Menu:

More Links:

Mining and Petroleum

Version: 0.1
(Aug 8, 2007)

Research

See the CCG Research page for up-to-date information.  This is an old summary.

My research work relates to theory and application of geostatistics to earth sciences data:

The following reflects my recent work in the field of petroleum reservoir characterization. I will continue in this area as well as diversifying more into applications to mining and environmental. Sometime in the future, geostatistics-based modeling techniques will be routinely applied to construct numerical geological models and flow models of reservoirs that incorporate both field data and appropriate analog reservoir and outcrop data.

These studies will provide the geologic basis for reservoir performance predictions and help assess the uncertainty of the predicted performance.

The first aspect of this is the creation of detailed numerical 3-D geologic models that combine a wide range of relevant geological, geophysical, and engineering data of varying degrees of resolution, quality, and certainty. 


The second aspect is to translate such finely gridded numerical geological models of rock properties to coarsely gridded flow models
that reproduce the historic production behavior of the reservoir and better predict its future performance.

The following seven study areas are my arbitrary attempt to organize research to achieve this vision: 

Reservoir Size and Compartmentalization:

Stochastic (sub-seismic scale) faults and their impact on compartmentalization and reservoir recovery: involves geological concepts to describe the physical heterogeneities, flow modeling to get impact on performance, and geophysics for indirect measurements Surface modeling constrained to soft data Multiple surfaces (isopach/isochore mapping); automated geometric modeling for stratigraphic framework and uncertainty assessment Bootstrap or Monte Carlo procedures for reservoir volume uncertainty.

Conceptual Geological Model / Correlation Structure:

Hierarchical pseudo-genetic techniques that mimic original sedimentation: build on fundamental geological principles with the goal of numerical models at the reservoir scale ``Deterministic'' models for known geological features (surfaces and trends) combined with stochastic models Stochastic surfaces (existing and restored); quantify uncertainty in trend maps/profiles Expert systems to quantify ``expert'' geological knowledge and appropriate conceptual data.

Rock-type Modeling:

Hybrid or multi-stage techniques that consider outcrop images or object-based realizations as sources for spatial statistics (use large collected experience from outcrops, guidelines for collecting relevant data, condition to local data) ``Pseudo-genetic'' techniques that mimic original sedimentation Object-based for deltaic, fluvial, and eolean sediments Measures of drainable and floodable volumes in carbonate reservoirs with little data Account for seismic data informing vertically averaged proportions of facies/lithology with uncertainty Better tools for uncertainty quantification (modeling approach, parameters, heterogeneity)

Assignment of Static Petrophysical Properties:

Prior integration of production data (use ``inverted'' coarse 2-D data to constrain 3-D modeling) Consider ``missing scale'' problem with non-linear low-entropy geological features Rigorous treatment of seismic data (scale and precision) Realistically honor trend maps/profiles Integration of well test data (a type of production data) Modeling fractured reservoirs Outcrop data (measurements of different type, optimal sample spacing and layout to describe spatial statistics)

Gridding for Flow Simulation:

Adaptive grid specification that minimizes off-diagonal tensor permeability terms (account for pressure field and original high resolution description of heterogeneities) Guidelines for gridding (1) relative sizes of blocks, (2) truncating layers, (3) non-orthogonality constraints, and (4) minimize magnitude of off-diagonal tensor Tensor properties for blocks not aligned with geologic continuity Numerical problems with dynamically changing grid specification.

Upscale Static Petrophysical Properties:

Dynamic scale-up in flow simulator - one way to honor the detailed geological description and adapt to changing boundary conditions Direct scale-up, with no spatial truncation error, that allows flexible boundary conditions (full and diagonal tensor) Scale-up from core plug to geological modeling cell with no bias due to implicit boundary conditions Accounting for connected extreme high and low permeabilities.

Dynamic Flow Properties:

Phase-effective-permeability model that is saturation and history dependent Sophisticated pseudos (match-derived pseudos that capture history and local resolution) Process-independent pseudo functions.