The collapse behavior of nanobubble in the vicinity of solid-liquid interface under impact loading critically influences material erosion, cavitation damage, and surface degradation in fluid-structure systems. This study employs the Reactive Molecular Dynamics (RMD) method to investigate the dynamic processes of nanobubble collapse near solid-liquid interface at realistic impact velocities. The simulation procedure involved initial Canonical ensemble (The NVT ensemble) equilibration of water molecules and iron atoms, followed by microcanonical ensemble (The NVE ensemble) impact loading via piston driving with an initial velocity of 5 km/s in the z-direction. We tracked the propagation of shock waves, the evolution of atomic velocity fields, and the spatiotemporal evolution of nanobubble radius, revealing complex deformation and collapse characteristics of nanobubble with different initial radius. Results show that nanobubble collapse morphology, energy localization effects, and momentum transfer to iron substrates are highly sensitive to nanobubble size. Notably, larger nanobubble induce local localized compressive-stress peaks that significantly alter substrate surface roughness and microstructural features, shedding light on potential mechanisms of surface erosion and fatigue initiation. This study elucidated the cavitation damage mechanism at the atomic scale, verified the reactive force field (ReaxFF) method's analytical capability in coupling chemical-mechanical responses under extreme conditions, and provided a theoretical basis for multi-scale modeling of cavitation erosion phenomena in engineering materials.
The study established a three-dimensional transient model of the entire continuous casting process for bloom. This model was developed using the finite element simulation software Fluent, in conjunction with the large eddy simulation (LES) turbulence model, solidification and heat transfer model, and discrete phase model (DPM). A ternary inclusion capture criterion, incorporating liquid phase volume fraction, inclusion velocity magnitude, and direction, was developed using user-defined functions (UDF). Accordingly, the study investigated the three-dimensional transient flow field, temperature field, shell solidification, and inclusion migration process during the continuous casting of bloom for rail steel. The model predicted the distribution patterns of inclusions with different particle sizes across the entire cross-section of the bloom. The result shows that slight transient changes occur in the flow and temperature fields within the mold, and the flow field exhibits a certain degree of asymmetric distribution. The impact of the molten steel jet within the mold significantly affects on the distribution of shell thickness. There are differences in the distribution of the removal locations of inclusions with different diameters by the upper surface. As the diameter of inclusions increases, they are more likely to float up to the meniscus and be removed. All inclusions are enriched in the center of the narrow side and near the corner of the wide side of the bloom, and the large-size inclusions are enriched in the range of 1/4 to 1/3 near the inner arc side. Moreover, the larger the inclusions, the closer they are to the inner arc.
The study of fuel droplet breakup processes in detonation waves is critical for building accurate models and improving the performance of liquid fueled detonation engines. However, there is currently a scarcity of experimental breakup studies that examine micrometer-scale droplets in gas-phase detonations. In this work, we present the first high-speed measurements of liquid fuel droplet breakup behind gas-phase detonation waves using digital holography and coherent imaging methods at sampling rates of up to 10 MHz. Here, individual 4.4 to 188 μm droplets of ethanol and n-dodecane are generated. These are then studied in atmospheric hydrogen-oxygen detonation waves with detonation wave speeds ranging from 2300 to 3200 m/s (Mach 5.2 to 5.3). Results show catastrophic breakup geometries and droplet lifetimes ranging from 0.2 to 17.7 μs. Non-dimensionalization with traditional droplet breakup analysis techniques successfully capture the drag and lifetime behaviors for the larger droplets in this range. Smaller droplets exhibit greater variability in nondimensional displacement and lifetime, potentially reflecting increased sensitivity to local flow-field variations and higher measurement uncertainty.
