Research

---

The growing challenges of data complexity, availability and uncertainty vary according to the E, D&P stages. This research theme addresses these challenges by focusing on R&D of:

(1)

Interactive data visualization (i.e. techniques for visualizing 2D, 3D and high-dimensional reservoir data and models) and analytics (i.e. analytical reasoning about the reservoir data), both facilitated by interactive visual interfaces.

(2)

Integrated visualization & visual analytics techniques, establishing a coherent workflow through complex processes, indicating the level of uncertainty in a range of features, visually highlighting data interpretations as well as anomales


Selected Projects w/ RT-2 as the primary Research Theme

Industry data are exponentially increasing in volume and complexity, representing diverse information, high dimensionality and varying levels of uncertainty. The data are acquired and structured at different stages of exploration, development and production (E, D&P) (Fig 1 and Fig 2), in different modalities, or types (e.g. grids, surfaces, volume data, point clouds) and scales, based on the various data sources. We focus on R&D of two main categories of projects:

(1)

Interactive Visualization of Reservoir Geology & Dynamics.
We focus on R&D of interactive reservoir visualization tools for E, D&P tasks to improve the understanding of how geological heterogeneities impact the flow behavior in petroleum reservoirs, with particular interest on carbonate reservoirs. Their characterization is associated with (1) a high degree of uncertainty, due to complex lithology, structure, pore network, and multi-scale heterogeneities; and (2) datasets from multiple sources such as seismic, wire-line logs, borehole imaging, petrophysics, well-tests, core analysis, multiple reservoir flow simulations, and outcrop analogue information.

(2)

Interactive Visual Analytics of High-Dimensional E, D&P Data. We focus R&D of interactive visual analytics tools for E, D&P tasks involving high-dimensional industry datasets - i.e., large number of data samples, in which each individual sample contains dozens or hundreds of attributes. Analysis of high-dimensional data typically takes the form of extracting and enabling interactive, exploratory visualization of correlations between data samples, discovering meaningful information in data, clustering data samples, and finding efficient representations of clustered data, classification and event association. Our projects and case-studies involve three categories of E, D&P industry data resulting from:

  • numerical simulations -- selected case-studies include reservoir simulation data, including post-processing, history matching, and geological modeling for flow simualtion case studies);
  • observed & processed data -- selected case-study include microseismic event monitoring data;
  • existing databases -- selected case-study include petrographic datasets from large petrological databases.


Selected Publications & Presentations

Amorim, E., Vital Brazil, E., Nonato, L. G., Samavati, F. & Costa Sousa, M. [2014]:
Multidimensional Projection with Radial Basis Function and Control Points Selection.
Proc. of the
7th IEEE Pacific Visualization Symposium (PacificVis '14), pp. 209--216
Yokohama, Japan, March 2014

Cevolani, J. T., Mostafa, A. E., Vital Brazil, E., de Oliveira, L. C., da Fonseca, L. G. & Sousa, M. C. [2013]:
Computational Methodology to Study Heterogeneities in Petroleum Reservoirs.
75th EAGE Conference & Exhibition / SPE EUROPEC 2013 (SPE 164865).
London, UK, June 2013

Hajizadeh, Y., Amorim, E. & Sousa, M. C. [2012]:
Building Trust in History Matching: The Role of Multidimensional Projection.
74th EAGE Conference & Exhibition / SPE EUROPEC 2012, (SPE 152754).
Copenhagen, Denmark, June 2012.

Mostafa, A. E., Carpendale, S., Vital Brazil, E., Eaton, D., Sharlin, E., Sousa, M. C. [2013 b]:
FractVis: Visualizing Microseismic Events.
Advances in Visual Computing, Lecture Notes in Computer Science
9th International Symposium on Visual Computing., Springer Berlin Heidelberg. Rethymnon, Crete, Greece, July 2013

-----

Somanath, S., Carpendale, S., Sharlin, E. & Costa Sousa, M. [2014]:
Information Visualization Techniques for Exploring Oil Well Trajectories in Reservoir Models
Proc. of
Graphics Interface (G I'14), pp. 209--216
Montreal, Canada, May 2014

Vital Brazil, E., Macedo, I., Costa Sousa, M., Velho, L. & de Figueiredo, L.H. [2011]:
Shape & Tone Depiction for Implicit Surfaces.
Computers & Graphics - NPAR 2010 Special Issue 35(1): 43--53.