Application of GA in optimization of pore network models generated by multi-cellular growth algorithms

Jamshidi, S ; Sharif University of Technology | 2009

121 Viewed
  1. Type of Document: Article
  2. DOI: 10.1016/j.advwatres.2009.07.007
  3. Publisher: 2009
  4. Abstract:
  5. In pore network modeling, the void space of a rock sample is represented at the microscopic scale by a network of pores connected by throats. Construction of a reasonable representation of the geometry and topology of the pore space will lead to a reliable prediction of the properties of porous media. Recently, the theory of multi-cellular growth (or L-systems) has been used as a flexible tool for generation of pore network models which do not require any special information such as 2D SEM or 3D pore space images. In general, the networks generated by this method are irregular pore network models which are inherently closer to the complicated nature of the porous media rather than regular lattice networks. In this approach, the construction process is controlled only by the production rules that govern the development process of the network. In this study, genetic algorithm has been used to obtain the optimum values of the uncertain parameters of these production rules to build an appropriate irregular lattice network capable of the prediction of both static and hydraulic information of the target porous medium. © 2009 Elsevier Ltd. All rights reserved
  6. Keywords:
  7. Genetic algorithm ; Irregular lattice networks ; L-system ; Multi-cellular growth algorithm ; Optimization ; Pore network modeling ; Porous media ; Genetic algorithms ; Porous materials ; Uncertainty analysis ; Three dimensional ; Fluid flow ; Porous medium
  8. Source: Advances in Water Resources ; Volume 32, Issue 10 , 2009 , Pages 1543-1553 ; 03091708 (ISSN)
  9. URL: https://www.sciencedirect.com/science/article/abs/pii/S0309170809001171