Kinetics of the Direct DME Synthesis: State of the Art and Comprehensive Comparison of Semi-Mechanistic, Data-Based and Hybrid Modeling Approaches

Abstract

Hybrid kinetic models represent a promising alternative to describe and evaluate the effect of multiple variables in the performance of complex chemical processes, since they combine system knowledge and extrapolability of the (semi-)mechanistic models in a wide range of reaction conditions with the adaptability and fast convergence of data-based approaches (e.g., artificial neural networks—ANNs). For the first time, a hybrid kinetic model for the direct DME synthesis was developed consisting of a reactor model, i.e., balance equations, and an ANN for the reaction kinetics. The accuracy, computational time, interpolation and extrapolation ability of the new hybrid model were compared to those of aumped and a data-based model with the same validity range, using both simulations and experiments. The convergence of parameter estimation and simulations with the hybrid model is much faster than with theumped model, and the predictions show a greater degree of accuracy within the models’ validity range. A satisfactory dimension and range extrapolation was reached when the extrapolated variable was included in the knowledge module of the model. This feature is particularly dependent on the network architecture and phenomena covered by the underlying model, andess on the experimental conditions evaluated during model development.

Kinetic Modeling of the Direct Dimethyl Ether (DME) Synthesis over Hybrid Multi-Site Catalysts

“…In direct DME synthesis from CO 2 and H 2 , three reactions (Equations ( 6)-( 8)) are involved [10,[29][30][31]:…”

A neural network with embedded stoichiometry and thermodynamics for learning kinetics from reactor data

“…Currently, machine learning modelling of chemical reactors focusses on determining input-output relationships for specific reactor setups [69]. While these models are very powerful for describing a single system, their application is limited to the system parameters that the model is designed for.…”

Direct Coupling of Microkinetic and Reactor Models Using Neural Networks

“…For these reasons, ANN have been considered as a potential tool for fitting of kinetic data. Initial work has primarily focussed on the application of ANN for interpolation of reaction rate data derived either from experiment or kinetic models [25,[33][34][35][36][37][38][39][40][41][42][43]. These ANN are applied typically in process optimisation studies and therefore focus on predicting target design variables (such as conversion or selectivity) from process conditions (temperature, pressure, feed ratio).…”

“…Conversion is typically considered as the predicted quantity to reduce the dimensionality of the problem when more complicated multi-component systems are considered. Only a select few studies focussed on the direct replication of reaction rates by considering simple reaction systems [41,44] or assuming reduced kinetics [37][38][39][40]43]. To achieve full representation of the reaction rates for a kinetic network in terms of the reactant or surface composition, other machine learning methods have seen significantly more attention [20,26,27,34,42].…”