Unveiling the relation between multiple chemical products and process conditions for trichloroethylene and perchloroethylene production via catalysis network analysis

Abstract

Trichloroethylene (TRI) and perchloroethylene (PER) are widely-produced in the chemical industry and used as solvents, varnishes, degreasers, and dry cleaning chemicals that involve complex process conditions. Data science and network analysis are used in order to unveil relationships between reactants, process conditions, and selectivities of select products with the aim to improve production efficiency. Data visualization and machine learning reveal the sets of conditions that have positive and inverse relations with TRI and PER selectivities, while transforming the data into networks reveals which sets of experimental conditions correlate with desired outcomes. Thus, it becomes possible to tailor experimental conditions in order to increase desired selectivities while avoiding production of undesirable selectivities.