Simplest mechanism builder algorithm (simba): an automated microkinetic model discovery tool.

Area of Science:

• Chemical Engineering
• Computational Chemistry

Background:

• Microkinetic models are essential for assessing industrial process efficiency and environmental impact.
• Manual construction of these models is labor-intensive and time-consuming, necessitating automated solutions.

Purpose of the Study:

• To introduce SiMBA (Simplest Mechanism Builder Algorithm), a novel automated approach for generating microkinetic models from kinetic data.
• To demonstrate SiMBA's capability in distilling complex kinetic behaviors into simple, accurate models.

Main Methods:

• SiMBA employs a four-phase process: mechanism generation, translation, parameter estimation, and model comparison.
• It utilizes matrix representations and a parallelized backtracking algorithm for systematic mechanism proposal and complexity management.
• Ordinary differential equations represent the microkinetic models, optimized against available data using information criteria for model selection.

Main Results:

• SiMBA successfully generated accurate microkinetic models for aldol condensation and fructose dehydration reactions.
• The algorithm correctly predicted reaction intermediates in all case studies.
• SiMBA streamlines mechanistic exploration, providing a robust starting point for chemical process modeling.

Conclusions:

• SiMBA significantly accelerates the development and modeling of chemical processes by automating microkinetic model generation.
• While effective, SiMBA requires expert input for the chemical identification of intermediates in complex systems.
• This data-first approach opens new research avenues in automated mechanism discovery for chemical engineering.