The SARS-CoV-2 pandemic highlighted the need to understand aerosol transport and associated disease transmission, and motivated many numerical flow studies using different numerical approaches to predict Lagrangian particle transport for infection risk modelling, with varying degrees of accuracy and computational cost. To evaluate the trade-off between these different flow simulation approaches, we compare particle concentration predictions based on solutions of the steady and unsteady Reynolds-averaged Navier-Stokes (RANS) equations with experimental data. A ventilated generic train entry segment is chosen because it is easy to set up for experiments and numerical flow simulations. Two heated dummies are placed in this ventilated space, one of which continuously exhales aerosol. The RANS approach predicts significant particle accumulations that are not observed in either the experiments or the URANS simulations. However, the averaged absolute deviation from the experimental data is reduced by a factor of 2.4 when URANS simulations are performed, albeit at an eightfold increase in computational cost.

Humid airflows with surface condensation, where latent heat and condensate accumulation interact with the near-wall flow, play an important role in vehicle and building ventilation. In this work, we investigate the combination of directly solving the single-phase flow equations while using a wall model for the thermal diffusivity to include surface condensate effects without the need to resolve droplet surfaces, and compare this approach to a super-droplet based simulation.

In pipe flows at high Reynolds numbers, more than 90% of the pumping power is dissipated by near-wall turbulence, making relaminarization a promising strategy for saving energy. Streamwise traveling waves of wall blowing and suction, previously proven able to relaminarize turbulent pipe flow at Re τ = u τ R/ν = 110, are here demonstrated to achieve relaminarization up to Re τ = 540. Here, u τ is the friction velocity, R the pipe radius, and ν the kinematic viscosity. Independent of the Reynolds number, the relaminarization requires 130 R/u τ for the drag reduction rate to converge. The resulting flow structure combines viscous-scaled half-vortices superimposed on the radius-scaled laminar profile. These results establish scaling relations that extend traveling wave–induced relaminarization to higher Reynolds numbers, offering a predictive framework for energy-efficient flow control.

Understanding the (particle) transport processes through the planar turbulent air curtain jet – a ventilation concept designed to reduce the spread of airborne particles and thus possible airborne infections – is important prior to its implementation in a future passenger cabin. In the present study, two-dimensional Particle Image Velocimetry is used to investigate the deflection of an exhalation jet by an air curtain as well as the deflection of the air curtain itself as an indicator of a possible breakthrough of the exhalation jet. The strength of the two opposing jets is varied by changing the air curtain momentum flux to achieve four different momentum flux ratios γ between 0.17 ≤γ≤ 6.88. Strong deflections of the exhalation jet were observed for γ = 3.47 and γ = 6.88, while less deflection and a (temporal) breakthrough of the air curtain were obtained for γ = 0.65 and γ = 0.17.

Computational Fluid Dynamics (CFD) methods are becoming increasingly important in the modern automotive design process, where the main objective is to predict the effect of geometry changes on drag and lift. To ensure the accuracy of the results, the methods and setups must be validated, using reliable experimental data from different geometries under identical boundary conditions. The DrivAer model allows the effects of these different geometric modifications to be analyzed with reduced complexity compared to a production car. Two geometry variants, the notchback and the estate back, were investigated in a full-scale wind tunnel including ground simulation. As the flow field of the notchback model has proven to be particularly difficult to reproduce in CFD, the experimental data presented focuses on this area. Additional investigations with the estate rear end configuration show the presence of a pronounced upwash from the underbody, resulting in an upward distortion of the wake of the vehicle. A strong dependence of the rear lift and drag on the diffuser angle is observed. For the present study, validation data were collected using different rakes, equipped with pitot-probes mounted directly on the model and on the traverse system of the wind tunnel.

In technical applications, pumping fluids through pipes often generates turbulent flows with high Reynolds numbers, where over 90 Reτ =110 ), reducing friction losses and energy consumption. This work extends the investigation to higher Reynolds numbers, demonstrating that traveling waves can trigger relaminarization up to Reτ =720 . A parametric study is conducted at Reτ =180 and Reτ =360 , examining upstream traveling waves (UTWs, c<0 ) and downstream traveling waves (DTWs, c>0 ) while varying amplitude a, celerity c, and wavelength λ . Consistent with channel flow studies, UTWs destabilize the flow yet can generate sublaminar drag; only low-speed UTWs with large amplitudes effectively reduce energy consumption. For DTWs, a wide range of parameters reduces drag, but significant net energy savings occur only for 0.067U_c,lam≲ a ≲ 0.1U_c,lam , c≈ U_c,lam , and λ≈ 360δ _ν , independent of Reynolds number. During relaminarization, the turbulent kinetic energy decays exponentially nearly to zero within 3D/u_τ , while the flow accelerates to its terminal velocity over 65D/u_τ . The relaminarized flow exhibits half-vortical structures scaling in viscous units superimposed on a laminar profile, effectively reducing the pipe cross section. Within 180≤Re_τ≤ 540 , over 97

