The shear flow induced by a jetting cavitation bubble within grooved structures is systematically investigated using a three-dimensional compressible volume-of-fluid (VoF) solver combined with ultra-high-speed photography of one million frames per second (fps). The structured surface comprises twenty parallel micro-channels, and local mesh refinement down to $0.1\ {\unicode[Arial]{x03BC}}\text{m}$ near the groove bottom resolves the linear tangential velocity profile within the boundary layer. The simulations accurately reproduce experimental observations for two representative configurations, namely bubbles initiated above the groove and above the protrusion. In general, a microjet formed from the distal bubble surface impacts the lower bubble interface and subsequently spreads both along and across the groove channel, generating strong shear flow. Spatio-temporal wall-stress distributions indicate that the maximum shear stress occurs near the instant of jet impact, while the subsequent evolution is further modulated by daughter-bubble fragmentation and re-expansion. For a bubble initiated above the groove, the microjet produces a single impact, whereas for a bubble initiated above the protrusion, the jet becomes oblique: it first strikes the protrusion, spreads into the trapped bubble inside the groove and finally pierces the lower bubble surface at an angle. The influences of three geometric parameters, namely the horizontal offset $d^*$, vertical stand-off distance $h^*$ and groove width $w^*$, are systematically examined, revealing their coupled effects on bubble deformation, jet orientation, fragmentation and the vortex dynamics. A peak wall-shear stress of approximately $150\,\text{kPa}$ is obtained under specific geometric conditions due to rapid jet spreading and flow refocusing within the groove. These results provide fundamental insight into the three-dimensional, unsteady wall-shear generation produced by jetting cavitation bubbles interacting with micro-structured surfaces.

Cavitation bubbles collapsing near capillary exits can generate fast microjets that penetrate into the channel and promote fluid exchange. Here, we combine high-speed imaging and axisymmetric compressible volume-of-fluid simulations to study laser-induced cavitation bubbles near the exit of a long capillary. By varying the stand-off distance $\gamma$, the normalised inner radius $R_i^*$ and the normalised outer radius $R_o^*$, we identify two distinct jetting mechanisms. Regular jets are driven primarily by the vertical collapse of the bubble surface distal to the capillary, whereas wave-driven jets arise from the propagation and convergence of an interfacial wave initiated at the proximal bubble surface, causing distal-surface splitting and strong local pressure amplification. Wave-driven jets occur for sufficiently small stand-off distances and outer radii, with formation velocities typically above $500\,\mathrm{m\,s}^{-1}$ and reaching up to $2000\,\mathrm{m\,s}^{-1}$ in the simulations. Regular jets dominate at larger stand-off distances or larger outer radii, where interfacial-wave effects weaken. The outer radius $R_o^*$ governs the jetting regime by controlling interfacial-wave formation, propagation and convergence, while the inner radius $R_i^*$ mainly affects the later jet dynamics, including the impact location and vapour-jet penetration into the capillary. Compared with the needle jet generated by a bubble collapsing near a smooth rigid wall at nearly zero stand-off distance, the wave-driven jet here occurs over a much wider parameter range and its velocity and size can be tuned geometrically. These results establish a physical framework for cavitation-driven jetting near capillary boundaries and suggest strategies for controlling confined microjets in microfluidics and biomedical applications.

A compact and low-cost multiple-input multiple-output (MIMO) antenna system is proposed to meet the demands of modern portable wireless communication devices. The antenna is manufactured on an FR4 substrate measuring 10 × 60 × 0.8 mm3 and operates efficiently in the LTE, Sub-6 GHz, and Wi-Fi 6E/7 bands. In particular, the system achieves a high peak gain of 9.32 dBi and a maximum radiation efficiency of 84%, covering operational frequency bands of 920–1040, 1810–2450, 3480–3780, and 5370–9000 MHz. The antenna structure does not rely on any passive matching components, yet it demonstrates excellent impedance matching, wide bandwidth, and strong inter-element isolation. To ensure the stable generation of multiple resonances, L-shaped grounding elements are integrated on both the top and bottom surfaces of the substrate. These grounding structures not only support the formation of multi-resonant modes but also contribute significantly to the suppression of mutual coupling, thereby enhancing MIMO performance reliability. To validate practical usability, the antenna system is integrated into two common platforms: a laptop and a smartphone. Both simulation and experimental results confirm that the proposed system offers robust wireless performance while satisfying the requirements for human-exposure-compliant design, making it a promising solution for high-band portable applications.

