Chair of Fluid Dynamics

To further develop physics-based forest fire models, generic pinewood model trees are being studied both experimentally and numerically under …

To further develop physics-based forest fire models, generic pinewood model trees are being studied both experimentally and numerically under controlled conditions. The precise design of the model trees, together with a standardised ignition procedure under dry conditions, allows for reproducible experiments. The numerical simulation is based on the coupling of thermally thick mesh-based solid particle models with a steady flamelet approach for the gas phase.
When using Lagrangian particles or the discrete element method to simulate solid biomass particles, an important factor is how these particles are coupled with the Eulerian gas phase. Using the classical particle-centroid method, if particles occupy more than 3% of a gas cell's volume, significant numerical errors and grid dependencies arise. Modern grid-independent methods enable the representation of particles between fully resolved (d_p>10?) and unresolved (d_pThe biomass present in the forest, such as branches, twigs and blades of grass, is represented by cylinders. These cylinders can vary greatly in diameter and length, reflecting the many different physical scales of forest fires. Grid-independent interaction is therefore extremely important. The Gaussian kernel was adapted to suit the cylinders and applied to the generic miniature tree. Figure 1 (left) shows the charcoal remaining after the pyrolysis of the generic miniature tree's biomass. The figure on the right shows the mass loss rate for cylinders of different lengths (slicing), using a constant Eulerian grid. The different cylinder lengths demonstrate a consistent trend. This ensures grid independence.

When this is applied to a LES simulation of a wildfire, the influence of multi-scale particle sizes is significantly reduced. This brings us closer to achieving an accurate wildfire simulation.

In a study recently accepted for the International Symposium on Combustion, Parsa Ghofrani examines the ignition dynamics of iron particle …

In a study recently accepted for the International Symposium on Combustion, Parsa Ghofrani examines the ignition dynamics of iron particle clouds. High-fidelity simulations reveal that the “rust,” or oxide layer, forming on iron particles acts as a critical barrier to efficient combustion and reliable ignition. As particles are heated within a turbulent flow, this oxide layer continues to grow; when heating is too slow, the layer thickens and significantly impedes ignition.

As a result, the required ignition temperature can rise above the surrounding gas temperature, effectively “smothering” the particles and preventing flame formation. One possible approach to overcome this is preheating. This enables the particles to ignite before the oxide layer becomes prohibitive, allowing a self-sustaining flame to develop. Understanding these microscopic mechanisms is key to designing next-generation, carbon-free industrial burners based on metal fuels.

M.Sc. Maximilian Schäfer has developed a new method to enforce elemental and mass conservation in neural networks predicting chemical source …

M.Sc. Maximilian Schäfer has developed a new method to enforce elemental and mass conservation in neural networks predicting chemical source terms. The approach applies a weighted projection as a final correction layer, ensuring strict physical consistency without modifying or retraining the neural network. Compared to current state-of-the-art conservation methods, the new technique reduces computational overhead by roughly an order of magnitude while maintaining high predictive accuracy. The work has been accepted for presentation at the 41st International Symposium on Combustion, which will take place this summer in Japan.

Simulations by J.S. Tampah Fossi and A. Karimi Noghabi have been presented at the DECHEMA conference on February 26th. The work, a collaboration …

Simulations by J.S. Tampah Fossi and A. Karimi Noghabi have been presented at the DECHEMA conference on February 26th. The work, a collaboration with Prof. Segets’ group (also EMPI), shows the effectiveness of a new rotating micro mixer that enables the transition from batch to continuous processing for the formation of nano-particles in the liquid phase.

A detonation cellular stability mechanism based on the dynamics of reactive decaying blasts is examined through detailed analyses of two-dimensional …

A detonation cellular stability mechanism based on the dynamics of reactive decaying blasts is examined through detailed analyses of two-dimensional (2D) numerical simulations of hydrogen-oxygen detonations. Different from previous blast-based examinations, we resolve the transient process of decoupling between shock and reaction fronts in decaying blasts, and correlate the size of unburnt gas mixtures behind decaying shocks to that of the subsequent blast kernels.
The impact on the stability mechanism of (1) chemical kinetics, (2) diffusive processes, and (3) boundary conditions are examined through a series of simulations. At a dopant level, ozone is known to reduce ignition delay without altering thermodynamic properties of the mixture, enabling investigation of the impact of ignition kinetics on the cellular stability. The addition of ozone leads to a stronger coupling between shock and reaction fronts and stabilizes the blast kernel to a smaller size. The resulting global cell size reduction in the ozonated detonation is well described by the stability analysis and in agreement with experimental cell measurements reported in Crane et al., Combust. Flame 200 (2019) 44–52.