Prediction of the partitioning of suspended particles at fracture bifurcations is a crucial issue in hydraulic fracturing, and so far there has been no model for particle partitioning. In this study, three-dimensional CFD–DEM simulations of the transport of neutrally buoyant spherical particles in a primary fracture connected with a branch are conducted. A model for the fluid partitioning of the single-phase flow is first established based on theoretical analysis and simulation data. We then establish correlations for the particle partitioning from fitting of our simulation data, as a function of bulk Reynolds number, branch-to-primary fracture width ratio, particle-to-branch size ratio and branch angle. The number ratio of particles entering the branch increases with increasing branch-to-primary aperture ratio, increasing branch angle, decreasing particle-to-branch size ratio and decreasing Reynolds number. The effects of the particle volume fraction at inlet on the fluid and particle partitionings are also examined. Our results indicate that as the particle volume fraction increases, the fluid partitioning monotonically increases, whereas the particle partitioning first decreases and then increases.
In previous work, a binary lattice Boltzmann model for simulations of a mixture of an ideal solute and an ideal soluble gas was constructed. This model made use of a combination of the well-balanced and binary free-energy lattice Boltzmann methods and was shown to recreate the correct equation of state at high density ratios and mass transfer between components. Here, we present an improved model that implements the free energy in terms of partial pressures, application of which guarantees conservation of momentum. This approach, combined with further cancellation of errors on the flow equations, can be used to accurately describe dynamic systems. This is demonstrated by simulation of the free rise of dissolving gas bubbles in three dimensions. The mass transfer and bubble rise velocity are benchmarked with correlations of Sherwood number and rise velocity. The model is shown to successfully simulate up to moderate Reynolds and Eötvös numbers. The capacity of the method to handle topological changes is demonstrated by simulations of electrolysis. The electrolysis simulations are quantitatively verified by demonstration of different regimes of bubble growth depending on the Damköhler number. Electrolysis is modelled through implementation of a mass conversion boundary condition which simulates the cathode/anode of an electrolysis cell and converts one component to the other. Finally, simulations that show bubble formation at varying conditions are presented.
The selective detachment of particles with differing compositions remains a significant challenge in industrial applications. In this study, we demonstrate the use of a simple gas-liquid interface to separate floating solid particles with distinct physical properties. Firstly, based on the potential energy equation governing the gas-liquid-solid three-phase system, analytical equilibrium equations representing the gas-liquid interface characteristics of a single particle are derived using a differential approach (energy minimization). We then present the evolution of the three-phase quasi-static structure with respect to particle density, size and surface wettability. In particular, utilizing a force-balance approach, the detachment criteria for particles of different sizes and densities are established through an analysis of structural and mechanical characteristics, showing good agreement with simulation results of Surface Evolver. For two particles situated at a gas-liquid interface, Surface Evolver simulations reveal that the maximal particle density ρmax that the gas-liquid interface can support increases with growing particle separation distance in the hydrophobic situation, whereas it remains nearly constant in the hydrophilic situation. This study provides insights into the mechanisms governing the detachment behaviors of floating particles and offers detachment strategies.
Precise control of ultrasound-driven bubble oscillations is important in many applications but remains challenging near solid boundaries, where wall-induced acoustic coupling strongly alters the bubble response. This work develops a physics-based control framework for near-wall bubble oscillation under bounded ultrasound. The framework integrates a wall-modified Rayleigh–Plesset (R–P) model with a time-varying fast terminal sliding-mode controller (TSMC), and is evaluated for both target-radius regulation and prescribed radius-trajectory tracking. For target-radius regulation, the controller drives the bubble radius to the target value from different initial phases, whereas excessively small targets become unreachable because of the rapid growth of internal gas pressure during compression. This mechanism yields a theoretical lower reachability threshold of 0.5646 times the initial radius. For radius-trajectory tracking, accurate following is obtained for low-frequency, small-amplitude oscillation trajectories. Increasing frequency or amplitude, however, induces phase lag and error accumulation near the minimum-radius stage. A non-monotonic reachability boundary with respect to target frequency is further derived. To examine the cross-model transferability of the controller, the control input is deployed in a high-fidelity multiphase flow solver, where the resulting bubble response remains consistent with the R–P prediction. Compared with conventional SMC and TSMC, the proposed controller achieves faster convergence and improved tracking accuracy. Overall, this work provides a framework for ultrasound-driven bubble control near a rigid wall while quantifying the physical reachability bounds.
