Search

O rastreamento de referências polinomiais é muito útil em aplicações como em sistemas mecânicos, químicos e de navegação. Em geral, este problema de rastreamento é difícil de executar. Através de técnicas de controle preditivo generalizado (GPC) é possível desenhar controladores para o rastreamento de referências polinomiais. Porém, a rejeição de perturbação também é um aspecto importante no desenvolvimento de sistemas de controle. O presente artigo aplica e avalia o comportamento de um controlador preditivo generalizado para rastreamento de referência polinomial (Poly-GPC) ante perturbações. Simulações foram efetuadas considerando a presença de perturbações tipo degrau tanto na entrada como na saída da planta. Os resultados indicam que o Poly-GPC possui a capacidade de rastrear referências polinomiais enquanto as perturbações são rejeitadas, sendo que a rejeição depende também da sintonia do controlador preditivo.

Recalcitrant compounds found in pulp and paper mill effluents, which are usually generated at high temperatures, have shown the potential for thermophilic treatment. Using bleached kraft pulp mill effluents, this study aims to evaluate the acclimation conditions of mesophilic aerobic sludge to thermophilic conditions and determine the kinetic parameters. Two feeding strategies were used for this purpose: in the first approach, the first reactor R1 was fed with bleached kraft pulp effluent, while in the second strategy the reactor R2 was fed with easily biodegradable synthetic effluent and then the gradual substitution for bleached effluent was performed after sludge acclimation to a thermophilic environment. Both reactors were operated in batch at 55ºC. The second strategy has a slight advantage in COD (chemical oxygen demand) degradation and a considerable higher biomass production. The kinetic parameters (KS, µmax, Y, Kd) for both approaches were obtained using the Monod model. Those values indicated that the approach used for reactor R2 provided higher reaction speed.

Resolver is an angular position sensor widely used in applications such as electric/hybrid vehicles, CNCs, antennas and robotics. However, the estimation of the angular position from resolver outputs is more difficult than the analysis of encoder signals, and it is still an open question. Most algorithms proposed in literature are based on type-I or type-II angle tracking ob- servers. Some type-III observers were proposed, but they require a high sampling frequency. This paper explores the use of synchronous demodulation of the resolver outputs to simplify the implementation of a type-III angle tracking observer. The resolver outputs are sampled at the peaks and valleys of the excitation resolver signal, being easy to get sine and cosine of the angular position. The proposed approach reduces the computational cost and the required sampling frequency to implement the type-III observer. Simulation and experimental results prove the accuracy of the proposed approach.

Sinusoidal reference tracking is required in important practical applications. Resonant controllers are suitable for that task. However, stable implementation of these controllers in a digital processor is difficult to achieve. On the other hand, quasi-resonant controllers are easier to implement, but they do not guarantee asymptotic reference tracking. This paper proposes a new resonant controller based on Generalized Predictive Control (GPC) and Internal Model Principle (IMP). The first-order and second-order difference operators are applied to create a GPC system whose structure embeds the Z-transform of the sinusoidal reference. Thus, according to IMP, the proposed controller, named Resonant Generalized Predictive Control (RGPC), asymptotically tracks sinusoidal references. Besides, the proposed RGPC has the advantages of GPC, such as fast response and easy implementation in digital processors. Experimental tests prove that the proposed RGPC has better performance than other controllers for sinusoidal reference tracking.