The quality of the glass melt is determined by thermal and compositional homogeneity across the feeding channel. Achieving that homogeneity with stirrers made from precious metals as well as from refractory is a widespread method used in the glass industry. Homogenisation of glass melts by stirring happens under laminar flow conditions due to the low Reynolds-numbers at typical stirring temperatures and viscosities. AS the glass viscosity conditions do not permit turbulence, good homogenisation is hard to achieve. Compositional and thermal differences such as cords cannot be fully dispersed, they can only be stretched and folded until they are virtually invisible. This poses the question "What stirrer geometry is fluid-dynamically efficient as well as considering a multitude of other requirements such as low precious metal weight, longevity and robustness?" In a project carried out by Umicore, the flow inside a stirring cell was studied by computational fluid dynamics and predominantly by physical simulation in order to develop an advanced stirring device. Cord was simulated by coloured tracer liquid at different locations in a glucose model and different stirring concepts were tested and compared. The exercise revealed valuable insight into how the glass flow should be affected and what geometry an efficient stirrer should have when it can be made from oxide dispersion strengthened PtRh-alloy. Finally, the study led to the development of the ARCtwister, a modular, self-supporting stirrer made from fine grain stabilized alloy. The stirrer creates a unique S-curved flow characteristic moving the glass volume vertically.
