In batch pharmaceutical manufacturing, understanding spatial variations in temperature is critical for scaling up temperature-sensitive processes such as cooling crystallization and synthetic chemistry. The spatial distribution of temperature can vary significantly during scale-up and impact process performance. Additionally, batch operations inherently exhibit process fluctuations, where variations in operating conditions and operator control can influence the transient temperature profile and its uniformity. While temperature probes provide pointwise measurements, it remains unclear whether these readings truly reflect the overall thermal state of the vessel. This study seeks to address two key questions: (1) How representative are temperature probe readings in large-scale vessels under different mixing and temperature conditions? (2) What happens when temperature profiles are scaled up from lab to plant scale—can they simply be applied as-is, or do they require adjustments?
To address these challenges, we developed a CFD-based modeling framework to simulate free-surface dynamics and transient heat transfer in stirred vessels. The model employs a Volume-of-Fluid (VOF) approach for free-surface tracking enabled by adaptive mesh refinement (AMR) and a multiphysics simulation strategy that enables long-time transient simulations. The consistent modeling and simulation strategy is applied from 250 ml lab scale to 600 L plant scale. An industrial case study on cooling crystallization scale-up was investigated, focusing on jacket temperature (Tj) overshooting—a commonly observed phenomenon where the external cooling system rapidly drives the Tj down, creating localized supersaturation zones within the vessel. We will demonstrate how the model captures how Tj overshooting leads to localized cold spots in the bulk fluid, potentially triggering undesirable primary nucleation events and polymorph variations, and how temperature control strategy should be adapted to mitigate primary nucleation in the plant scale.