To address the urgent thermal management requirements of high-heat-flux electronic devices, a gradient IWP wick design strategy for vapor chamber applications is proposed and systematically investigated. Based on the IWP triply periodic minimal surface (TPMS) sheet architecture, uniform wick structures with volume fractions of 10%, 20%, 30%, and 40% were first established; based on these baseline configurations, three gradient transition strategies were then designed. Transient gas–liquid two-phase simulations were conducted in COMSOL Multiphysics to evaluate the capillary transport behavior. The influence of the wick volume fraction and gradient distribution was assessed through the temporal evolution of the liquid volume fraction, together with the velocity field, pressure field, and cell Reynolds number. The results show that volume fraction is a dominant factor governing the capillary performance of IWP sheet structures. Low-volume-fraction structures exhibit faster liquid penetration, higher liquid occupancy, and stronger capillary replenishment capability, whereas high-volume-fraction structures suffer from reduced pore connectivity, increased flow resistance, and more pronounced vapor entrapment. Compared with uniform structures, gradient architectures effectively reconstruct internal transport pathways, improve liquid-front propagation, and optimize local flow organization, thereby achieving a better balance between capillary driving force and flow resistance. Among the investigated cases, the 20%→80% gradient structure exhibits the best overall liquid-filling capability and flow activity, while the unidirectional 80%→20% gradient structure shows a stronger pressure response and more evident directional transport characteristics. Steady-state thermal simulations further reveal that the 80%→20% gradient structure achieves the highest total heat flux and the most uniform temperature distribution, driven by enhanced convection through hierarchically enlarged pores, despite its slower capillary penetration. These findings provide theoretical guidance for the design of high-performance, lightweight, and structurally integrated wick structures for next-generation vapor chambers.
In this work we report on the Infra-Red temperature measurements of an ethanol evaporating meniscus interface formed at the mouth of a capillary tube of 1 mm internal diameter with changing axis orientation respect to the gravity vector. Five tube orientations have been analyzed, namely: 0degree (vertical tube with meniscus interface facing upwards); 45degree; 90degree (horizontal tube); 135degree; and 180degree (vertical tube with meniscus interface facing downwards). What we found is that when the tube is vertical (with the meniscus facing upwards or downwards), the temperature profiles along vertical diametrical sections of the tube are broadly symmetrical with respect to the tube axis. Instead with the other tube orientations, the symmetry of the temperature profiles found for the vertical tube is lost and the upper parts of the temperature profiles are locally at higher temperature than the lower part. This clearly shows that the physical presence of the tube in the gravitational field distorts the buoyancy driven cold plume in the external airflow that is ultimately responsible for the asymmetry found in the temperature and velocity fields in the internal liquid. We adapted a simplified dimensional analysis for the inclined tube, which was originally developed for isothermal tubes, that also takes into account the changing external tube wall temperature and the predicted results are broadly in line with the experimental trends. The results reported in this experimental study on evaporation are very relevant to engineering applications such as evaporative coolers, vapor compression refrigeration systems, drying of drops, combustion, heat pipes, and porous media.
Motivated by the application of spray combustion in the aviation industry, this work investigates the dispersion of non-spherical droplets in turbulence. The most common strategy for modeling sprays relies on the Lagrangian particle tracking (LPT) method, which represents the spray as a discrete collection of spherical particles. One limitation of this approach is that it neglects the influence of droplet deformation on the spray dynamics. Prior studies have highlighted the importance of non-sphericity in droplet vaporization, combustion and drag coefficient. However, these works are restricted to idealized configurations such as an isolated droplet in a uniform flow. To study droplet deformation in a more realistic configuration, we adopt homogeneous isotropic turbulence (HIT) as the framework to investigate its effect on droplet dispersion. Droplets of various Stokes number are studied to investigate the interplay between deformation and inertia. Analysis of droplet statistics reveals that the impact of droplet deformation on both dispersion and clustering is dependent on the inertia regime. For weakly-inertial droplets, deformation weakens both dispersion and preferential concentration, whereas for strongly-inertial droplets, deformation tends to enhance preferential concentration while weakening dispersion. The results also suggest that to achieve the same level of clustering, deformed droplets require a higher Stokes number. Interestingly, for non-inertial droplets, the deformation seems to induce an effective inertia. This is verified by a comparison between the full unsteady Taylor Analogy Breakup (TAB) model and its steady-state limit, which suggests that unsteady shape dynamics affect temporal correlation statistics, but leave the mean clustering pattern unchanged. These findings demonstrate that accounting for droplet deformation and its unsteady shape oscillation is essential for accurately predicting droplet dispersion and clustering in turbulence.
