Forecasting extreme climate events, for instance extreme heat waves, is key for society and a scientific challenge. In this paper we propose a novel machine learning approach that successfully forecasts extreme heat waves up to 45 days before the end of the event. The approach allows for dynamical process studies. A key message is that optimal machine learning forecasts require a large amount of data. The image shows temperature (colors) and geopotential height (lines) anomalies for a typical atmospheric situation.

Increasing the ventilation airflow speed is commonly assumed to decrease the probability of airborne infectious disease transmission, since higher airflow means more fresh air and lower concentrations of airborne pathogens. Here we demonstrate experimentally that increasing the airflow speed by a factor of 10 had no impact on the probability of influenza transmission between guinea pigs. We use perfect-mixing and Gaussian plume models to interpret this result as evidence that the virus is primarily carried on virus-contaminated dust, rather than in expiratory aerosols as commonly assumed.

Pulsed jetting through cyclic refilling and discharging of a pressure chamber is an effective method to achieve high swimming speed. Inspired by squids, we demonstrate that an auxiliary inlet that opens during refilling and closes during discharging greatly reduces the energy expenditure and improves efficiency. The opening and closing actions of this inlet is controlled by a passive valve which is mechanically simple.

The dispersion of red blood cell transit times in capillaries and microvascular networks is strongly correlated with their deformability. In this paper, we show that a pulse of red blood cells undergoes axial dispersion as it flows along a narrow channel and that the evolution of the pulse length is mainly determined by the transverse migration of cells. This simple macroscopic axial dispersion measurement can be used to derive microscopic cell migration parameters which are a signature of red blood cell deformability.

The dynamics of a single Newtonian drop in a turbulent microscopic flow are strongly influenced by the shear thinning behavior of the continuous phase. The central aim of this study is to map and describe the various dynamic modes of a single Newtonian drop in terms of both the rheology of the continuous phase and the Reynolds number. In the presence of shear thinning the drops are elongated in the form of long filaments, and the stronger the shear thinning behavior of the matrix, the longer their characteristic break time is. Single drop dynamics is found to be dominated by local extension. Our findings could be useful for addressing the turbulent emulsification of non-Newtonian fluids.

By studying the spreading behavior of aqueous glycerol and polymer solutions on smooth surfaces, we found that shear-thinning solutions behave similarly to water in the initial spreading phase, while glycerol solutions show significantly slower spreading. For the glycerol solutions, an increase in glycerol concentration effectively increases the contact-line friction, resulting in increased resistance to wetting. For the polymeric solutions, however, an increase in polymer concentration does not modify contact-line friction. As a consequence, the energy dissipation at the contact line can not be controlled by varying the amount of additives for shear-thinning fluids.

Partitions in enclosures change transport rates significantly and play an essential role in engineering applications. We have shown that heat transport can be increased by up to 250% using partitions in thermal convection. Pressure-driven forced convection is established on the conduction wall due to the increased plume ejection and the impact zones; thus, high heat transport. The theoretical framework and numerical simulations developed in this work show that heat transport is proportional to the cubic power of the constriction gap between the partitions and the conduction wall.

An experiment is conducted investigating the dynamics of a droplet as it impacts a fiber array. To remove gravitational effects, experiments are conducted in the unique low gravity environment of a drop tower. The droplet dynamics are described with a model showing the dominating role of contact line dissipation. The model shows good agreement with experiments for important metrics such as the droplet penetration length.

Our work investigates cleaning surfaces coated with protein solutions or bacterial biofilms using continuous collisions and sliding air bubbles in an aqueous medium. Air bubbles between 0.5-1 mm in radius perform best when tilted at 20-25 degrees with respect to the horizontal plane. Based on our model, the interplay between the steady sliding speed and the steady film thickness between the bubble and the surface yields the best cleaning at 22.5 degrees. The technique offers a safe and environmentally friendly way of cleaning fresh agricultural produce without damaging it or reducing its freshness period.

In this work, we simulated the coarsening of magnetic particles under a shear flow and for different magnetic fields. We found that particle assembly varied in fully three-dimensional simulations and constrained two-dimensional simulations, which affected the corresponding rheological properties.

We experimentally investigate dissolution morphologies in a radial Hele-Shaw cell. We observe a unique phenomenon where the dissolution front shifts from stable to unstable and uncover the role of buoyancy-driven convection in forming radial stripes (dissolution “flower”) on the solid surface. We propose a phase diagram to predict the transitions of dissolution patterns.

Three consequent bifurcations of buoyancy convection in a laterally heated cube with perfectly thermally insulated boundaries are studied. With the growth of the Grashof number, the flow becomes oscillatory unstable, then the stability reinstates at a larger Grashof number, with further increase of which the flow finally becomes unstable. It is shown that all three transitions result from an interaction between the destabilizing centrifugal mechanism and the stabilizing effect of thermal stratification.

