Modeling and Analysis of Long-Term Particle Deposition on a Cylinder
Licentiate thesis, 2025
Long-term particle deposition studies are challenging to conduct due to both their experimental and computational difficulties. In terms of computational challenges, the main difficulty is the extreme computational cost of accurate particle deposition simulations, where simulating the carrier fluid represents a large fraction of the total computational cost. This project investigates the potential of using the recurrence computational fluid dynamics (rCFD) method for time-extrapolating the carrier flow fields, eliminating a large fraction of the computational cost of particle deposition simulations.
Firstly, we investigate the feasibility and computational cost of using rCFD for performing a particle deposition study on a benchmark system of flow around a cylinder. In this study we show that deposition rates can be accurately obtained with rCFD at a fraction of the computational cost of conventional computational fluid dynamics (CFD). Special effort is focused on the cylinder back-side deposition rates, a benchmark case that is particularly challenging due to the turbulent wake interactions.
Secondly, we investigate the time-dependence of particle deposition rates on the back of the cylinder using direct numerical simulation (DNS) simulations. The results of this study indicate that particle deposition rates are highly time-dependent, with observed short-term impact rate fluctuations of up to a factor 27 for flow at Re = 6600. To the best of our knowledge, this effect has not been observed before. This study emphasizes the importance of choosing an appropriate rCFD database, while at the same time highlighting the challenges in constructing such a database.
The aim of this project is to reduce the computational cost of performing particle deposition studies. Such a reduction in cost would be useful in academia and industry alike. Examples of applications include sensor soiling in the car industry, icing on aircraft and ash build-up in boilers.
Firstly, we investigate the feasibility and computational cost of using rCFD for performing a particle deposition study on a benchmark system of flow around a cylinder. In this study we show that deposition rates can be accurately obtained with rCFD at a fraction of the computational cost of conventional computational fluid dynamics (CFD). Special effort is focused on the cylinder back-side deposition rates, a benchmark case that is particularly challenging due to the turbulent wake interactions.
Secondly, we investigate the time-dependence of particle deposition rates on the back of the cylinder using direct numerical simulation (DNS) simulations. The results of this study indicate that particle deposition rates are highly time-dependent, with observed short-term impact rate fluctuations of up to a factor 27 for flow at Re = 6600. To the best of our knowledge, this effect has not been observed before. This study emphasizes the importance of choosing an appropriate rCFD database, while at the same time highlighting the challenges in constructing such a database.
The aim of this project is to reduce the computational cost of performing particle deposition studies. Such a reduction in cost would be useful in academia and industry alike. Examples of applications include sensor soiling in the car industry, icing on aircraft and ash build-up in boilers.
particle deposition
recurrence CFD
data-assisted simulation
Author
Johannes Hansson
Chalmers, Mechanics and Maritime Sciences (M2), Fluid Dynamics
J. Hansson, T. Lichtenegger, S. Pirker, S. Sasic, H. Ström. Recurrence CFD for efficient predictions of long-term particle deposition on a cylinder
J. Hansson, S. Sasic, H. Ström. Low-frequency wake modulation governs particle back-side deposition on cylinders
Virtual real-time prediction of sensor soiling
VINNOVA (2021-05061), 2022-04-01 -- 2025-12-31.
Subject Categories (SSIF 2025)
Fluid Mechanics
Infrastructure
Chalmers e-Commons (incl. C3SE, 2020-)
Publisher
Chalmers