The immersed boundary method (IBM) is a renowned approach for numerical simulations of fluid–structure interactions, whose methodology and numerical techniques have been refined and are commonly integrated into classical Navier–Stokes solvers such as the finite difference, finite element, and finite volume methods. However, the effects of kernel functions and their implementation are seldom discussed within the lattice Boltzmann method (LBM). The theoretical foundation of LBM has distinctions from classical Navier–Stokes solvers, requiring careful implementation of IBM in the LBM framework. The streaming and collision processes in LBM restrict fictitious LBM fluid particles from traveling more than one nearest node per time step. For direct-forcing IBM-LBM, commonly used kernel functions span more than a single fluid node, potentially causing inaccurate hydrodynamic forces in dense particle-fluid systems. In this work, we investigate the effect of kernel functions, including commonly used Dirac delta functions and new kernels introduced in this work, on the accuracy of fluid–particle interactions. We find that commonly used Dirac delta functions have limitations in accurately capturing fluid–particle interactions when particles are close to each other. To overcome this issue, we propose more compact kernel functions, including an adjusted cosine function and a combined linear-shape/error-function kernel. The proposed kernels are validated through single particle sedimentation and two-particle drafting-kissing-tumbling (DKT) benchmark tests. Compared with existing DKT simulation studies, the proposed compact kernels perform better than the existing Peskin kernel.
A static bubble confined within a channel serves as a fundamental topic. Nevertheless, significant gaps remain in the fundamental understanding of the structural and mechanical characteristics of such trapped bubbles. Here, we formulate the model's energy equations and minimize the system's energy to determine the quasi-equilibrium state of the trapped bubble in a rectangular channel. As the bubble size increases, four distinct regimes characterize the evolution of bubble pressure, whereas at a fixed bubble volume, three regimes describe the pressure variation with respect to the square cross-sectional dimension of the channel. Analytical approximations for the bubble extension along the confined channel are also developed. Moreover, in a rectangular channel featuring a square cross-section and uniform wettability, it is observed that half of the total area of flat gas-liquid interfaces on the top and the bottom walls approximately equals the area of a flat gas-liquid interface on one of the side walls; this relationship similarly applies to the forces exerted by the channel walls on the bubble at low Bond number. Furthermore, under partial wettability conditions, a gas-solid interface forms on the channel wall, and when the sum of the contact angles on two adjacent walls reaches or exceeds 90°, the liquid menisci near the channel corner vanish.
The separation of a moving structure driven by an underwater supersonic gas jet involves complex fluid-structure interactions, among which the bidirectional coupling mechanism between the unsteady jet dynamics and the moving body remains poorly understood. In this study, a high-fidelity numerical model rigorously validated against experiments is established to investigate the coupling mechanisms under different jet expansion conditions. The results reveal that a self-sustained instability cycle of the jet—necking, bulging, breaking, and back-attack—dominates the pressure pulsations acting on the separating structure surface. Meanwhile, the motion of the separating body, constrained by a narrow radial gap, actively intensifies the jet unsteadiness, thereby establishing a strong bidirectional feedback that significantly controls jet evolution and separation efficiency. To further elucidate the underlying mechanism, an interface entrainment coefficient is introduced to quantify the mass and momentum exchange across the gas-liquid interface, providing a unified scaling for the evolution of jet instability. This study establishes a predictive framework for fluid-structure coupling in underwater supersonic separation systems.
