Using Multiagent System to Build Structural Earth Model

  • Authors:

  • Beiting Zhu;Zahia Guessoum;Michel Perrin;Bertrand Braunschweig;Pierre Fery-Forgues;Jean-François Rainaud

  • Affiliations:

  • Institut Français du Pétrole, Rueil-Malmaison, France 92852;Laboratoire d'Informatique de Paris 6, Paris, France 75016;Ecole des Mines de Paris, Paris, France 75272;Institut Français du Pétrole, Rueil-Malmaison, France 92852;Institut Français du Pétrole, Rueil-Malmaison, France 92852;Institut Français du Pétrole, Rueil-Malmaison, France 92852

  • Venue:
  • Agent Computing and Multi-Agent Systems
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Mainly based on seismic data, structural earth models are commonly used for oil & gas reservoir engineering studies. The geometry and the topology of a structural earth model strictly depend on the geological characteristics of its various surfaces. These characteristics constitute the geologists' interpretations of the earth model to be built. In the 3D modeling applications currently used in industry, these interpretations are often registered after the completion of the structural earth model. Therefore further changes in the interpretations can hardly be introduced without reconstructing the structural model from scratch. This situation is dramatic since the only true geometrical information concerning subsurface is that issued from drillings, which daily bring new data.For the above reasons, there is a great interest in providing computer-aided tool for validating and improving geological interpretations to facilitate the geologists' modeling work. This paper presents a distributed problem solving approach, which enables geologists to quickly validate and improve structural earth models by considering new drilling data. This approach is based on a self-organizing multiagent system and on domain-specific expertise.