Process analytical technology (PAT) has developed significantly since its introduction in pharma where many in situ analytical probes and measuring devices are now commercially available, replacing the use of off-line quality control measurements that are typically laborious and time intensive. The use of PAT instrumentation should not interfere with the process itself and subsequently should have no effect on the product whilst measuring representative samples. Implementation of these devices is typically arbitrary using empirical means. Therefore, the objective of this study is to highlight the use of computational fluid dynamics modeling to investigate the effect of interfacing parameters and process parameters of an inline near-infrared (NIR) probe used to determine the viscosity of a non-Newtonian micellar liquid. The parameters investigated for the probe were immersion depth, immersion angle, gap size, and fluid velocity. The results conclude that the immersion angle and depth should both be optimized to prevent stagnant fluid accumulating in the measuring gap ensuring that the NIR measurements are representative of the bulk. The gap size determines the optical pathlength and therefore was also investigated against an existing predictive viscosity model showing no changes in model performance with varying gap size. The use of computational modeling to develop a digital twin prior to PAT implementation at the equipment design stage ensures the technology can perform at its best and will also aid in calibration transfer studies.
Keywords: CFD; NIR; Near-infrared; computational fluid dynamics; optimization; process probe.