Lithologic Inversion

Transforming Traditional Seismic Inversion Approaches

Advanced Reservoir Property Estimation Without LFM Dependency

TraceSeis has developed a groundbreaking methodology for estimating reservoir properties that overcomes key limitations of conventional approaches. Our innovative technique delivers more reliable results with significantly reduced user input and simplified parameterization.

Traditional Reservoir Property Estimation Challenges

Conventional seismic inversion typically follows a two-stage process:

  1. First Stage: Elastic Inversion
    • Computes rock properties (P- and S-impedances, density) through complex mathematical processes
    • Requires minimizing differences between observed and modeled data
    • Incorporates AVO, offset-varying wavelet, and Low Frequency Models (LFM)
    • Uses non-intuitive parameters in available software
    • Creates sensitivity issues that are poorly understood by typical users
    • Often requires iterative parameter modifications and well-log comparisons
  2. Second Stage: Reservoir Property Estimation
    • Converts rock properties to reservoir properties through various approaches
    • May use qualitative methods based on cross-plots or multivariate clustering
    • Sometimes employs model-based deterministic approaches using effective media relationships
    • Often relies on empirical relationships fitted to well-log or core measurements
    • Results are applied to 3D volume for reservoir property estimates

The traditional approach is heavily analyst-dependent, with reliability strongly tied to geoscientist experience. When different specialists perform the rock property and reservoir property analyses, additional uncertainty is introduced.

The LFM Dependency Problem

The Low Frequency Model (LFM) in traditional methods:

  • Provides low frequency components (including DC) on which seismic-derived relative changes are superimposed
  • Operates in frequency bandwidths outside seismic ranges
  • Remains mostly unchanged during inversion
  • Is created from non-reflectivity data (well-logs, seismic velocities)
  • Has magnitude several times larger than relative rock property changes
  • Can introduce significant errors (small LFM inaccuracies can create errors as large as the range of variation
Figure 1. Customary and proposed inversion workflows

.

TraceSeis’s Innovative Solution: Direct Relative Property Analysis

Our methodology revolutionizes this process by using relative rock properties to compute reservoir properties directly:

Key Advantages:

  • Simplified workflow with reduced user input requirements
  • Enhanced accuracy through separate execution of elastic inversion components
  • Modified execution order of processing elements
  • Reduced LFM dependency by incorporating it only when necessary in later stages
  • Direct computation of reservoir properties without absolute rock property estimation
Figure 2. QC for reservoir properties estimation
 .

Our Refined Process:

  1. Modified Pre-conditioning:
    • Seismic wavelet is offset-equalized and phase-corrected in initial data conditioning
    • Processing occurs prior to computing relative properties
  2. Relative Property Focus:
    • We bypass absolute rock property estimation entirely
    • The LFM is incorporated only if necessary during reservoir property computation
    • This approach relaxes the need for a rigorous LFM
  3. Linear Relationship Determination:
    • Reservoir properties are computed as linear combinations of relative rock properties
    • Parameters are established through least-squares fitting (regression analysis) of well-log data
    • When required, the LFM is input as one of the properties in regression analysis
    • The LFM can be any low-frequency property that mimics the reservoir property of interest
    • For example, P-wave velocity from seismic often serves as LFM when computing total porosity
  4. Volume Property Calculation:
    • The linear relationship from well-log analysis is applied to the seismic volume
    • Relative properties from seismic are obtained by integrating reflectivities of corresponding rock properties
    • These reflectivities are derived through analytical transforms of AVO attributes

Comprehensive Applications and Results

This methodology can quantitatively estimate any reservoir property that can be represented as a linear combination of relative rock properties, including:

  • Effective porosity
  • Brittleness
  • Mineralogy (including volume fraction of kerogen)

TraceSeis provides comprehensive quality control displays that validate the estimation process at both the well-log analysis and seismic application stages.

By bypassing absolute rock property estimation and relaxing the need for rigorous low frequency models, TraceSeis delivers more reliable reservoir characterization with simplified workflows and reduced analyst dependency.