Linha divisórias de águas e fraturas de caminhos ótimos em meios desordenados

• Ano de defesa: 2012
• Autor(a) principal: Oliveira, Erneson Alves de
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Física estatística; Fractais; Percolação; Statistical Physics; Fractals; Percolation
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/9656

In the nature all material breaks down depending on the value of stress applied. Depending of kind, shape and other characteristics of the material or even the stress point, we can produce distinct {it fractures}, like a tear on stressed sheet of paper, a congestion in the network traffic of a city or cracked soils by arid climates. Such fractures are economically related with the extraction of oil from the underground reservoirs, with the extraction of heat and steam from geothermal reservoirs and even the preservation of the groundwater. Phenomenologically, we can imagine that fracture processes are the ones that divides the system in two or more parts, destroying its global connectivity. In this context, we built two computer models to study, characterize and elucidate the behavior of natural phenomena similar to fracture processes. In the first model, we explored concepts of invasion percolation applied to description of the irregular geometry of the ridge of mountains that divides hydrographic basins. We shown robustly the self-similar nature of the watershed lines, with fractal exponent $D=1.21pm0.001$ for artificial uncorrelated landscapes and, $D=1.10pm0.01$ and $D=1.11pm0.01$, for real correlated landscapes of the Swiss Alps and the Himalaya Mountains, respectively. In the second model, we used optimal paths that are cracked sequentialy providing the collapse of the system, producing a percolating fracture. In the two-dimensional case, we considered artificial uncorrelated landscapes in the weak and strong disorder. In both regimes, we obtained the same fractal exponent for the backbone fracture, $D=1.22pm0.01$. For artificial correlated landscapes, we found that the fractal dimension of the backbone decreases with increasing of the {it Hurst} exponent. In the three-dimensional case, we considered only artificial uncorrelated landscapes with strong disorder. In this case, we obtained a percolating surface with fractal dimension $D=2.47pm0.05$ that cracks the system in two parts.