Mixing behaviour of cohesive and non-cohesive particle mixtures in a ribbon mixer

Abstract

Abstract. Ribbon mixers are used in a wide range of applications involving pharmaceuticals, ceramics and cosmetics, to name a few. Here, the discrete element method is used to investigate the effect of impeller speed on the mixing behaviours of cohesive as well as non-cohesive particle mixtures in a ribbon mixer, which has a horizontal cylindrical vessel. The mixing behaviours are characterized by particle-scale and macroscopic mixing indexes. Simulations show that the mixing rate increases with the impeller speed for both the cohesive and non-cohesive mixtures up to a certain speed, beyond which it showed a reduction. There is a possibility that the mixture quality becomes poorer at higher impeller speeds for the non-cohesive particles, but it was not the case with the cohesive particles. Inspection of velocity fields shows that many local recirculation regions exist in the case of non-cohesive particle mixing, preventing the overall mixing. By contrast, in the case of the cohesive mixture, there exists a circumferential motion about the shaft and a convective motion in the horizontal axial direction, improving the particle mixing. Force analyses are also carried out, which show that the particle contact forces increase with the impeller speed for non-cohesive particles, but in the case of cohesive particles, they increase initially with the impeller speed, and then show a reduction after a certain speed. The results will be useful in selecting operation conditions of a ribbon mixer.

A soft-sensor approach to mixing rate determination in powder mixers

“…Since an excessive or insufficient dose of a pharmaceutical could have severe health consequences, thorough particle mixing is crucial [1]. This mixing may by conducted as a continuous process, such as in many food production applications, or as a batch process [2][3][4]. Batch mixers are commonly found in particle mixing for pharmaceutical goods, where they are more suited to the regulations placed upon the manufacture of such goods [5].…”

“…Fill level is of great interest in particle mixing due to the throughput maximisation problem that arises by balancing the short mixing time and small volumes of low fill levels with the long mixing time and large volume of high fill levels [6]. DEM Simulation studies by Halidan et al [2,3] focus on the effect of cohesiveness in particles when being mixed in ribbon mixers. Their work showed that cohesiveness, characterised by the Bond number, decreased the rate of mixing.…”

“…They showed that mixing was fastest for intermediate impeller speeds (100 RPM), where avalanching was dominant. At lower speeds the effect of avalanching was too small to achieve fast mixing, and at higher speeds the particles were forced into the wall where mixing is poor [3]. The particle scale mixing was compared to a macroscopic mixing index, and the disagreement in the data showed that the macroscopic mixing index being well mixed does not ensure that particles are well mixed [3].…”

“…At lower speeds the effect of avalanching was too small to achieve fast mixing, and at higher speeds the particles were forced into the wall where mixing is poor [3]. The particle scale mixing was compared to a macroscopic mixing index, and the disagreement in the data showed that the macroscopic mixing index being well mixed does not ensure that particles are well mixed [3]. Such studies are motivated by the need to maximise throughput in ribbon mixers and generate sufficiently well mixed products.…”

“…This notion of mixing has been expressed as a first order differential equation with respect to time, following an exponential curve asymptoting towards 100% [12]. Several studies show curves of mixing with similar features to first order systems [2,3]. In practice, however, a first order model is often insufficient to capture the initialisation behaviour.…”