Advanced search
Start date
Betweenand


Modelling and optimization processes for real time implementation

Author(s):
Mylene Cristina Alves Ferreira Rezende
Total Authors: 1
Document type: Master's Dissertation
Institution: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Química
Defense date:
Examining board members:
Paulo Jorge da Silva Bartolo; Caliane Bastos Borba Costa; Edson Tomaz; Valdir Apolinario de Freitas
Advisor: Rubens Maciel Filho
Abstract

Control and optimization algorithms can be used separately or may be altematively integrated arranged in a hierarchical strategy, leading to a Real Time Integration in order to solve on-line optimization and cçmtrol problems. Integration of Control and Optimization in Real Time can be carried out basically through two main strategies, usually named one-Iayer approach and two-layer approach. In the case of two-Iayer approach, the upper level is the optimization one, which determines the optimal set-points of the variables at the steady-state which are then used in the control of the system. In the one-layer approach, the economical optimization problem and the problem of control are solved simultaneously. The main case study employed in this work is a three phase catalytic reactor in which the hydrogenation of o-cresol cresol producing 2-methyl-cyclohexanol occurs. The deterministic model that represents the reactor, as well as the dynamic behavior of the reactor, showing the influence of input variables on output variables, is presented. The non optimized steady-state of the reactor in terms of productivity and reagent conversion is also showed. This reactor is characterized by a high dimensionality and non-linearity mo deI, which is difficult to be optimized by conventional methods, since not always convergence is achieved. This justifies the use of an evolutionary method, based on the Genetic Algorithms, to deal with this process. In this way, in order to optimize the process, the Genetic Algorithm code is coupled with the rigorous model of the reactor. The aim of the optimization through Genetic Algorithms is the searching of the process inputs that maximizes the productivity of 2-methyl-cyclohexanol subject to maximal conversion of o-cresol. The Genetic Algorithms are used in binary and real encoding forms. For both, a study of the Genetic Algorithm input parameters through factorial design is proposed in order to determine the significant parameters that have influence on the performance of GA to optimize the reactor. The variables optimized by Genetic Algorithms are used as possible manipulated variables and as set-points in the control of the reactor with the aim of to maintain the reactor on the optimized steady state. For this, Dynamic Matrix Control is employed. The present work presents an approximation of Takagi-Sugeno Fuzzy models to the deterministic models...Note: The complete abstract is available with the full electronic digital thesis or dissertations (AU)