In reactive CFD simulations, evaluating chemical source terms with traditional ODE solvers is computationally expensive, especially for detailed …

In reactive CFD simulations, evaluating chemical source terms with traditional ODE solvers is computationally expensive, especially for detailed mechanisms. Neural networks offer a faster surrogate, but their raw outputs can violate mass and element conservation. Our work introduces a lightweight, model-agnostic post-processing layer that applies a stoichiometry-based projection into the valid solution space, minimally adjusting each species by its relative importance. The proposed method corrects conservation violations without altering the network architecture or training process, delivering physically valid outputs even in mechanisms involving both major and minor species and is applicable to any system requiring mass or element conservation.

This study examines the over-pressures achieved from shock focusing and the resulting ignition in hydrogen-air
mixtures within three-wall …

This study examines the over-pressures achieved from shock focusing and the resulting ignition in hydrogen-air
mixtures within three-wall corner geometries. Numerical simulations of compressible flows are validated against
experiments performed by the group at Warsaw University of Technology, in which a planar shock wave in a
stoichiometric hydrogen-air mixture propagates through a circular tube and is focused in the corner. Pressures
are measured using a small pressure sensor embedded at the tip of the corner. The simulation has been set up to
include the sensor geometry and to permit a comparison to the measured pressure as an average over the sensor
surface.

The impact of flame propagation speed in a hydrogen-fueled internal combustion engine with port-fuel injection is investigated by comparing …

The impact of flame propagation speed in a hydrogen-fueled internal combustion engine with port-fuel injection is investigated by comparing two species transport models: the Mixture-Averaged model and the Unity Lewis Number model. This study aims to evaluate the role of preferential diffusion—an effect characteristic of hydrogen combustion—on key combustion characteristics. Results indicate that simplified transport models fail to capture the acceleration and deceleration dynamics of flame front development, highlighting the importance of accurate species transport modeling.

At high Schmidt or Prandtl numbers, scalar length scales are much smaller than velocity scales, requiring fine grids and careful management …

At high Schmidt or Prandtl numbers, scalar length scales are much smaller than velocity scales, requiring fine grids and careful management of numerical diffusion. A previously developed Euler-Lagrangian decomposition (ELD) method addresses this by splitting the scalar field into a low-wavenumber component handled in an Eulerian framework and a high-wavenumber component reconstructed with Lagrangian particles. However, particle-based reconstruction can introduce noise, increasing computational cost or obscuring smooth scalar structures.
This paper introduces an improved approach that uses anisotropic diffusion as an edge-detection method to target steep gradient regions for particle addition. A novel particle removal strategy is also proposed to reduce computational cost and noise in scalar smooth regions, and it has been tested here for the first time.

Bei hohen Schmidt- oder Prandtl-Zahlen sind die skalaren Längenskalen deutlich kleiner als die Geschwindigkeitsskalen, was feine Gitter und eine sorgfältige Kontrolle numerischer Diffusion erfordert. Die zuvor entwickelte Methode Euler-Lagrangian-decomposition (ELD) begegnet diesem Problem, indem das Skalarfeld aufgeteilt wird in eine niederfrequente Komponente, die in einem Euler'schen Koordinatensystem behandelt wird, und eine hochfrequente Komponente, die mit Lagrange'schen Partikeln rekonstruiert wird. Die partikelbasierte Rekonstruktion kann jedoch Rauschen verursachen, was entweder die Rechenkosten erhöht oder glatte skalare Strukturen überdeckt.
Diese Arbeit stellt einen verbesserten Ansatz vor, bei dem anisotrope Diffusion zur Erkennung von Kanten verwendet wird, um in Bereichen mit steilen Gradienten gezielt Partikel hinzuzufügen. Außerdem wird eine neuartige Strategie zur Entfernung von Partikeln vorgeschlagen, um die Rechenkosten und das Rauschen in Bereichen glatter Skalarfelder zu reduzieren, die hier erstmals getestet wurde.

Paper link: https://doi.org/10.1016/j.compfluid.2024.106490

The important role of confinement on detonations is examined by detailed numerical simulations in 1D, 2D and 3D, for circular and rectangular …

The important role of confinement on detonations is examined by detailed numerical simulations in 1D, 2D and 3D, for circular and rectangular channel.

Paper link: https://doi.org/10.1016/j.proci.2024.105388