Browsing by Autor "Kevin Vargas"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item type: Item ,
    Analysis of Fractality in Water Distribution Networks Using Hydraulic Criteria
    (2019) Kevin Vargas; Juan Saldarriaga
    The fractal dimension in water distribution networks (WDN) can be seen as an indicator or a scaling factor of how much the network changes topologically, according to the scale in which it is analyzed. The fractal dimension of a considerable number of networks with different characteristics was calculated in order to study its behavior and have a better understanding of this feature. Three different criteria were used to calculate the weight of each node in the box covering algorithm used to determine the fractal dimension: topology, SumQ, and HGL*SumQ. All of the studied networks resulted to be fractal networks since the coefficient of determination (R2) of the linear fit after applying the box covering algorithm was always above 0.95. The largest values for the fractal dimension were obtained with the topology weight criterion. There was not a clear tendency in the variation of fractal dimension when modifying the networks’ demands and calculating the weights with the proposed hydraulic criteria (SumQ and HGL*SumQ).
  • Thumbnail Image
    Item type: Item ,
    Potential District Metered Area Identification in Water Distribution Networks Using Hydraulic Criteria and the Box Covering Algorithm
    (2019) Kevin Vargas; Juan Saldarriaga
    The box covering algorithm, used to calculate the fractal dimension in water distribution networks (WDNs), was proposed as an alternative to identify possible district metered areas (DMAs) in WDNs. By using hydraulic criteria (SumQ and HGL*SumQ) in the method’s node weight calculations, tests on Cazucá, Exnet, and Santa Marta networks were carried out in order to establish possible divisions into DMAs. The quality of these divisions was measured using the modularity index. These divisions were then compared to those obtained with the community detection algorithm which guarantees the maximum modularity possible starting from every node as a single independent DMA. Divisions with relatively high modularity were obtained using the box covering algorithm, however, their modularity was never as high as the ones obtained with community detection. Therefore, the methodology proposed could be a viable alternative to identify potential divisions into DMAs in large and complex WDNs.