Modelagem, simulação e controle de reatores cataliticos de leito fixo




In this work, pseudo-homogeneous bi- and uni-dimensional fixed bed catalytic reactors are presented. Such models incorporate the fluid and solid thermal capabilities, (pCp)f e (pCp)s, respectively. The proposed models reproduce a dynamic behaviour expected in a fixed bed catalytic reactor, including the inverse response phenomenon. The bi-dimensional models solution usually requires excessive computing (CPU) time since a detailed description of the variables is considered in the formulation of the model, which results in mathematical models of difficult solution. To circumvent this difficulty, techniques of model reduction were used, which permitted the generation of smaller space dimensions (with dimension reduction) or whose solution is less complex (without dimension reduction). Therefore, a comparative work of both techniques performance is also presented in this work. Some of the reduced models presented a good agreement with the bi-dimensional model, thus enabling their safe and fast utilization for studies of control and optimization. Another study performed in this work was the parametric sensitivity analysis of the fixed bed catalytic reactor, with the purpose of knowing the primary effects and the interaction of the reactor operating parameters, which are necessary for the implementation of a suitable control strategy. A method of factorial planning was used to acquire this knowledge. It was capable of getting the primary and interaction effects on the reactor dynamic behaviour. Another important detail is that the complete factorial method also provided some effects against time; this kind of knowledge is not available when the method RGA is used. Last, with the development of dynamic models for applications in real time and with the knowledge of the dynamic behaviour and the parametric sensitivity analysis, an analysis of the performance of some advanced control algorithms was made. In the study of the analysis of the performance of advanced controllers, the control algorithms used were predictive, DMC and GPC, and adaptive predictive, STGPC (GPC coupled to the parameters identifier RLS), both in the monovariable (SISO) and multivariable (MIMO) forms. These controllers were also compared to the classic controller PI. It was observed that the controllers GPC and STGPC provided the best results, showing the need of using advanced controllers to obtain a reliable and efficiente behaviour for the fixed bed catalytic reactor


controle de processos quimicos reatores quimicos simulação (computadores)

Documentos Relacionados