[MAD-SIM] Modelling powder flows in Agitated Devices: from rheological studies to predictive SIMulation at process scale

This PhD thesis aims to model and simulate powder flows in agitated industrial equipment, a critical challenge given that powder behaves variably as a solid, a fluid or a gas. The design and operation parameters choice of these devices are often determined empirically from experiments or relying on default rheological models. Rationalizing these choices is paramount for ensuring product quality and process safety. To that purpose, existing modelling techniques face significant limitations. While dimensional analysis requires established invariants and precise knowledge of the phenomena, numerical methods like Discrete Element Modelling (DEM), though providing accurate access to particle trajectories and velocities, demand substantial computing time, making them unrealistic for large scale equipment using real powders [1]. As more macroscopic approaches, continuum mechanics models, such as CFD-type, allow for larger-scale simulations but are dependent on knowing the constitutive laws specific to the powder rheology [2][3]. Accurate modelling of the phenomena involved in powder rheology is therefore essential for the rational design of equipment.

Recent work [4] observed dense-phase flow in horizontal mixers that can be described by the µ(I) rheology [5][6][7]. This relationship reveals the transition between frictional and pre-collisional regimes and can be used as invariant basis for scaling up processes [8]. However, these findings need consolidation and extension from model granular media to real powders [9]. The thesis aims to explore rheological approaches and continuum mechanics models to simulate these flows. An instrumented pilot mixer will be used as a systemic rheometer. Both physical and digital experiments will be conducted to identify the geometry of the shear zone and the mechanisms causing its formation, linked to particle properties and agitator geometry. DEM will be specifically utilized to characterize phenomena near the agitated zone, providing details difficult to measure in a laboratory setting. The identified effective viscosity will mimic microscopic fluid behaviour and contribute to the continuous macroscopic model that will be implemented in CFD software, allowing for large-scale simulations. Initial validation will occur on the pilot mixer, with the long-term goal of adapting the model to volumes of several hundred or thousand litres. A key deliverable is establishing the principles for a prototype rheometer capable of generating data for dense flows in a semi-confined environment, addressing a current market gap. This ambitious project seeks to apply advances in granular physics to real industrial situations, ultimately aiming for an integrated approach to powder characterization and industrial-scale flow simulation. This will facilitate the development of digital twins to optimize the eco-design of industrial equipment and determine operating parameters.