Optimization-based framework for modeling and kinetic parameter estimation

Abstract

To understand the mechanism of a catalytic system, computational analysis is essential. Additionally, process simulators play a fundamental role in the reproduction of models associated with various unit operations such as chemical reactors. However, due to the complexity that some chemical and kinetic reactions may present, it is often complicated to design a reactor considering stoichiometric and chemical kinetic relationships resorting to yield-based or Gibbs-type reactors. This barrier has limited on several occasions the adequate reproduction of kinetic processes, their dynamic study, and even the direct application of process intensification strategies. This work presents a generic optimization framework for solving parameter estimation and catalyst design problems. Using a stochastic optimization method in conjunction with Aspen Plus and experimental data, kinetic data values were predicted. The direct advantage of obtaining kinetic parameters with this methodology is to be able to represent the kinetics according to the Arrhenius equation, which is widely used in commercial simulators. To validate this method, a previously presented case of dimerization of isobutane to produce isooctane is used as a baseline. Once the proposed methodology was used, it was possible to obtain kinetic parameters like those obtained by the authors of the case study with an error lower than 0.1%. Finally, the strategy to determine the kinetic parameters of a more complex case, the kinetic values of the oligomerization and hydrogenation reactions present in the aviation biofuel production process by the ATJ method were developed obtaining results like those reported experimentally with an error percentage lower than 1%, and with statistical metrics with relatively good approximation.