Filtration is a broadly utilized procedure in chemical, mining, and manufacturing industries, with special representation in Minnesota’s industrial sector. Instead of a universal filtration media that performs well in all applications, a tailored approach is needed to optimize the structure of the fiber media to a targeted application. Computational simulations of the fluid filtration process are necessary for guiding optimization, but the extraordinary structural complexity of the filter media poses unsurmountable challenges to the established computational methods of atomistic (molecular dynamics) and continuum (finite element) scales.

Professor Traian Dumitrica (Mechanical Engineering; MSI PI) is working on a project called “Digital twin informatics for de novo design of fibrous filtration media,” that will develop a “coarse grained” distinct element method (DEM) representation of the electrospun polyamide nonwoven fiber media, such as those currently manufactured at Donaldson. The DEM network will be a visual digital replica of the corresponding experimental network observed by SEM, and empowered with the mechanics of the real material. Once established, a CFD-DEM capability will utilize the created DEM Digital Twin to simulate the particle-filter collision processes, the mechanics of capture, and the potential fiber rupture. For validation purposes, results will be compared with the available experiments. This will lay the grounds for a DEM-based Digital Twin de novo design of fibrous filtration media. The Digital Twin represents a large dataset of DEM mechanical models, each corresponding to an SEM-imaged network portion made from a specific synthetic fiber. CFD-DEM simulations will be performed ahead of experimentation to determine the role of the mesoscale structure onto the desired filter characteristics, and to inform manufacturing with the optimal structure for the desired application.

This project recently received a Research Computing Seed Grant. RC Seed Grant funds are intended to promote, catalyze, accelerate and advance U of M-based informatics research in areas related to the MnDRIVE initiative, so that U of M faculty and staff are well prepared to compete for longer term external funding opportunities. 

This Seed Grant falls under the Robotics research area of the MnDRIVE initiative.

schematic of filter screening viruses and dust out of air