We present a tool chain to analyze the internal processes of turbulent flows, measured by Lagrangian Particle Tracking (LPT) and Particle Image Thermometry (PIT). It is based on data assimilation by means of Physics-Informed Neural Networks and a subsequent construction of joint distributions of the kinetic energy, velocity vector field curvature and temperature variance as well as the evaluation of their time derivatives. We apply this tool chain to a measured data set of Rayleigh-Bénard convection (RBC) at a Rayleigh number Ra = 3.4 × 10^7 and a Prandtl number Pr = 10.6 resulting in a curvature-based energy-spectrum and an investigation of the links between the physical and phase space.

We compare two methods to assimilate temperature and pressure fields based on given three-dimensional velocity fields of moderately turbulent Rayleigh-Bénard convection in a cubic geometry with system parameters Rayleigh and Prandtl number: Ra =1· 10^6 , Pr =0.7 . The first method describes a direct solution of the problem using the fractional step of the associated Navier-Stokes equation. The second method uses a physically informed neural network approach that learns the associated temperature and pressure fields by minimizing the residual loss of the set of equations governing the flow. The ground truth temperature, pressure, and velocity fields originate from a direct numerical simulation.

We predict the SARS-CoV-2 infection risk in aircraft cabins by simulating the aerosol transport with computational fluid dynamics and taking medical parameters into account. A recently presented new measurement technique allows us to measure the rapid virus inactivation after exhalation with high temporal resolution. In addition, much higher airborne SARS-CoV-2 inactivation rates than in previous studies were obtained. This raises the question of how the new knowledge of SARS-CoV-2 stability affects the prediction of infection risk. To answer this question, we evaluated 70 Lagrangian particle simulations with an index person sitting in all possible seats in an aircraft cabin. We then estimated the infection risk for the other passengers based on the old and new SARS-CoV-2 stability data. For typical transmission events, we found that the predicted infection risk is reduced by about 50 _2 (500 ppm). However, elevated ambient CO _2 concentrations of 3000 ppm protect the virus from inactivation and increase infection risk by about 50 _2 . In addition, a high relative humidity of the ambient air, e.g., from exhaled breath, delays the rapid inactivation by a few seconds, increasing the risk of infection for immediate neighbors.

PurposeThe purpose of this study is to perform direct numerical simulation (DNS) and unsteady Reynolds-averaged Navier-Stokes simulation (URANS) of cough-induced flow and particles in a large-scale circulation (LSC) to assess the performance and accuracy of the latter.Design/methodology/approachBoth simulations were performed in a 12.5 m(3) room for 30 s, with a background flow defined by a lid-driven LSC. In the URANS, the particles were modelled using a stochastic dispersion model to account for turbulent fluctuations. Initial flow fields were obtained from LSC simulations.FindingsThe URANS predicted a larger cough jet entrainment, resulting in a shorter but wider jet flow, leading to underprediction of the horizontal displacements, especially of the small particles, during the jet and early puff phases. The cough puff in the URANS was overly influenced by the downward background flow, resulting in faster particle descent. The wider jet spread led to an overprediction of particle dispersion during the jet and early puff phases, but subsequently the shorter puff spread led to an underprediction of particle dispersion during the mid and late puff phases.Originality/valueUnlike similar studies, this research includes DNS of a cough-induced flow and particles in a larger domain over a longer period of time within a background flow characterised by an LSC, highlighting the need for a better representation of the flow fields and cough-induced particle dynamics resolved in URANS.

The use of partially scale-resolving CFD (computational fluid dynamics) techniques represents a state-of-the-art approach to aerodynamic design, which is increasingly being adopted by the industry. It still remains crucial to validate the available tools by means of experimental data. The availability of experimental data for the full-scale DrivAer model is still limited when ground simulation is included. New experimental data, including flow field measurements, will be presented in this study. The performance of two commonly used CFD solvers, the finite volume method (FVM)–based open-source Toolbox OpenFOAM and the commercial lattice Boltzmann method (LBM)–based solver PowerFLOW will be analyzed and validated against experimental data. In particular, the flow field around the notchback configuration of the DrivAer model, which includes a shallow detachment region, has proven to be challenging to reproduce in CFD. The results obtained from delayed detached eddy simulation (DDES) exhibit a significant discrepancy with respect to experimental data and unsteady Reynolds-averaged simulations (URANS) in this region. The flow detaches prematurely when DDES is applied. This is caused by an incorrect behavior of the limiter, which is designed to shield the URANS mode from intrusion of large eddy simulation (LES) mode. The recently published enhanced protection (EP) method provides more favorable results in this area. LBM, while generally in good agreement with experiments, predicts a premature separation, most likely due to the Cartesian grid not allowing sufficient resolution. In another challenging area, namely the wheels, the scale-resolving methods show the best results.

