Effect of Operating Parameters on Powder Mixing in Horizontal Stirred Bed Reactors using Discrete Particle Method

  • Pourandi, Sahar (University of Twente)
  • Weinhart, Thomas (University of Twente)
  • Ostanin, Igor (University of Twente)
  • Luding, Stefan (University of Twente)
  • Thornton, Anthony (University of Twente)

Please login to view abstract download link

Gas-phase polymerization has become more efficient due to the use of horizontal stirred bed reactors (HSBR), which provide effective axial powder mixing [1]. These reactors typically comprise a bed of particles agitated by a series of paddles connected to a central shaft. In this study, we aim to gain insight into the mixing of polypropylene powder in a lab-scale HSBR. To achieve this, we used a scaled-down model of an industrial HSBR and simulated it with a discrete particle model using the open-source software MercuryDPM [2]. Spherical particles were assumed, and the Hertz-Mindlin contact law was used for particle interactions. We investigated the influence of impeller speed and reactor filling on the mixing of polypropylene powder in the angular, radial, and axial directions. The effectiveness of the angular mixing was measured through the cycle time- the time of a full angular revolution of a particle around the central shaft. Our results indicated that increasing the fill level and impeller speed results in a shorter cycle time, which is desirable for faster mixing. Furthermore, we measured Lacey mixing index to investigate mixing quality in the radial direction. This is because a higher Lacey mixing index shows better mixing. Our results showed that increasing fill level increases the Lacey mixing index, but changing the speed has no effect. Last, we studied axial mixing by measuring the axial dispersion coefficient. This is because high axial dispersion leads to better mixing. Our results showed that the axial dispersion coefficient increases linearly with increasing impeller speed. However, there was no clear relation between reactor filling and the axial dispersion coefficient. Therefore, to optimize the HSBR, we should increase the fill level (in the range studied) and rotation speed. Finally, by applying an advection-diffusion model, we used the measured dispersion coefficient to predict the mixing quality in the HSBR mixer for two components with equal particle properties. We validated our prediction by measuring the Lacey mixing index directly from the DPM simulations using particle tracking, which demonstrated good agreement. This indicates that the measured dispersion coefficient can be effectively used to predict the mixing quality in an HSBR mixer when no segregation is present. Our findings provide valuable insights into the mixing of polypropylene powder in HSBRs and can help improve their efficiency in gas-phase polymerization.