in seismic inversion:
  • Convolutional Neural Networks (CNNs) for facies classification
  • Support Vector Machines (SVMs) for lithology prediction
  • Random Forests for porosity estimation

Rock Properties Estimation with Seismic Inversion

Primary Rock Property Estimation

• Acoustic impedance derived from seismic inversion infers:
  • Lithology (sandstone vs. shale)
  • Porosity (high impedance often indicates low porosity)
  • Fluid content (gas-filled pores typically show lower impedance than water-filled pores)
• Elastic properties estimated using pre-stack inversion methods:
  • P-wave velocity (Vp)
  • S-wave velocity (Vs)
  • Density
• Porosity estimation from inverted acoustic impedance requires:
  • Integration of well log data
  • Appropriate rock physics models (Wyllie's time-average equation, Raymer-Hunt-Gardner equation)

Advanced Property Estimation and Modeling

• Fluid substitution modeling applies Gassmann's equations to inverted elastic properties
  • Predicts changes in seismic response due to different pore fluids (oil, gas, water)
• Seismic attributes derived from inverted data characterize unconventional reservoirs:
  • Brittleness (ratio of Young's modulus to Poisson's ratio)
  • Fracture density (from azimuthal analysis of inverted properties)
• Multi-attribute analysis and neural network approaches predict additional reservoir parameters:
  • Permeability (using relationships between porosity, clay content, and pore structure)
  • Water saturation (combining resistivity logs with inverted acoustic properties)
• Time-lapse (4D) seismic inversion monitors changes in reservoir properties during production:
  • Identifies bypassed pay zones
  • Tracks fluid movement (water flood fronts, gas cap expansion)

Limitations of Seismic Inversion Results

Fundamental Limitations

• Non-uniqueness presents a fundamental limitation of seismic inversion
  • Multiple subsurface models can produce the same seismic response
  • Requires additional constraints from well data or geological knowledge
• Low-frequency component of the earth model (typically below 10 Hz) lacks constraint from seismic data
  • Requires additional information from well logs or geological models
  • Impacts the absolute values of inverted properties
• Vertical resolution limits by the bandwidth of the seismic data
  • Thin layers below tuning thickness (typically 1/4 of the dominant wavelength) challenge to resolve
  • Affects the ability to detect and characterize thin reservoirs or stratigraphic features

Data Quality and External Factors

• Noise in seismic data significantly impacts the quality and reliability of inversion results
  • Random noise (background fluctuations)
  • Coherent noise (multiples, acquisition footprint)
• Anisotropy and attenuation effects, if not properly accounted for, lead to biased estimates of elastic properties and rock parameters
  • Vertical Transverse Isotropy (VTI) in shales
  • Horizontal Transverse Isotropy (HTI) in fractured reservoirs
• Quality and distribution of well control influence:
  • Accuracy of the low-frequency model
  • Calibration of inversion results
• Uncertainties in rock physics relationships and petrophysical interpretations propagate through the inversion process
  • Affect final estimates of reservoir properties
  • Require careful validation and sensitivity analysis

Seismic Inversion Interpretation for Applications

Reservoir Characterization and Modeling

• Inverted acoustic impedance volumes delineate:
  • Stratigraphic features (channel systems, carbonate buildups)
  • Facies distributions (sand-shale transitions)
  • Potential hydrocarbon accumulations (low-impedance anomalies)
• Integration of inverted elastic properties (Vp, Vs, density) enables:
  • Lithology discrimination (through Vp/Vs ratios)
  • Fluid discrimination (using elastic impedance crossplots)
• Geobody extraction from inverted properties aids in:
  • Quantification of reservoir volumes
  • Identification of sweet spots in unconventional plays (areas of high porosity and hydrocarbon saturation)
• Seismic inversion results provide high-resolution input for:
  • Geostatistical reservoir modeling
  • Property population in 3D geological models (porosity, permeability trends)

Production Optimization and Monitoring

• Time-lapse seismic inversion results interpret to track:
  • Fluid movements (water encroachment, gas cap expansion)
  • Pressure changes (depletion zones, injection-induced pressure increases)
  • Compaction effects during reservoir production
• Inverted properties and their derivatives optimize:
  • Well placement (targeting high-quality reservoir zones)
  • Horizontal well trajectories (staying within the most productive layers)
  • Infill drilling opportunities (identifying undrained compartments)
• Seismic inversion results contribute to:
  • Reduction of uncertainty in reserves estimation
  • Economic evaluation of hydrocarbon prospects (by providing more accurate volume estimates)
• Examples of production applications:
  • Monitoring CO2 injection for enhanced oil recovery
  • Identifying bypassed pay in mature fields
  • Optimizing hydraulic fracturing designs in shale reservoirs