Projeto multicritério do dimensionamento de redes de gás natural sob condições de incerteza da evolução da carga

Abstract

Natural gas has been increasingly used as a source of energy and presents itself as a strong trend for the future. In this context, taking into account the high cost of installing pipelines, the design of these networks requires quality solutions, relating not only financial indicators but also reliability and security in relation to demand. Thus, this thesis proposes an approach for the design of natural gas networks under conditions of uncertainty of the evolution of the load over a time horizon. A predefined network topology is assumed, where the pipe diameters define the design variables. We propose multiobjective algorithms based on the Multiobjetive Variable Neiborhood Search (MOVNS), which are contrasted with the NSGA-II considering a set of instances defined from the TSPLIB library data. The proposed methodology is applied to a real case study, and the results are compared to those obtained by three engineers of a gas company with six years of experience on average. The solutions are compared from a dominance analysis, considering the criteria: installation cost, minimum pressure, feasibility rate, average cost of failure and sensitivity. The results indicate solutions relatively different from those obtained by the engineers, presenting more robust and safe networks under conditions of uncertainties of load evolution.