arXiv subjects

Lihao Zhao

Publications and source records attributed to Lihao Zhao.

13 recordsLinked to original sources

arXiv2026

Multi-agent rendezvous in fluid flows via reinforcement learning

Rendezvous is a critical task for multi-agent systems, requiring agents to coordinate to meet at an unspecified location. However, achieving this in fluid environments presents a challenge, as it remains unclear how agents can exploit underlying fluid kinematics to facilitate convergence. In this study, we adopt a multi-agent reinforcement learning (MARL) approach to develop physics-informed rendezvous strategies in vortical flows. Compared to a naive strategy, where agents navigate toward their counterparts, MARL strategies significantly improve the rendezvous rate. MARL strategies also show transferability across varying vortex intensities, vortex scales, and swarm sizes. By breaking the symmetry of the state-action map, MARL strategy leverages a non-intuitive mechanism that prevents agents from becoming trapped in separate vortices, thereby enhancing rendezvous success. Additionally, a heuristic strategy is extracted from the learned strategy and also outperforms the naive strategy. Furthermore, a theoretical analysis demonstrates that fluid deformation impedes the rendezvous process. Large finite-time Lyapunov exponents identify where fluid effects separate adjacent agents, suggesting that targets should be planned in weak-deformation regions. Our findings reveal the important role that agent-fluid interactions play in multi-agent tasks and highlight the MARL capability to explore swarm intelligence in complex flow environments.

cs.MA↗

arXiv2026

Log-linear law of the mean streamwise velocity in turbulent boundary layers with moderate adverse pressure gradients

An essential feature of canonical zero-pressure-gradient (ZPG) turbulent boundary layers (TBLs) is that the mean streamwise velocity exhibits a logarithmic dependence on the wall-normal distance, known as the log law. In this study, we demonstrate that this conventional log law is not suitable for turbulent boundary layers subjected to pressure gradients (PGs). Instead, a log--linear law is theoretically derived for TBLs under moderate adverse pressure gradients (APGs), based on the total shear-stress balance and a rescaled eddy-viscosity model. The validity of the proposed log--linear law is assessed using available datasets of incompressible APG TBLs with the Clauser pressure-gradient parameter $\beta$ ranging from 0.73 to 9.0. Compared with the conventional log law, the present log--linear formulation shows significantly improved agreement with the measured mean velocity profiles. In the limiting case of $\beta \to 0$, the proposed law naturally recovers the classical log law.

physics.flu-dyn↗

arXiv2025

Algorithms for the Shortest Vector Problem in $2$-dimensional Lattices, Revisited

Efficiently solving the Shortest Vector Problem (SVP) in two-dimensional lattices holds practical significance in cryptography and computational geometry. While simpler than its high-dimensional counterpart, two-dimensional SVP motivates scalable solutions for high-dimensional lattices and benefits applications like sequence cipher cryptanalysis involving large integers. In this work, we first propose a novel definition of reduced bases and develop an efficient adaptive lattice reduction algorithm \textbf{CrossEuc} that strategically applies the Euclidean algorithm across dimensions. Building on this framework, we introduce \textbf{HVec}, a vectorized generalization of the Half-GCD algorithm originally defined for integers, which can efficiently halve the bit-length of two vectors and may have independent interest. By iteratively invoking \textbf{HVec}, our optimized algorithm \textbf{HVecSBP} achieves a reduced basis in $O(\log n M(n) )$ time for arbitrary input bases with bit-length $n$, where $M(n)$ denotes the cost of multiplying two $n$-bit integers. Compared to existing algorithms, our design is applicable to general forms of input lattices, eliminating the cost of pre-converting input bases to Hermite Normal Form (HNF). The comprehensive experimental results demonstrate that for the input lattice bases in HNF, the optimized algorithm \textbf{HVecSBP} achieves at least a $13.5\times$ efficiency improvement compared to existing methods. For general-form input lattice bases, converting them to HNF before applying \textbf{HVecSBP} offers only marginal advantages in extreme cases where the two basis vectors are nearly degenerate. However, as the linear dependency between input basis vectors decreases, directly employing \textbf{HVecSBP} yields increasingly significant efficiency gains, outperforming hybrid approaches that rely on prior \textbf{HNF} conversion.