Confined capillary flows can have dramatic effects on systems containing structures soft enough to be deformed by surface tension. In this work, we study the drainage of a defending oil phase by an immiscible water phase through channels textured with arrays of elastic beams, which constitute a compliant porous medium. The coupling of flow dynamics with elastocapillarity causes the entrapment of the defending phase within periodic self-assembled bundles.

Surfactants can have a detrimental effect on the drag reduction in shear flow over superhydrophobic surfaces in the Cassie state. Assuming an incompressible, inviscid surfactant phase at the gas-liquid interfaces, we study liquid flow over an array of narrow gas-filled grooves embedded in an otherwise planar surface, with the gas-liquid interface protruding above or below the plane. A recirculating flow develops at the gas-liquid interfaces, such that the slip length for flow over such surfaces is much smaller than at corresponding surfactant-free interfaces.

In this work, high-fidelity numerical simulation is employed for the quantitative study of an aerated cavity induced by a surface-piercing plate. Numerical simulation shows that a large-scale aerated pocket with a conical shape is formed at the fringe of the plate, many bubbles pinching off from it. Numerical results reveal that in the process of free-surface flow around a semi-submerged structure, the aerated cavity contributes a majority of air entrainment in the wake.

The locomotion of microorganisms such as Paramecium is driven by helical metachronal waves. Motivated by that, we propose a simple model for the swimming motion of microorganisms based on an infinite cylinder with helical surface waves. The model predicts both the rotational motion of the microorganisms, and anomalous oscillations normal to their axes for the first mode of a purely azimuthal wave.

We numerically investigate the rheology of concentrated suspensions of fibers by modeling them as continuous slender bodies suspended in an incompressible fluid, including surface roughness. We model contact dynamics and friction in the suspension using a normal load-dependent friction coefficient and investigate the shear rate-dependent rheology of suspensions for different aspect ratios, volume fractions, roughness, and flexibility values.

In this work, we numerically investigate the loads on a finite-length cylinder embedded in a steady flow. We demonstrate that Khayat \& Cox (JFM 1989) slender-body theory can predict with reasonable accuracy the drag force on the cylinder for a large range of aspect ratios. However, for moderately long cylinders, this theory is not accurate in predicting lift force and torque. Therefore, we develop semi-empirical models based on small but finite inertia theory to improve the match between our numerical predictions and theoretical results. Finally, we compare our entire model, with experimental data on settling cylinders.

We investigate the dynamics of a single flexible fiber settling in a viscous fluid embedded with obstacles of arbitrary shapes. We show that the presence of obstacles may lead to two types of events: gliding and trapping. In particular, we observe nontrivial trapping conformations on sharp obstacles that result from a subtle balance between elasticity, gravity, and friction. In the gliding case, a flexible fiber reorients and drifts sideways after sliding along the obstacle. We show how these effects can be leveraged to propose an efficient strategy to sort fibers based on their size and/or elasticity.

In two-fluid coaxial atomization, the average length of the liquid jet is governed by the gas-to-liquid dynamic pressure ratio. We use high-speed high-resolution back-lit imaging to measure the time series of the liquid jet length over a broad range of operating parameters, resulting in over four orders of magnitude of the dynamic pressure ratio. We introduce a framework to characterize the full statistics and the temporal dynamics of the liquid jet length, which underlines a change of behavior, that stands as a quantitative signature of change in atomization regimes, from membrane break-up to fiber-type atomization.

We experimentally investigate the problem of a propulsive rotor submerged in water with flexible blades in both spanwise and chordwise directions. When we rotate the propeller and set an incoming flow, the blades bend and twist. We report measurements of forces and associated blade deformations. We show how the direction and magnitude of the deformation can be understood through simple geometric arguments, and highlight a relationship between the thrust and bending angle.

Minimal dimensional models are desirable for reduced computational costs in simulations as well as for applications such as model-based control. Long-time dynamics of flows often evolve on a low-dimensional manifold M in the full state space. We use neural networks to estimate M and the dynamics on it for two-dimensional Kolmogorov flow in a chaotic bursting regime. Outcomes include: a minimal dimension estimate, good short-time tracking and long-time statistics, as well as accurate predictions of bursting events.

We provide an experimental benchmark for modeling nondilute miscible mixing in porous media, which emerges in CO${}_2$ sequestration and solvent-based enhanced oil recovery. In a near-fracture porous medium, we show that conventional Darcy-Fickian coupling fails to predict nondilute fluid-fluid displacement, although it works perfectly for dilute systems. Nondilute fluid systems show “diffusive” mixing kinetics, wedge-like mixing front, and significant convection at the front that is perpendicular to the concentration gradient. These above abnormal phenomena are rationalized by Korteweg stress, which is still challenging in modeling.

