(119a) Analysis of Two Multi-Phase Stirred Vessels Utilizing Computational Fluid Dynamics

A study was completed to determine potential root causes for performance differences between a pilot-plant stirred vessel and a commercial stirred vessel. Time-resolved computational fluid dynamics (CFD) simulations were utilized to determine the flow patterns and residence time distributions (RTDs) of fluid and different particle sizes in these two stirred vessels utilizing M-Star CFD software. These models predicted time-resolved turbulent fluid flow inside the vessel as well as particle dynamics within the flow. To determine RTDs, tracers and 3 different particle sizes (representing the P10, P50, and P90 particle sizes) were introduced into the inlet. Pilot-plant CFD results with constant inlet and outlet flows showed that the fluid RTD closely aligned with its equivalent continuously stirred tank reactor (CSTR) RTD, but the particle RTDs varied significantly from the fluid RTD. The smallest particle size had a mean residence time ~3x the fluid mean residence time while the mean residence time of the largest particle size could not be determined with any degree of certainty because so few particles exited over the computational run time. Actual pilot-plant operation is known to have a periodic pulsed outlet flow. An approximation of this pulse was applied as a transient boundary condition. When compared with the constant outlet flow simulation, the pulsed outlet flow simulation showed significant differences in the RTDs for all particle sizes, but only minor differences in the fluid RTD. The particle RTDs still varied significantly with the fluid, with the largest particle size having a mean residence time ~55x the fluid’s mean residence time. In contrast to the pilot-plant simulations, the commercial vessel simulation showed that the fluid RTD closely matched the CSTR RTD and the particle RTDs showed much less variation and closer agreement with the CSTR RTD, with the mean residence time of the largest particle size ~2x the fluid mean residence time. These results indicate that the performance differences between the pilot plant and commercial scales are likely significantly impacted by the particle RTDs. These results also indicate the need for accurate time-resolved operational data for the pilot plant, especially the outlet flow pulse, to be able to accurately determine the RTDs of the different particle sizes utilizing a CFD model.