Particle dispersion models (PDMs) are essential to capture the influence of unresolved turbulent eddies on particle transport in computational fluid dynamics (CFD) simulations. However, the validation of these models remains challenging, especially when relying on experimental data or CFD simulations that are based on turbulence models. In this work, we use time-averaged data obtained in a direct numerical simulation (DNS) instead of relying on turbulence models to model particle dispersion. In addition, a new particle dispersion model is presented, referred to as the limited particle–eddy interaction time (LPI) model. For a detailed and systematic evaluation of the new LPI model, we compare its performance with that of other commonly used models, such as the mean particle–eddy interaction time (MPI) model implemented in OpenFOAM® and the randomized particle–eddy interaction time (RPI) model from the literature. The MPI model shows good agreement with the DNS for the largest particles tested (Stokes number, St = 0.2) but exhibits erratic and unphysical trajectories for smaller particles (St ≤ 0.05). To mitigate this erratic behavior, we have adjusted the eddy interaction time in the new LPI model.

Rayleigh-Bénard convection at high Rayleigh number exhibits turbulence superimposed on large-scale circulation. While buoyancy forces drive the flow at certain scales, how kinetic energy is transfers across the scales is not understood. Here, utilizing a Lagrangian description of the kinetic energy flux, we present experimental evidence of a split cascade where energy flows downwscale at small scales and upscale at large scales. The flow topology of these energy transfer events differ profoundly, and the transition between them occurs gradually, over a broad range of scales.

Three-dimensional velocity fields of a large-scale reorientation in turbulent Rayleigh-B & eacute;nard convection at $ {\rm Ra} = 2.5\cdot 10<^>9 $ Ra=2.5 & sdot;109 in a 300 mm cubic water cell are measured using particle tracking velocimetry. Reorientations are rare events occurring about once every three days in our setup, involve the large-scale circulation switching between cell diagonals. The dominant flow structures of the reorientation are extracted from the measurement data using POD supported by symmetries of the cubic cell to mimic long time series of reorientation events. The decomposition reveals degenerate mode pairs. The first six modes of the decomposition account for about 70% of the total energy and contain the major coherent structures. Modes 1-3 reflect the orientation and dynamics of the large-scale circulation, and modes 4-6 provide the orientation and dynamics of the corner circulations. A pure Y roll structure is observed in the middle of the reorientation event.

We use the local curvature derived from velocity vector fields or particle tracks as a surrogate for structure size to compute curvature-based energy spectra. An application to homogeneous isotropic turbulence shows that these spectra replicate certain features of classical energy spectra such as the slope of the inertial range extending towards the equivalent curvature of the Taylor microscale. As this curvature-based analysis framework is sampling based, it also allows further statistical analyses of the time evolution of the kinetic energies and curvatures considered. The main findings of these analyses are that the slope for the inertial range also appears as a salient point in the probability density distribution of the angle of the vector comprising the two time evolution components. This density distribution further exhibits changing features of its shape depending on the Rayleigh number. This Rayleigh number evolution allows to observe a change in the flow regime between the Rayleigh numbers 10^6 and 10^7. Insight into this regime change is gathered by conditionally sampling the salient time evolution behaviours and projecting them back into physical space. Concretely, the regime change is manifested by a change in the spatial distribution for the different time evolution behaviours. Finally, we show that this analysis can be applied to measured Lagrangian particle tracks.

In technical applications, more than 90% of the energy required to pump the fluids through pipes is dissipated by turbulence near the wall. In this respect, streamwise traveling waves of wall blowing and suction have been used to relaminarize turbulent pipe flow at a low friction Reynolds number of Re_τ=110, considerably reducing friction losses and energy consumption. Here, we demonstrate that streamwise traveling waves of wall blowing and suction applied to initially turbulent pipe flow can trigger relaminarization up to friction Reynolds numbers of Re_τ=720. Furthermore, we perform a parametric study comprising both upstream traveling waves (UTWs, c<0) and downstream traveling waves (DTWs, c>0) by varying the traveling wave amplitude a, celerity c, and wavelength λ at Re_τ=180 and Re_τ=360 in order to investigate the scaling of the maximum drag reduction and of the net energy saving rate in direct numerical simulations. Consistent with channel flow studies in the literature, we found that UTWs destabilize the flow, while generating sublaminar drag due to the pumping effect. However, only low-speed UTWs with large amplitudes were discovered to decisively reduce energy consumption. For DTWs, a large range of wave parameters lead to a conspicuous drag reduction. Nevertheless, only a subgroup of these wave parameters are associated with siginificant net energy savings, i.e. 0.067U_c,lam≲ a ≲0.1U_c,lam, c ≈ U_c,lam, and λ≈ 360δ_ν, independent of the Reynolds number. Here, U_c,lam=1/2Re_τu_τ is the centerline velocity of the corresponding laminar flow, and δ_ν= ν/u_τ is the viscous length scale.

