2024 AIChE Annual Meeting
Studies in Powder Segregation — a Computational & Experimental Approach
Segregation is the physical phenomenon in which a multicomponent powder mixture separates into its components due to differences in size, shape and density. Preventing segregation is of special interest to the Procter & Gamble Company in applications where a powder with a specific composition is required. To understand how segregation occurs in an industrial hopper on the production lines of certain P&G products involving powders, an existing numerical simulation process, based on solving an advection-diffusion-segregation PDE using gPROMS software, has been used with literature-validated results. However, due to necessary manual calculation steps, work was required to reduce manual errors and enhance useability. Additionally, a potential new powder binary mixture for a P&G product was identified, requiring an experimental assessment of segregation tendency of these materials on the production line. On the computational side, work was carried out regarding the segregation simulations to make it more usable, and develop capabilities for automation of gPROMS simulations in Python and numerical simulation implementations in Python and MATLAB. Subsequently, a Python-based tool which can interface with any gPROMS-based simulation was developed then applied to the powder segregation simulations, plus additional code eliminating manual calculations and implementing automatic postprocessing of simulation results. On the experimental side, the segregation tendency of the new proposed binary mixture was assessed in pilot scale experiments: a specialized “modified heap” segregation testing apparatus was used to assess sifting segregation, a pilot-scale loss in weight feeder was used to simulate real production line conditions, and a twin-shaft paddle mixer was used to assess mixing difficulty and segregation in handling. Results suggest a weak tendency for this binary mixture to segregate in real manufacturing conditions. However, achieving a well-mixed sample of this specific mixture is very difficult in the first place, with an abnormally long time required in the mixer to reach a well-mixed steady state, and segregation readily occurs in post-mixing handling. These experiments also served as method development for future experiments with other mixtures. Future research on the computational side will largely focus on other applications of numerical PDE simulations, including alternate uses of the Python-gPROMS interfacing code as well as applications of simpler systems in Python or MATLAB. On the experimental side, rheology work will focus on determining the impact of certain amounts of segregation of this proposed mixture on the consumer appeal of this P&G product.