cs.CG↗

arXiv2024

Interactions between the near-wall turbulent structures and heavy particles in compressible turbulent boundary layers

In the present study, we conduct direct numerical simulations to investigate the near-wall dynamics of compressible turbulent boundary layers at the free-stream Mach number of 6 laden with heavy particles. By inspecting the instantaneous near-wall flow structures, Reynolds stresses and the impacts of particle forces on solenoidal and dilatational motions, we observed that higher particle mass loadings lead to the less meandering yet almost equally intense velocity streaks, but the weakened wall-normal velocity fluctuations induced by vortices and near-wall dilatational motions organized as travelling wave packets. The strong correlation between the particle force and dilatational velocities indicates that particles are accelerated/decelerated while travelling through these travelling wave packets composed of expansive and compressive events, and in return, leading to the weakened dilatational motions of the fluid during this process. This correlation further supports the elucidation by Yu et al. (J. Fluid. Mech., vol. 984, 2024, pp. A44) that dilatational motions are generated by the vortices that induce strong bursting events, rather than the evolving perturbations beneath the velocity streaks. Nevertheless, the variation of skin friction in the presently considered cases with moderate mass loadings, either increased or decreased by the presence of particles, is found to be primarily attributed to the solenoidal Reynolds shear stress as in incompressible turbulence, suggesting the essentially unaltered nature of wall-bounded turbulence populated by vortical and shear motions instead of gradually switching to the state dominated by dilatational motions.

physics.flu-dyn↗

arXiv2024

Short term vs. long term: optimization of microswimmer navigation on different time horizons

We use reinforcement learning to find strategies that allow microswimmers in turbulence to avoid regions of large strain. This question is motivated by the hypothesis that swimming microorganisms tend to avoid such regions to minimise the risk of predation. We ask which local cues a microswimmer must measure to efficiently avoid such straining regions. We find that it can succeed without directional information, merely by measuring the magnitude of the local strain. However, the swimmer avoids straining regions more efficiently if it can measure the sign of local strain gradients. We compare our results with those of an earlier study [Mousavi {\em et al.} Phys. Rev. Res. {\bf 6}, L022034 (2024)] where a short-time expansion was used to find optimal strategies. We find that the short-time strategies work well in some cases but not in others. We derive a new theory that explains when the time-horizon matters for our optimisation problem, and when it does not. We find the strategy with best performance when the time-horizon coincides with the correlation time of the turbulent fluctuations. We also explain how the update frequency (the frequency at which the swimmer updates its strategy) affects the found strategies. We find that higher update frequencies yield better performance.

physics.flu-dyn↗

arXiv2024

Prolate spheroids settling in a quiescent fluid: clustering, microstructures and collisions

In this study, we investigate the sedimentation of prolate spheroids in a quiescent fluid by means of the particle-resolved direct numerical simulation. With the increase of the particle volume fraction $\phi$ from $0.1\%$ to $10\%$, we observe a non-monotonic variation of the mean settling velocity of particles, $\langle V_s \rangle$. By virtue of the Voronoi analysis, we find that the degree of particle clustering is highest when $\langle V_s \rangle$ reaches the local maximum at $\phi=1\%$. Under the swarm effect, clustered particles are found to preferentially sample downward fluid flows in the wake regions, leading to the enhancement of the settling speed. As for lower or higher volume fraction, the tendency of particle clustering and the preferential sampling of downward flows are attenuated. Hindrance effect becomes predominant when the volume fraction exceeds 5\% and reduces $\langle V_s \rangle$ to less than the isolated settling velocity. Particle orientation plays a minor role in the mean settling velocity, although individual prolate particles still tend to settle faster in suspensions when they deviate more from the broad-side-on alignment. Moreover, we also demonstrate that particles are prone to form column-like microstructures in dilute suspensions under the effect of wake-induced hydrodynamic attractions. The radial distribution function is higher at a lower volume fraction. As a result, the collision rate scaled by the particle number density decreases with the increasing volume fraction. By contrast, as another contribution to the particle collision rate, the relative radial velocity for nearby particles shows a minor degree of variation due to the lubrication effect.