Eulerian-Lagrangian CFD (Computational Fluid Dynamics) simulations of liquid ammonia sprays remain challenging, particularly in consistently predicting spray morphology and droplet size across operating conditions. This is because ammonia's physical properties differ significantly from those of common hydrocarbons, and so traditional models do not work effectively. At the same time, ammonia is attracting increasing interest as a fuel for combustion applications such as propulsion systems, gas turbines, and industrial burners. In particular, for thermal-power applications, the direct injection of liquid ammonia represents an attractive solution due to its potential to improve both volumetric and overall thermal efficiency. In this context, we propose a comparative study of two different Eulerian-Lagrangian breakup modeling approaches for simulating evaporative and flash-boiling ammonia sprays. Rather than resolving the micro-scale internal nozzle flashing, we focus on capturing the macroscopic downstream spray behavior, plume expansion, and vaporization in a computationally efficient framework suitable for application-scale simulations. In the first approach, we used correlations for the time and size constants of the KH-RT breakup model, as well as for the spray cone angle. The second approach is based on an effervescent breakup model (fbBreakup), which accounts for bubble growth within droplets and is applied here to ammonia. We evaluated the numerical results against experimental data for the spray issued by a multi-hole injector, including liquid and vapor tip penetrations, overall spray morphology, and local Sauter Mean Diameter (SMD) values. We also quantified the significant cooling effect induced by ammonia during the phase change process. The results show that while the calibrated correlation approach reproduces the spray structure well, the fbBreakup model behaves autonomously without case-specific tuning. We conduct an extensive discussion to highlight the advantages, limitations, and downstream thermal characteristics of both methodologies.
In this study, the collapse dynamics of shock wave interaction with tandem bubbles near a solid boundary is investigated via the numerical approach. Two dimensionless parameters, including the dimensionless distance δ between tandem bubbles and the dimensionless standoff distance γ between the proximal bubble and the solid boundary are examined. The simulation results revealed that the collapse of tandem bubbles can be categorized into three distinct patterns based on δ and γ, including bubble coalescence, tandem bubble collapse, and bubble shielding. The maximum jet velocity umax and wall pressure pwall during cavitation bubble collapse is affected by three factors, including the intensity of the incident shock wave, the dimensionless distance δ and γ. Quantitative analysis shows that due to the bubble shielding effect, the dimensionless wall pressure pwall,max/pshock decreases by an order of magnitude compared to other collapse patterns. Furthermore, three peaks caused by the water hammer shock wave and secondary shock wave, with the highest water hammer pressure value occurring near the solid surface. Finally, as the distance between the proximal bubble and the solid boundary increases, the collapse time of the tandem bubbles extends accordingly, and the incident shock wave significantly reduces the collapse time of the tandem bubbles. Overall, the main findings of this study are believed to benefit a better understanding of the synergistic collapse mechanism of the shock wave interaction with bubble clusters.
The flow past a fully submerged rotating cylinder beneath a free surface involves complex multiphase interactions, including free-surface deformation, air entrainment, and wave breaking. Using large-eddy simulation coupled with the volume-of-fluid method, we investigate these phenomena at Re=6×104 and spin ratio α=2 over a range of submergence depths H/D=3–100, Froude numbers Fr=2.0–3.2, and both clockwise and anticlockwise rotations. The free surface exhibits markedly different responses depending on the rotation direction: clockwise rotation produces a deep wave trough accompanied by intense air entrainment, whereas anticlockwise rotation generates a wave crest with substantially less entrainment. The two-phase mixing region volume is quantified and found to exhibit local extrema, reflecting the shifting balance between gravity- and inertia-dominated regimes for the parameter set investigated. The lift coefficient exhibits a significant reduction that is strongly correlated with the occurrence of intense air entrainment, with an optimal submergence of H/D=5 providing maximum lift under moderate Froude numbers. The streamwise velocity field reveals how air entrainment alters the wake structure and promotes energy dissipation. These findings provide a fundamental quantitative understanding of multiphase Magnus-effect flows.
