Novos métodos para reconfiguração das redes de distribuição a partir de algoritmos de tomadas de decisão multicriteriais

Abstract

The electric networks reconfiguration has as purpose: to reduce the energy losses, to improve the voltage profile for the customers, to increase the reliability levels or to re-establish the energy supply in contingency situations. In spite of the complexity of this combinatorial problem of large dimensions, there is an effort of the researchers in the direction of efficient optimization methods, once the classic methods present difficulties for its resolution. The disadvantage of these methods is that a major part of researchers ignores the totality of formal information that is available in the power utilities. The efficient solution should also consider the knowledge and experience of power utilities professionals that can be accomplished starting from the development of Specialist Systems. For these reason, it becomes more recommended to consider algorithms that make possible the use of formal and informal analysis methods. This context motivated this thesis whose proposal is the development of new methods and software for distribution networks reconfiguration starting from multiple criteria decision-making methods, aiming at a more adequate processing of all available information in power utilities. To solve the approached problem, this paper uses the Bellman-Zadeh s method that was shown efficient in the solution of several problems of multiple criteria decision, besides allowing the analysis, at the same time, of quantitative and qualitative criteria. In general, the reconfiguration can not be solved in an efficient way without the analysis of the following problems: the appropriate modeling of the equipments and electrical loads; the creation of algorithm for representation of the electrical topology that allows fast changes in the electric configuration; calculations of the load flow; the definition of composition and presentation forms for the target functions and restrictions; the development of optimization methods and software to define the best electric configuration in normal operation state and in contingency situations. The proposed methodologies were customized according to the power utilities needs, helping the operators in the decision-making process. Besides that the algorithms meet the demand for fast processing of the efficient solutions, which is fundamental mainly in the re-establishment of energy supply during contingencies. Among the several advantages of the developed software in relation to the conventional softwares, the most important are: the representation of the electric topology considers only the information of the connections that really exist, dispensing the use of matrixes; integration with the electric load modeling systems for maximum demand as for load typical curves; the use of the initial data reliability to coordinate the tranformers load; application of multiple criteria optimization algorithms for distribution networks reconfiguration, in normal operational state and in contingency situations; composition of the target functions and restrictions considering quantitative and qualitative characteristics; evaluation of the impact of the reconfiguration in the power systems energy losses. As results, case studies are presented with real data from the power utilities, in order to prove the efficiency of the proposed methods.