physics.flu-dyn↗

arXiv2024

Transport of inertial spherical particles in compressible turbulent boundary layers

In the present study, we perform direct numerical simulations of compressible turbulent boundary layers at the free stream Mach number of 2 ~ 6 laden with dilute phase of spherical particles to investigate the Mach number effects on particle transport and dynamics. Most of the phenomena observed and well-recognized for inertia particles in incompressible wall-bounded turbulent flows, such as the near-wall preferential accumulation and clustering beneath the low-speed streaks, the flatter mean velocity profiles and the trend variation of the particle velocity fluctuations, are identified in the compressible turbulent boundary layer as well. However, we find that the compressibility effects are significant for large inertia particles. As the Mach number increases, the near-wall accumulation and the small-scale clustering are alleviated, which is probably caused by the variation of the fluid density and viscosity that are crucial to particle dynamics. This can be affected by the fact that the forces acting on the particles with the viscous Stokes number greater than 500 are modulated by the comparatively high particle Mach numbers in the near-wall region. This is also the reason for the abatement of the streamwise particle velocity fluctuation intensities with the Mach numbers.

physics.flu-dyn↗

arXiv2023

Efficient survival strategy for zooplankton in turbulence

Zooplankton in a quiescent environment can detect predators by hydrodynamic sensing, triggering powerful escape responses. Since turbulent strain tends to mask the hydrodynamic signal, the organisms should avoid such regions, but it is not known how they accomplish this. We found a simple, robust, and highly efficient strategy, that relies on measuring the sign of gradients of squared strain. Plankton following this strategy show very strong spatial clustering, and align against the local flow velocity, facilitating mate finding and feeding. The strategy has the potential to reconcile competing fitness pressures.

physics.flu-dyn↗

arXiv2022

Fluid inertial torque is an effective gyrotactic mechanism for settling elongated micro-swimmers

Marine plankton are usually modeled as settling elongated micro-swimmers. For the first time, we consider the torque induced by fluid inertia on such swimmers, and we discover that they spontaneously swim in the direction opposite to gravity. We analyze the equilibrium orientation of swimmers in quiescent fluid and the mean orientation in turbulent flows using direct numerical simulations. Similar to well-known gyrotaxis mechanisms, the effect of fluid inertial torque can be quantified by an effective reorientation time scale. We show that the orientation of swimmers strongly depends on the reorientation time scale, and swimmers exhibit strong preferential alignment in upward direction when the time scale is of the same order of Kolmogorov time scale. Our findings suggest that the fluid inertial torque is a new mechanism of gyrotaxis that stabilizes the upward orientation of micro-swimmers such as plankton.

physics.bio-ph↗

arXiv2021

Active gyrotactic stability of microswimmers using hydromechanical signals

Many plankton species undergo daily vertical migration to large depths in the turbulent ocean. To do this efficiently, the plankton can use a gyrotactic mechanism, aligning them with gravity to swim downwards, or against gravity to swim upwards. Many species show passive mechanisms for gyrotactic stability. For example, bottom-heavy plankton tend to align upwards. This is efficient for upward migration in quiescent flows, but it is often sensitive to turbulence which upsets the alignment. Here we suggest a simple, robust active mechanism for gyrotactic stability, which is only lightly affected by turbulence and allows alignment both along and against gravity. We use a model for a plankton that swims with a constant speed and can actively steer in response to hydrodynamic signals encountered in simulations of a turbulent flow. Using reinforcement learning, we identify the optimal steering strategy. By using its setae to sense its settling velocity transversal to its swimming direction, the swimmer can deduce information about the direction of gravity, allowing it to actively align upwards. The mechanism leads to a rate of upward migration in a turbulent flow that is of the same order as in quiescent flows, unless the turbulence is very vigorous. In contrast, passive swimmers with typical parameters of copepods show much smaller upward velocity in turbulence. Settling may even cause them to migrate downwards in vigorous turbulence.

physics.flu-dyn↗

arXiv2021

Turbulence modulations and drag reduction by inertialess spheroids in turbulent channel flow