This study combines linear stability analysis (LSA) and direct numerical simulations (DNS) to investigate the coupling mechanisms of rotation, stratification and shear in two flow geometries: spanwise rotating stratified plane Couette flow (between two infinite parallel plates) and sheared annular centrifugal Rayleigh–Bénard convection (between two concentric cylinders). First, LSA reveals a universal scaling law with respect to $(1-\eta )$, where $\eta$ denotes the radius ratio, for the deviation between annular and planar system results, quantitatively establishing the dynamical connection between annular and planar geometries. Building on this, DNS results indicate that the breakdown of the twin-force analogy in the turbulent regime originates from mode competition: buoyancy tends to drive three-dimensional chaotic plumes via baroclinic effects, whereas the Coriolis force tends to sustain quasi-two-dimensional longitudinal vortices. The Prandtl number determines the dominant mechanism in the near-wall region by regulating the relative thicknesses of the thermal and momentum boundary layers, thereby leading to a systematic dissimilarity between heat and momentum transport. Furthermore, this study provides a new theoretical basis for transport prediction in complex flows from the perspectives of enstrophy and mode competition.

In this work we propose a physics-constrained machine learning approach to enhance the generalisability of neural-network-represented turbulence models across different flow regimes. Machine learning-augmented turbulence modelling has emerged to improve the predictive accuracy of the Reynolds-averaged Navier–Stokes method based on high-fidelity data. However, such data-driven models face difficulties in generalising well in different flow scenarios outside training datasets. Here we show that this issue can be alleviated by incorporating physical constraints of the Reynolds stress during the learning process. It is achieved with a physics-constrained learning approach based on the regularised ensemble Kalman method, which allows for model learning from limited velocity data and is flexible in imposing physical constraints. The method is used to learn a nonlinear eddy viscosity model, represented by a tensor-basis neural network, from velocity data in separated flows over periodic hills. Two physical constraints are incorporated during the model learning, i.e. the baseline coefficients to capture the log law of the wall and the algebraic Reynolds stress equation to inform the Reynolds stress anisotropy. Moreover, an adaptive penalty method is developed to enforce physical constraints in the feature space associated with canonical flows, avoiding the incompatibility with the training data from separated flows. Our results show that the proposed method can learn a generalisable turbulence model across different flow scenarios, including similar separated flows, canonical wall-bounded flows and complex flow configurations. Finally, the learned neural-network-based model function is discussed, demonstrating its consistency with the imposed physical constraints.

Prospective evidence linking overall diet quality to gastroesophageal reflux disease (GERD) is limited, and metabolic pathways remain unclear. In a cohort of 183 878 adults free of GERD at baseline, we examined nine dietary pattern scores derived from repeated 24-h dietary recalls and assessed associations with incident GERD over a median follow-up of 13·2 years, during which 15 401 cases occurred. Higher adherence to healthy patterns including Mediterranean diet, Mediterranean-Dietary Approaches to Stop Hypertension Diet Intervention for Neurodegenerative Delay, Dietary Approaches to Stop Hypertension and dietary index for gut microbiota was associated with lower GERD risk (highest v. lowest tertile hazard ratios 0·87–0·89), whereas a pro-inflammatory pattern (Energy-adjusted Diet Inflammatory Index) was linked to higher risk (hazard ratio 1·11). In a metabolomics sub-cohort (N 98 669), diet-related NMR metabolomic signatures, derived from 168 biomarkers using Elastic Net regression, reflected similar associations and mediated 11·1–32·2 % of the diet–GERD relationship. Associations were largely consistent across subgroups and robust in sensitivity analyses, supporting a potential role of systemic metabolic profiles in mediating the effects of diet on GERD and highlighting metabolic mechanisms underlying diet-related disease risk.

