Modelagem, estimação de parâmetros e incertezas na síntese de propionato de propila em um reator cromatográfico de leito fixo

Abstract

This dissertation addresses the modeling and parameter estimation, including uncertainties and confidence regions, for the Propyl Propionate (ProPro) synthesis process. The study employs a rigorous model that considers the heterogeneous nature of the investigated system, involving a fixed-bed chromatographic reactor with the solid catalyst AmberlystTM 46. The parameter estimation approach combines Particle Swarm Optimization (PSO) and a Gradient Method (GM) to estimate both thermodynamic and kinetic parameters in two distinct steps. Initially, the estimation of thermodynamic parameters occurred using literature equilibrium data. In the second step, for the estimation of kinetic parameters with the heterogeneous model, author’s experimental data were employed. Uncertainties of the experimental data were evaluated and calculated for use in the estimation of kinetic parameters. Through PSO, parameters were estimated and their confidence regions determined, providing a comprehensive understanding of their uncertainties. The uncertainties of thermodynamic and kinetic parameters are evaluated and propagated to the activities in the thermodynamic model and to the Arrhenius constants and reaction rate of the heterogeneous kinetic model. The results demonstrate good agreement between the model predictions with the parameters estimated in this work and the experimental data, adequately reflecting the statistical variability of the experimental observations with an expanded uncertainty of 95%. Moreover, the heterogeneous Langmuir-Hinshelwood model presents an improved representation of the experimental data, both in transient and steady states, compared to the Pseudo-Homogeneous model used in the literature. The Root Mean Square Deviation (RMSD) for parameter estimation shows a significant improvement compared to the values found using parameters available in the literature.