Previous studies on nonspherical particle-fluid interaction were mostly confined to elongated fiber-like particles, which were observed to induce turbulence drag reduction. However, with the presence of tiny disk-like particles how wall turbulence is modulated and whether drag reduction occurs are still unknown. Motivated by those open questions, we performed two-way coupled direct numerical simulations of inertialess spheroids in turbulent channel flow by an Eulerian-Lagrangian approach. The additional stress accounts for the feedback from inertialess spheroids on the fluid phase. The results demonstrate that both rigid elongated fibers (prolate spheroids) and thin disks (oblate spheroids) can lead to significant turbulence modulations and drag reduction. However, the disk-induced drag reduction is less pronounced than that of rigid fibers with the same volume fraction. Typical features of drag-reduced flows by additives are observed in both flow statistics and turbulence coherent structures. Moreover, in contrast to one-way simulations, the two-way coupled results of spheroidal particles exhibit stronger preferential alignments and lower rotation rates. At the end we propose a drag reduction mechanism by inertialess spheroids and explain the different performance for drag reduction by fibers and disks. We find that the spheroidal particles weaken the quasistreamwise vortices through negative work and, therefore, the Reynolds shear stress is reduced. However, the mean shear stress generated by particles, which is shape-dependent, partly compensates for the reduction of Reynolds shear stress and thus affects the efficiency of drag reduction. The present study implies that tiny disk-like particles can be an alternative drag reduction agent in wall turbulence.

physics.flu-dyn↗

arXiv2020

Effects of the quiescent core in turbulent channel flow on transport and clustering of inertial particles

The existence of a quiescent core (QC) in the center of turbulent channel flows was demonstrated in recent experimental and numerical studies. The QC-region, which is characterized by relatively uniform velocity magnitude and weak turbulence levels, occupies about $40\%$ of the cross-section at Reynolds numbers $Re_τ$ ranging from $1000$ to $4000$. The influence of the QC region and its boundaries on transport and accumulation of inertial particles has never been investigated before. Here, we first demonstrate that a QC is unidentifiable at $Re_τ= 180$, before an in-depth exploration of particle-laden turbulent channel flow at $Re_τ= 600$ is performed. The inertial spheres exhibited a tendency to accumulate preferentially in high-speed regions within the QC, i.e. contrary to the well-known concentration in low-speed streaks in the near-wall region. The particle wall-normal distribution, quantified by means of Voronoï volumes and particle number concentrations, varied abruptly across the QC-boundary and vortical flow structures appeared as void areas due to the centrifugal mechanism. The QC-boundary, characterized by a localized strong shear layer, appeared as a \emph{barrier}, across which transport of inertial particles is hindered. Nevertheless, the statistics conditioned in QC-frame show that the mean velocity of particles outside of the QC was towards the core, whereas particles within the QC tended to migrate towards the wall. Such upward and downward particle motions are driven by similar motions of fluid parcels. The present results show that the QC exerts a substantial influence on transport and accumulation of inertial particles, which is of practical relevance in high-Reynolds number channel flow.

physics.flu-dyn↗

arXiv2020

Swimming strategy of settling elongated microswimmers by reinforcement learning

Particular types of plankton in aquatic ecosystems can coordinate their motion depending on the local flow environment to reach regions conducive to their growth or reproduction. Investigating their swimming strategies with regard to the local environment is important to obtain in-depth understanding of their behavior in the aquatic environment. In the present research, to examine an impact of the shape and gravity on a swimming strategy, plankton is considered as settling swimming particles of ellipsoidal shape. The Q-learning approach is adopted to obtain swimming strategies for smart particles with a goal of efficiently moving upwards in a two-dimensional steady flow. Strategies obtained from reinforcement learning are compared to those of naive gyrotactic particles that is modeled considering the behavior of realistic plankton. It is found that elongation of particles improves the performance of upward swimming by facilitating particles' resistance to the perturbation of vortex. In the case when the settling velocity is included, the strategy obtained by reinforcement learning has similar performance to that of the naive gyrotactic one, and they both align swimmers in upward direction. The similarity between the strategy obtained from machine learning and the biological gyrotactic strategy indicates the relationship between the aspherical shape and settling effect of realistic plankton and their gyrotactic feature.

physics.flu-dyn↗