Controle adaptativo robusto de estrutura variável por modelo de referência aplicado a filtros ativos de potência

• Ano de defesa: 2010
• Autor(a) principal: Stefanello, Márcio
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; BR; Engenharia Elétrica; UFSM; Programa de Pós-Graduação em Engenharia Elétrica
• Palavras-chave em Português: Controle adaptativo; Controle de estrutura variável; Filtro ativo de potência; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA
• Link de acesso: http://repositorio.ufsm.br/handle/1/3664

This works presents a Robust Variable Structure Model Reference Adaptive Control scheme denoted VS RMRAC. In this adaptive control strategy, each parameter of the control law is obtained by the sum of two other adaptive terms. One of them is associated with the RMRAC algorithm, while the other is composed of an adaptive gain rated by an approximated sign function, being associated with the VS controller. The conventional RMRAC has the property of robustness with respect to certain types of unmodeled dynamics, while the conventional VS improves the system performance and can use certain information about the plant uncertainties to attain a sliding surface and invariance of the states of the closed loop plant with respect to these uncertainties. The main motivation for combining the RMRAC and VS schemes is to obtain an adaptive control strategy with composite properties, namely, robustness and fast response, which are conflicting each other in conventional adaptive algorithms. The method of combination was already proposed in the literature, however, unlike the original proposal, the adaptation laws are modified to cope with the effect of the unmodeled dynamics, which are by hypothesis, of multiplicative and additive types. The modification includes normalization for robustness with respect to such dynamics. For the discrete time case it is also necessary a modification on the error equation for causality of the adaptive algorithm. This modification implies a suitable adaptive algorithm, designed according to the Gradient approach. The stability proofs are stated for the continuous time case, as well as for the discrete time case. The performance of the overall controller can be verified quantitatively, from the analysis of the robust stability, as well as qualitatively, from simulations and analysis of experimental results. Regarding to the last, the performance of the RMRAC and VS RMRAC algorithms are tested for the current control of an Active Power Filter connected in parallel with the grid. Once the filter is connected to the grid, its is subject to parametric uncertainties, unmodeled dynamics and disturbances. Besides, the references for the compensation currents are complex waveforms, including several harmonic components of low order. For these reasons, stability and performance are mandatory in applications of Active Power Filters and this system was elected to experimentally test the performance of the VS RMRAC adaptive algorithm.