Pore-size graded multilayered membrane filters have shown improved performance in terms of foulant retention and filtrate production. In this work, we model such a filter as a network of circularly cylindrical pores, whose radii decrease linearly across several layers of membrane material. Under two different filtration modes - flux extinction and flux threshold - we find that there exist performance-optimizing radius gradients, which correlate strongly with the depth of foulant penetration into the membrane.

Shock-wave boundary layer interactions are of utmost importance for both aeronautical and aerospace applications. The dynamics of high-speed compressible turbulent boundary layers becomes tightly coupled with strong gradients of thermodynamic properties, leading to an increase of thermomechanical loads. In the contextual presence of hypersonic and high-enthalpy regimes, thermochemical nonequilibrium effects must be considered as well. This work is a first step toward the fundamental understanding of this configuration.

In Burgers turbulence driven by an Ornstein-Uhlenbeck process, for which the characteristic time scale is much larger than the characteristic time scale of energy transfer, we find the spatial-temporal third-order structure function follows a separation-of-variables form: the spatial dependence follows the classic results calculated from the Karman-Howarth-Monin equation, and the temporal dependence is determined by the temporal correlation of the Ornstein-Uhlenbeck process. This result confirms the forcing dependence of the spatial-temporal structure function and provides an intuitive way of analyzing sparsely measured spatial-temporal turbulence data.

We generalize Townsend’s classical attached-eddy hypothesis (AEH) to eddies with arbitrary fractal dimension. New scaling laws for the Reynolds stress components are obtained. Preliminary evidence from high-resolution simulations of adverse-pressure-gradient turbulent boundary layers and wing boundary layers are provided. Furthermore, we investigate the cascade self-similarity of general attached eddies, and obtain relationships between the power-law exponents of the probability density, population density, area coverage, volume fraction, and fractal dimension of general attached eddies, leading to insights into the connection between turbulent coherent structures and statistics.

The spectral transport equations of turbulence kinetic and internal energy (TKE and TIE) are rigorously derived in the compressible turbulent flows, exhibiting a full structural similarity in terms of two-point correlations of velocity and a sound-speed-like variable. The scale dependence of TKE and TIE flow across space and scales, and among components is revealed comprehensively in a similar manner, in hypersonic turbulent boundary layers. Special attention is paid to the effect of wall cooling on the energy evolution, which is observed to exist throughout the boundary layer primarily by means of modifying the small-scale motions.

We perform wind tunnel tests on the turbulent flow around a square-back Ahmed to analyze the use of rear flexibly hinged parallel plates as a control strategy to reduce the drag in a self-adaptive manner under changing flow conditions. Those rigid flaps are able to rotate, adopting an angle, θ. The observed fluid-structure interaction phenomena results in an important decrease of the drag coefficient of the original body, outperforming purely rigid flaps in aligned and yaw conditions. Additionally, hinged flaps are shown to interact with the reflectional-symmetry-breaking (RSB) modes, typically present in the wake of three-dimensional bodies.

In this paper, we demonstrate that the problem of turbulent bursts can be tackled through a complex systems approach. Specifically, by considering both duration and intensity of the bursting events, one can incorporate the effect of bursts on the turbulence statistics at any specified scale of the flow, thereby allowing one to connect the origin of bursts to the presence of organized eddy motions in turbulent flows called coherent structures. Through our approach we discover a particular aspect of universality associated with turbulent bursts and contribute toward the development of next-generation turbulence models by creating a union between complex systems science and fluid mechanics.

A novel concept of vortical cleaning in porous surfaces is studied experimentally. A vortex ring of various strengths is made to interact with oil-impregnated porous surfaces with the aim of understanding the mechanism of oil ejection from the porous surface. The vortex dynamics involves different phenomena like vortex cancellation and Kelvin-Helmholtz instabilities. The cleaning takes place from both sides of the porous surface through an intricate interaction process characterized by Rayleigh-Taylor and Rayleigh-Plateau type instabilities that is studied in three different regimes: i) Penetration, ii) Bag formation, and iii) Bag breakup.

When emptying a water filled container through a bottom hole, a “bathtub vortex” may appear and deform the free surface. Torricelli’s law, which predicts the flow rate as a function of the water height, has been known for 400 years, but deviations are observed in the presence of such a vortex. This study focuses on the impact of the bathtub vortex on the emptying velocity. Through an experiment on the unsteady draining flow in a rotating tank, we show that Torricelli’s law is modified as a function of the surface deformation, and that the draining time is mainly determined by a nondimensional parameter corresponding to the ratio of the outlet size to the Ekman boundary layer thickness.

We show experimentally that shear thinning can cause suppression of vortex-induced vibration (VIV). Increasing shear thinning in a fluid flow causes a decrease in the critical Reynolds number at which VIV occurs. VIV occurs within a range of reduced velocity, called the lock-in range. The reduced velocity at which lock-in begins, the width of the lock-in range, and the oscillation amplitude decrease for a shear thinning fluid, and beyond a critical concentration, oscillations are completely suppressed.