Estudo de propagação harmônica em redes de distribuição considerando a característica estocástica das cargas

• Ano de defesa: 2015
• Autor(a) principal: Marcio Melquiades Silva
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Método de Monte Carlo; Qualidade da energia elétrica; Simulação de sistemas elétricos; Medições em sistemas elétricos; Harmônicos em sistemas elétricos; Modelo estocástico de cargas; Redes de distribuição da energia elétrica; Engenharia elétrica; Energia eletrica Distribuição
• Link de acesso: http://hdl.handle.net/1843/BUBD-A2DP6F

The main aim of the thesis is to formulate and develop of a new methodology for modelling and handling with aggregated harmonic loads and linear loads in harmonic power flow analysis. This method, so-called Variable Load Method, can be used in balanced and unbalanced systems and for deterministic and stochastic studies. The main types of harmonic loads and their participation coefficients inside the facility are estimated by inspection or based on usual values. In addition, a new type of load model that represents the load variation and minor loads is also defined based on field measurements, in order to build a more adherent harmonic load representation. A stochastic analysis, based on Monte Carlo simulations, is also performed. This sequence of steps makes possible to evaluate if the harmonic distortion has a significant probability of exceeding the limits set by standards due to load variations. This analysis is performed jointly with the bootstrapping method for obtaining the probability density functions of the voltage harmonic distortions. Tests in real plants show that the proposed approach is able to represent measured data with high precision. For these studies new software was development in Matlab environment with graphical interface. The program allows the study of harmonic propagation in radial electric network, in both balanced and in unbalanced networks, and showed similar results to the other referral program when simulating typical IEEE systems. The program also allows multiple simulations using different loads to the same network settings obtained at random. Simulation of cases examples shows realistic results.