Model driven adaptive design with concentration profiles

Abstract

Effective kinetic models of heterogeneous catalytic processes are an indispensable tool for reactor design, optimization, and control. Under the assumption of using functional forms like power laws, model parameters are traditionally fitted to kinetic data measured along local line scans. A local line scan involves systematically varying one individual reaction parameter, such as a reactant concentration or temperature, at a time. This approach typically involves numerous separate kinetic measurements and is susceptible to the uncertainty of these line scans in determining the model’s parameters. Here, we explore the use of profile reactors in combination with a fully automated adaptive design approach for an efficient identification of effective kinetic models. Originally developed to provide operando information along the axis of tubular reactors, profile reactors provide a complex line scan that encapsulates kinetic information across all reaction conditions probed along the tube. The proposed Model-Driven Adaptive Design with Profiles algorithm harnesses this extensive dataset to strategically guide the selection of initial reaction conditions for subsequent profile reactor measurements. This approach ensures that each line scan provides maximally complementary information, thereby significantly enhancing the efficiency and accuracy of kinetic model identification.