Análise matemática e aplicação da contribuição dinâmica na técnica para projeto de malhas de controle de processos.

Abstract

Large-scale industrial production has only become possible thanks to technological evolution and the creation of process control structures suited to this demand. In this sense, increasingly

sensitive sensors combined with high-performance and high-speed data processors and high- precision actuators have made possible the emergence of control configurations for industrial

processes that would have been unimaginable a few decades ago, such as multivariable

predictive control. However, several industrial plants still use control strategies based on multi- loop control, either because of the lack of an adequate process model, or because this

application already meets the robustness and performance requirements expected for the process. This work aimed to analyse the inclusion of dynamic process information in the definition of control pairing to be used in a multi-loop control strategy. Parameters such as the stability of the control system, the RGA, the system response in the frequency domain and the conditions for a control system to be considered decentralized were used as metrics to compare the proposal with other pairing procedures found in the literature. It was observed that the systems proposed by the technique met the necessary stability requirements, not fulfilling, however, the sufficient requirements that ensure generalized diagonal dominance. It was also possible to verify that the proposed systems present suitable pairs to the servo characteristic of the controller, evidenced by the PRGA matrix, as well as a good disturbance rejection seen in the CLDG matrix and in the response in the frequency domain.