Dynamic Model of an Industrial Steam Reformer and Its Use for Multiobjective Optimization

Optimization of a Steam Reforming Plant Modeled with Artificial Neural Networks

The objective of this research is to improve the hydrogen production and total profit of a real Steam Reforming plant. Given the impossibility of tuning the real factory to optimize its operation, we propose modelling the plant using Artificial Neural Networks (ANNs). Particularly, we combine a set of independent ANNs into a single model. Each ANN uses different sets of inputs depending on the physical processes simulated. The model is then optimized as a black-box system using metaheuristics (Genetic and Memetic Algorithms). We demonstrate that the proposed ANN model presents a high correlation between the real output and the predicted one. Additionally, the performance of the proposed optimization techniques has been validated by the engineers of the plant, who reported a significant increase in the benefit that was obtained after optimization. Furthermore, this approach has been favorably compared with the results that were provided by a general black-box solver. All methods were tested over real data that were provided by the factory.

Conceptual Approach in Multi-Objective Optimization of Packed Bed Membrane Reactor for Ethylene Epoxidation Using Real-coded Non-Dominating Sorting Genetic Algorithm NSGA-II

An isothermal plug flow reactor model with extended Fick diffusion model for transport through the porous membrane is utilized for simulation of ethylene oxide formation in a packed bed membrane reactor (PBMR). The model was verified and validated using published experimental data from an existing lab-scale unit. Sensitivity analysis was performed to determine robustness of the model. A conceptual approach on operation and design stage multi-objective optimization study is discussed. Real-coded NSGA-II is used and effect of its parameters on optimization of reactor performance is also studied. The results of three two-objective operation-stage (with 4 decision variables) and one two-objective design-stage (with 6 decision variables) optimization case studies are presented. Good convergence to a Pareto optimal solution is achieved for all cases. Significant improvement over current experimental operation is observed in terms of increase in conversion of ethylene, selectivity to ethylene oxide and ethylene oxide product flow rate.

Kinetic Modeling and Reactor Simulation and Optimization of Industrially Important Polymerization Processes: a Perspective

The field of what is now referred to as polymer reaction engineering started in the early 1930s with Staudinger’s discovery of macromolecules. Though the earlier work was related primarily to synthesis and kinetics, the field started growing at increasing rates, possibly in the late 1950s or early 1960s. In the early years, this field provided a challenging area of research. It has evolved from the modeling of simple polymerizations to that of more complex systems, to experimentation for filling the gaps in our knowledge, to optimization. This mini-review summarizes a small sampling of the literature in polymerization reaction engineering over the last about four decades using a personal perspective. The concepts in this area are now being applied in a variety of specialized domains, e.g., polymerization at the nano scale, design and control of chain macrostructure and rapid optimal switch-over of grades being manufactured, molecular simulation and computational fluid dynamics, etc.

Improvement of an overloaded, multi-component, solvent gradient bioseparation through multiobjective optimization

Solvent gradient chromatography is quite often used in analytical studies for decreasing the analysis time of samples having components with widely different retention behaviour. Several studies, both theoretical and experimental, have been reported on the optimization of gradient profiles in improving analytical separation performance, suggesting various linear and non-linear gradients. In preparative chromatography, on the other hand, though solvent gradient is being increasingly used (especially in bioseparation) to improve the product yield and productivity, there is a dearth of literature and clearer understanding of the effect(s) of modifier gradients on the separation performance. For this, the gradients used in applications are of relatively simple profiles like step or linear gradients, obtained through hand optimization based on experience and intuition. Significant improvements, however, can be expected using the state-of-the art modelling of chromatographic processes and optimization routines running on widely available hi-speed desktop computers. In this work we are reporting such an optimization procedure to improve the purification of an industrial multi-component mixture, containing 65.8% of Calcitonin as the main product, in an overloaded reversed-phase column. The work comprises both theoretical simulations and their experimental validation using multilinear gradients as optimization variable. The study produced interesting insights for modifier gradient design, like using peak deformation of the target peptide to increase yield and productivity, and improved our understanding of the effect of modifier gradients in non-linear separations.