Simulation of the methanol synthesis with artificial neural networks

The conversion of sustainable COx (e.g. from biomass or captured CO2) and green H2 to methanol is a key reaction to integrate renewable energy with the chemical industry and the mobility sector, as methanol is a base chemical for the production of several valuable fuels (e.g. gasoline, diesel, jet fuel) and chemicals (e.g. formaldehyde, dimethyl ether) [1]. In order to better understand the methanol synthesis and allow possibilities for optimization, we previously developed a microkinetic model with extensive experimental validation, allowing detailed simulation of the surface reactions [2]. While this type of model is useful to gain insights into the chemistry behind the process, its complexity and high computational effort requirements hinders its use in practical applications, such as reactor optimization, scale-up, and techno-economic analysis. An interesting alternative, explored in this work, is the development of a machine learning technique to extract the information of the theoretical model and handle it to the appropriate reactor model, drastically reducing the computational effort [3]. In the present work, 113,000 data points of reaction rate were generated with the theoretical model, covering a wide range of operating conditions. ... mehrThe data was randomly divided into three groups (train, validation, and test), and used to train an artificial neural network (ANN), which adequately reproduced the theoretical data. A highlight of this method is that the ANN only learned to simulate the kinetics of the methanol synthesis, and known effects, such as thermodynamic and balance equations, being separately added in the reactor model, with experiments being adequately reproduced by the simulations. The ANN is highly flexible with this approach, allowing the modeling of novel reactor designs, detailed CFD simulations, the inclusion of long-term catalyst deactivation models, and the use in the aforementioned practical applications.

[1] Bozzano and Manetti (2016). Prog. Energy Combust. 56: 71-105.
[2] Lacerda de Oliveira Campos et al. (2021). Reac. Chem. Eng. 6: 868-887.
[3] Park et al. (2020). Catalysts 10, 655.