An important route of transmission for potentially harmful bacteria is the spread of bioaerosols in indoor environments. In a chamber specially developed for particle dispersion tests, we created a defined bioaerosol to study the performance of two methods commonly used in biology and engineering studies: airborne bacterial detection and particulate matter (PM) analysis. A total of five ventilation cases were investigated in which an air curtain, operated at Reynolds numbers Re < 11, 000, shielded the particles in one half of the test chamber from the other half. In two of these five cases, a HEPA filter was also installed to specifically reduce the particle concentration in the test chamber. In addition to active and passive air sampling measurements of bacteria, we took PM measurements in front of, beneath, and behind the air curtain under constant air temperature and relative humidity conditions. The bioaerosol contained nine bacterial species, evenly distributed in artificial saliva. Two species in the bioaerosol, Staphylococcus capitis DSM 111179 and Burkholderia lata DSM 23089 T , were selected for evaluation due to their antibiotic resistance, which makes them distinguishable from other species. The results show a similar trend in the concentrations of the detected particles and bacteria. The survival rates of the evaluated bacterial species differed; S. capitis exhibited a greater agreement with the PM measurements than B. lata did, which emphasizes the importance of using a various model organism in such experimental setups. We evaluated the effectiveness of the air curtain in reducing particle and bacterial spread, with values reaching up to 66% for both measurement approaches. This study highlights the key differences between the two detection methods and confirms the reproducibility and suitability of the standardized bioaerosol for future research applications. Both methods have demonstrated their potential for use in more realistic scenarios.

This computational fluid dynamics (CFD) study examines the comfort parameters of an innovative air vent concept for car cabin interiors using a reduced order model (ROM) and proper orthogonal decomposition (POD). The focus is on the analysis of the influence of geometric and fluid mechanical parameters on the resulting jet, in particular on the deflection angle of the airflow and the total pressure difference along the outlet geometry. Different parameters of the investigated system, such as the surface orientation, the outlet height, the separator distance, and the separator height, lead to different effects on the airflow structure. The results show that changes in the air vent surface orientation are always accompanied by an increase in the deflection angle and the total pressure difference. In contrast, the variation of the outlet height ratio positively influences the deflection angle and the total pressure difference in terms of the requirements for air vent geometries. The study also examines the interaction of the geometric parameters and reveals complex correlations that influence the resulting air jet. A comprehensive understanding of these influences makes it possible to adapt the design and implementation of new and innovative air vent concepts to meet specific requirements. By balancing design considerations and technical requirements, optimized solutions are characterized by a high deflection angle and a reduced overall pressure difference for improved system performance and efficiency. Therefore, this evaluation provides a final framework for the design and implementation of an innovative air vent concept based on the volume flow vectoring that is tailored to specific application requirements.

We present a method to infer temperature fields from stereo particle image velocimetry (PIV) data in turbulent Rayleigh-B & eacute;nard convection (RBC) using physics-informed neural networks (PINNs). The physical setup is a cubic RBC cell with Rayleigh number Ra = 107and Prandtl number Pr = 0.7. With data available only in a vertical plane, the residuals of the governing partial differential equations are minimized at a set of collocation points in an enclosing 3D domain of finite thickness along the direction perpendicular to the plane. Dynamic collocation point sampling strategies are used to overcome the lack of 3D labeled information and to optimize the overall PINN convergence. In particular, in the out-of-plane direction, the collocation points are distributed according to a normal distribution, in order to emphasize the region where data is provided. Along the vertical direction, we leverage direct numerical simulation (DNS) meshing information and sample points from an optimized kernel-density estimation. This sampling approach balances labeled information by pointing greater attention to critical regions, particularly in areas with high temperature gradients within the thermal boundary layers. Using DNS planar three-component velocity data, we successfully validate the accurate reconstruction of the temperature fields in the PIV plane. We evaluate the robustness of our method with respect to characteristics of the labeled data used for training: the data time span, the sampling frequency, some noisy data, and omission of boundary data, aiming to better accommodate the challenges associated with experimental data. Developing PINNs on controlled simulation data is a crucial step towards their effective training and deployment on experimental data. The key is to systematically introduce noise, gaps, and uncertainties in simulated data to mimic real-world conditions and ensure robust generalization.