Andre Luiz Alberton

Fluid catalytic cracking (FCC) is a major conversion process in the production scheme of a refinery. Among the products, Light Cycle Oil (LCO) is an important middle-cut stream that may become a focus of attention if a refinery is required to increase the diesel/gasoline production ratio. The present work presents an empirical model for the FCC yield profile, exploring the low conversion region (favoring higher LCO formation). Data were taken from experimental runs performed in a large FCC pilot plant using five different batches of vacuum gasoil from a commercial FCC unit at a 200 kg/h rate and a catalytic system aimed at maximizing LCO. With regard to the empirical model, the explanatory variables associated with the reaction severity level in the Riser were: reaction temperature (T_rx), catalyst-to-oil mass ratio (CTO), and contact time between the feed and catalyst in the Riser reactor (t_c); while the light hydrocarbons content of the feed (HC_light), i.e., the wt % of feed with a boiling point lower than 343 °C, was considered as the factor associated with the feedstock characterization. The empirical model describing the catalytic cracking yield profile was developed through combinations of these factors normalized between −1 and +1, using a cross-validation Monte Carlo-based method to avoid model overfitting. Factors were accepted in the model only if their associated parameters were significant according to the statistical tests. Restricted least squares were also employed to ensure closed mass balances. Results indicated LCO as an intermediate product, with much lower yield variation compared to naphtha or slurry oil. Although there is the disadvantage of an increased slurry oil yield, operating the FCC at low conversion presents a potential opportunity to increase the diesel/gasoline production. The empirical model suited well with experimental data for most of the lumped groups (R² between 0.95 and 0.97), with a higher dispersion for LPG (R² = 0.90) and coke (R² = 0.88), and may be a useful tool to further evaluate the more profitable operation mode of an FCC unit.

In the present work, adaptive orthogonal collocation and a Monte Carlo method are used to compute the molecular weight distributions (MWD) of ethylene/1,9‐decadiene copolymers produced with a constrained geometry catalyst. Predictions from each model are compared to each other and to the experimental MWDs, allowing for the evaluation of relative strengths and weaknesses of each mathematical modeling method. Comparisons with experimental results indicate that the rate of macromonomer incorporation in the growing polymer chains decays with the macromonomer radius of gyration. In all cases, the proposed models are able to fit appropriately the available experimental MWDs.

Long chain branches improve the viscoelastic properties of polyolefins and make them easier to process. The frequency of long chain branching in polyethylenes made with coordination catalysts can be substantially increased by copolymerizing ethylene with small amounts of nonconjugated dienes using an adequate catalyst. In this work, the kinetics of copolymerization of ethylene and 1,9‐decadiene is investigated using a constrained geometry catalyst in a solution polymerization semi‐batch reactor, and a novel mathematical model is proposed to describe the resulting molecular weight distributions. A hybrid approach, combining particle swarm optimization, parameter identifiability procedures, and the Gauss–Newton method is applied to estimate the model parameters. The proposed mechanism includes macromonomer reincorporation through pendant double bonds that result from diene incorporation. The predicted long chain branching frequencies are validated by Monte Carlo simulation. Experimental average molecular weights and ethylene feed flow rates are successfully predicted by both methods. image

Monte Carlo methods are heuristic algorithms that use probabilities to select an outcome among several possible events in a given process. Monte Carlo methods are useful in polymer reaction engineering because they can predict the molecular architecture of polymers with details that cannot be easily captured by any other modeling technique. One of the advantages of Monte Carlo simulation is that one does not need to solve differential or algebraic equations to predict the microstructures of polymers. This article reviews the literature on steady‐state and dynamic Monte Carlo methods in polymer reaction engineering. We hope to convince the readers that playing dice regularly can be a great asset to polymer reactors engineers.

Different values of the kinetic rate constants have been reported for initiator decompositions carried out in microwave and conventional batch reactors. The present work analyzes the confidence regions of kinetic parameters estimated for the thermal decomposition of two initiators, potassium persulfate (KPS) and 2‐2′‐azobis(2‐methylpropionamidine) dihydrochloride (V‐50), which confirm the difference between the parameters obtained in microwave and conventional reactors. Moreover, by assuming no changes in kinetics under microwave heating, the temperature and volume of hot spots required to explain the apparent increase of kinetic rate constants are calculated. The results suggest that formation of hot spots may explain the observed effects without the occurrence of additional phenomena.

This work proposes a procedure for simultaneous parameters identifiability and estimation in metabolic networks in order to overcome difficulties associated with lack of experimental data and large number of parameters, a common scenario in the modeling of such systems. As case study, the complex real problem of parameters identifiability of theEscherichia coliK-12 W3110 dynamic model was investigated, composed by 18 differential ordinary equations and 35 kinetic rates, containing 125 parameters. With the procedure, model fit was improved for most of the measured metabolites, achieving 58 parameters estimated, including 5 unknown initial conditions. The results indicate that simultaneous parameters identifiability and estimation approach in metabolic networks is appealing, since model fit to the most of measured metabolites was possible even when important measures of intracellular metabolites and good initial estimates of parameters are not available.