UCD Model the High Shear Process Developing Optimised Design for Mixing Head
Dr. Mingming Tong, Dr. David Browne
22/06/15
School of Mechanical and Materials Engineering, University College Dublin, Ireland
The UCD modelling group has developed a model to simulate the high shear process (HSP) developed by Brunel to ‘clean’ contaminated molten aluminium scrap material. The HSP breaks up large oxide films/clusters in the melt into a fine distribution of particles which can dramatically enhance the nucleation and growth of harmful iron containing particles which then can be more easily removed from the melt. The modelling problem can be classified into 2 sub problems. One problem is how the particles are sheared and hence change in size, and the other problem is how the bulk melt is redistributed across the whole crucible at the macroscale.
UCD simulations found the reduction in size of the oxide clusters (Fig.1), in the close vicinity (points 1,2,3) of the HSP mixing head, is dramatically more than that in the far field (points 4,5).
Fig.1 Plot of oxide particle size with processing times at different locations in the HSP mixing crucible.
By modelling different designs of mixing head in the computer simulation we have found that the rotor speed has dominant influence on the size of oxide clusters. While the rotor-stator gap and the size and number of holes of the stator have relatively less influence (Fig.2).
Fig.2 Figure showing how changing the rotor speed, hole size and rotor-stator gap affects the mean particle size at the location of point 1 in Figure 1, where shearing was most effective
Looking at the overall crucible in 3D (Fig.3), the high shear mixer redistributes the molten alloy throughout the crucible.
Fig.3 Illustration of the parameters used in 3D CFD modelling.
The computational fluid dynamics (CFD) modelling that we carried out formulated the complicated flow pattern of the molten alloy in 3D, both in close vicinity and far field of the mixer (Fig.4). We found that there is a very localized recirculation of melt in the close vicinity of the mixing head through the large opening of the mixing head at its bottom.
Fig. 4 Flow pattern of the melt in the close vicinity and far field of the mixer.
This localized recirculation of the melt only agitates the melt inside very limited volume and leaves the rest of the melt relatively stagnant in the crucible. This means the only a small fraction of the melt is undergoing the HSP. UCD used the modelling results to optimise the design of the mixing head to evenly redistribute the melt throughout the crucible, enabling the whole melt to be processed. CFD simulations by UCD predict the improved design of the mixer head should improve performance of the HSP by very well agitating the meld throughout the whole crucible (Fig.5).
Fig. 5 Volume of the well agitated melt in the crucible based on the original design and optimised design of the mixer.
UCD have an ongoing collaboration with Brunel University as part of the RecycAl project and the teams are currently working on validating the predictions of the model and using the predictions to improve the processing of aluminium.