Boundary slip can strongly influence the wall shear stress generated by a collapsing cavitation bubble near a solid surface. Here, we investigate this effect by systematically varying slip length $l$ and bubble stand-off distance $\gamma$. We show that boundary slip has negligible influence on bubble dynamics and jet formation prior to jet impact, but markedly alters the near-wall flow during jet spreading. Increasing $l$ substantially reduces both outward and inward peak shear stresses, suppresses flow separation during fast jet spreading and reorganises the spatio-temporal stress distribution. We derive a scaling for the maximum outward stress $\tau _m$ with $\tau _m\approx {\mu {U}_{0}{Re}^{0.5}(U_e/{U}_{0})^{1.5}}/ ({0.3R_{\textit{max}}+l(U_e {Re}/{U}_{0})^{0.5}})$, which captures the coupled effects of slip length, maximum bubble radius $R_{\textit{max}}$ and the outer-edge velocity $U_e$ of the spreading wall jet. Combined with a scaling for $U_e$ in terms of $\gamma$ and $l$, this yields a semi-empirical predictive framework for the maximum wall shear stress. These results clarify how interfacial slip regulates extreme cavitation-driven shear and offer a basis for its prediction and control.

Current flying-car designs lack scalability for diverse missions. This paper presents a modular design platform for developing reconfigurable flying-cars, embedding modularity across structural, electrical, and flight control domains. A full-scale sightseeing prototype demonstrates the platform’s feasibility and flexibility. The work contributes to design methodology by illustrating how modular architectures improve cross-mission adaptability, scalability, and lifecycle efficiency in complex mechatronic systems. (project introduction video available at https://www.aidilab.ai/flying-car)

In this study, we develop epidemic reaction-diffusion models by incorporating the dependency of the diffusion rate of susceptible individuals on new infection cases, employing both Fickian and Fokker–Planck-type diffusion laws. As the first part of a two-part series, we focus on epidemics driven by frequency-dependent incidence. We explore linear, exponential and algebraic relationships between diffusion rate of the susceptible population and new infection cases to provide deeper biological insights. Our analysis establishes the global existence of solutions and characterizes the threshold dynamics using basic reproduction numbers. We find that in quasilinear parabolic systems, the Fokker–Planck-type diffusion law tends to induce spatial segregation of susceptible and infected individuals, while the Fickian law favours spatial homogenization of susceptible individuals. Additionally, the Fokker–Planck-type model, where the diffusion rate of infected individuals depends on new recovery cases, more accurately captures the cognitive diffusion behaviour of individuals.

We propose a novel flow-control strategy to suppress low-frequency unsteadiness in shock wave–turbulent boundary layer interactions. Our analysis reveals that the breathing motion of the separation bubble is governed by the leading two-dimensional global mode through a modal resonance mechanism, with its growth rate primarily concentrated in a narrow region near the shock foot. By introducing a micro-bump at this critical location, we achieve targeted stabilisation of the global mode, resulting in a substantial attenuation of the low-frequency breathing motion. Consequently, the turbulent kinetic energy and wall-pressure fluctuations associated with the breathing motion are reduced by 85 % and 70 %, respectively. This study demonstrates an effective and precise approach for controlling low-frequency unsteadiness in shock-induced separation, while minimally disturbing the flow.

In extreme environments, hexapod robots are highly susceptible to locomotor failure due to leg structure damage or electronic circuit interruptions. Such failures are particularly critical in ipsilateral two-leg loss, where most conventional hexapod designs fail to maintain stability. This paper introduces a hexapod robot platform equipped with a functional arm (FA) and proposes a fault-tolerant strategy based on FA compensation. The FA serves as a task-oriented module under normal conditions and as a compensatory leg during instability caused by leg loss, enabling rapid recovery of stability and walking capacity without complex adjustments to the remaining legs. An optimization algorithm was developed for gait adjustment, using the longitudinal stability margin as the optimization objective with the compensation median angle θ and swing angle α as key parameters. The optimal compensatory positions of the FA were identified for both the ipsilateral middle-hind leg and front-hind leg failure scenarios. Experimental results demonstrate a marked improvement in both walking stability and capability after FA compensation. For both failure modes, the robot’s attitude fluctuations were significantly suppressed, and the stride length was substantially increased, affirming the feasibility and effectiveness of the proposed approach in enhancing the environmental adaptability of multi-legged robots.

