Advanced process control for salvianolic acid A conversion reaction based on data-driven and mechanism-driven model

Abstract

On-line monitoring and process control are the most efficient techniques to ensure high quality, low cost, and effective resource utilization in modern process engineering. The high-temperature and high-pressure operating conditions make the process control challenging during conversion reaction. In this study, the contents of main substances produced during chemical conversion of salvianolic acid A were monitored and controlled by an advanced process control method. At first, a kinetic model was established according to the reaction mechanism. Then, a high-precision partial least squares calibration model was established based on near-infrared spectroscopy. Finally, the conversion reaction was controlled in real-time by combining the data-driven model with the mechanism model. The correlation coefficients obtained by the data-driven model were all > 0.9. The fitting correlation coefficients of reaction rate constants were all > 0.9. The concentration of the reaction product was controlled within the limit by adjusting the reaction temperature based on the concentration-temperature curve. Combining the data-driven model with the mechanism model, the reaction process was controlled in real-time. This study emphasized the advantages of advanced process control for high temperature and high-pressure reaction processes and exemplified it as an important step in implementing intelligent production.