(186l) Evaluation of Genetic Algorithm Coding to Optimization of a Large Scale Dynamic Systems | AIChE

(186l) Evaluation of Genetic Algorithm Coding to Optimization of a Large Scale Dynamic Systems


Victorino, I. R. D. S. - Presenter, State University of Campinas (UNICAMP)
Morais, E. R. - Presenter, State University of Campinas (UNICAMP)
Freitas Jr, B. B. - Presenter, State University of Campinas (UNICAMP)
Maciel Filho, R. - Presenter, University of Campinas, UNICAMP

Genetic Algorithms (GAs) have obtained great popularity in the last decades, since this have shown great potential and ability to solve complex problems of optimization in diverse industrial fields, including chemical engineering process. Several optimization classic techniques have been used with this objective, but generally these techniques are not efficient, mainly when the complexity of the problem is large due to non-linearity and process variable interactions. The genetic algorithms are based on the genetics and natural evolution principles of the species. The mechanism of the Genetic Algorithms technique occurs with successive modifications of the individuals or chromosomes (artificial structures) of population through the application of selection, crossover, and mutation operators. The application of Genetic Algorithms needs to develop a representative objective function of the reactor model, where this objective function evaluates the quality of determined solutions, being used as main criterion in approaches of great potential in industrial applications. As a case study is considered an industrial hydrogenation multiphase catalytic reactor to obtain a cyclic alcohol (CA). The system is analyzed for different level of operational requirements. The reactor consists of concentrical tubes, with the reactants flowing through the internal tubular part, while the coolant circulates through the inner anules. The mathematical equations of the deterministic model are based on conservation principles (mass, energy and momentum) for the reactants and for the coolant fluid and validated with real operational data and developed for the dynamic regimen. Industrial data of temperature are available and they will be used in the development and validation of the model. The main task is to develop and implement a GA code coupled in this reactor mathematical model, considering the analysis and comparison among two coding types inherent the Genetic Algorithms (binary and floating-point coding), in the optimization of reactor operational parameters. This analysis considers the use of specific genetic parameters of each coding type, and the respective influences in the reaction performance (CA productivity increase). The choice of an appropriate structure (representation or coding) for solutions (individuals) of a determined problem is a major determining factor to GAs success. The intention is to show that this technique is suitable for cyclic alcohol production maximization obtaining good results with operational improvements (reduction catalyst rate, reduction main reactant rate ? benzilic alcohol and undesired product rate ? cycloalcane), analyzing two coding types and the influence that these have in the reactor performance (CA productivity increase). The reduction main reactant rate must be considered, mainly when the emissions occur in high concentrations causing damages to the environment. The results shown that, the method is very robust procedure which allows the reactor to be operated of high level of performance, restricting the emissions of toxic reactant and reducing the undesirable products.