Implementação de uma estrutura de controle indireto da concentração de etanol em uma coluna de destilação híbrida

• Ano de defesa: 2013
• Autor(a) principal: Gehlen, Cindi de Oliveira
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; BR; Engenharia de Processos; UFSM; Programa de Pós-Graduação em Engenharia de Processos
• Palavras-chave em Português: Controlador PID; Inferência; Identificação; Destilação; PID controller; Inference; Identification; Distillation; CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA
• Link de acesso: http://repositorio.ufsm.br/handle/1/7958

The unit operation of distillation demands the largest amount of energy in the ethanol separation. It is of utmost interest to minimize the energy expended in the process, playing the control an important role on this task. Here a Proportional-Integral- Derivative (PID) controller is proposed to ensure the fuel ethanol composition at desired value, i. e., above 92,5°INPM as the same time as save energy through the minimal use of reboiler power. It is well known that the composition analyzers are expensive and have a high response time. Due to such limitation, we have developed an inference model in which the product composition is inferred from the product temperature and thermodynamics relations, enabling the monitoring and the indirect control a posteriori of this variable. In addition, the inference of composition allows defining the reference value of product temperature which ensures the fuel ethanol production at desired composition. The chosen control structure takes into account the reference value of product temperature as set-point of a SISO (Single Input Single Output) controller, i. e., the product ethanol composition is indirectly controlled by the direct control of product temperature. The control design depends on a good dynamic model between the manipulated variable (reboiler power) and the controlled variable (product temperature). Such relationship was obtained by process data identification with discrete parametric models, being Output Error and ARMAX the models with the best performance. Several tuning methods were compared for the controller design. Among these methods, the Internal Model Control (IMC) has presented the best relationship between robustness and performance to be used in the distillation process of this study. Finally, the proposed controller was validated in the process and the results have shown its good performance to obtain the product composition at desired value. Thus, we can conclude that the proposed controller, through the simple PID control strategy, performed as expected within the limitations of available instrumentation and the process particularities.