Neuropsychiatric disorders (NPDs) are a leading cause of disability worldwide. The predominantly plant-based EAT–Lancet diet has been proposed to confer neuropsychiatric benefits, yet evidence remains limited. This study synthesized associations between adherence to the EAT–Lancet diet and neuropsychiatric outcomes. We searched PubMed, Web of Science, Embase, Scopus, and ProQuest Dissertations and Theses Global through September 4, 2025. Observational studies reporting associations between EAT–Lancet adherence and NPDs were included. Binary outcomes were pooled as hazard ratios (HRs) or odds ratios (ORs), and continuous outcomes as regression coefficients (β). Subgroup, sensitivity, and publication-bias analyses were performed. Certainty of evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation framework (GRADE). Twenty-two cohort and six cross-sectional studies were included. Higher adherence to the EAT–Lancet diet was associated with lower risks of depression (OR 0.76; 95% CI 0.71–0.81), anxiety (OR 0.82; 0.76–0.89), stroke (HR 0.84; 0.76–0.92), and dementia (HR 0.96; 0.93–1.00), whereas no significant association was observed for global cognitive function (β 0.02; −0.01 to 0.06). Sensitivity analyses supported robustness. Certainty of evidence was very low for anxiety, depression, and cognition, and low for stroke and dementia. Greater adherence to the EAT–Lancet diet was associated with lower risks of depression, anxiety, stroke, and dementia. However, given the low certainty of evidence, findings should be interpreted cautiously. Further large prospective studies and randomized controlled trials are warranted to improve evidence quality and clarify the potential role of the EAT–Lancet diet in neuropsychiatric disease prevention.

We show a high-slope-efficiency Ho:CaYAlO4 laser using the conventional water-cooling method. Under an L-shaped laser resonator cavity with an overlapping efficiency greater than 98%, a maximum output power of 18.64 W at 2081 nm was achieved when the absorbed pump power was 26.83 W. By linear fitting, the slope efficiency relative to the absorbed pump power is 90.6%. Under the maximum output power, the beam quality factor is lower than 1.25 and the root-mean-square stability is about 0.42%. In addition, we validated the output performances of the laser in acousto-optic Q-switched mode under the high-slope-efficiency condition. At a pulse repetition rate of 10 kHz, the maximum average output power of 18.26 W and the narrowest pulse width of 40.1 ns were achieved, corresponding to a pulse energy, peak power and slope efficiency of 1.83 mJ, 45.54 kW and 87.7%, respectively. To the best of our knowledge, this laser may have the highest slope efficiency for a Ho-doped solid-state laser using conventional water-cooling.

The study of matter under extreme conditions, particularly non-repeatable ultrafast processes such as those in high-energy-density physics and inertial confinement fusion, urgently demands innovations in ultrafast imaging diagnostics. To address this, we propose a novel multi-pulse, multi-angle imaging scheme that leverages a beam-splitting design applied to a single ultra-intense laser pulse to generate multiple, precisely delayed sub-pulse beams. This method not only achieves a tunable temporal resolution range from femtoseconds to picoseconds and spatial resolution in the micrometer scale, providing a powerful tool for capturing single-shot ultrafast dynamics, but also overcomes the low repetition rate limitation in petawatt-class laser facilities, enabling previously unattainable continuous dynamic imaging. Experimental validation is scheduled on the 1 PW/0.1 Hz laser system at the Shanghai Superintense Ultrafast Laser Facility. Feasibility of the source is verified by FBPIC simulations and combined with SynchRad for radiation calculations, while Geant4 simulations demonstrate imaging performance.

Continuous-wave diamond Raman lasers (DRLs) operated at high pump power commonly suffer from power roll-off and output instability, even when high-thermal-conductivity gain media are employed. These degradations originate from thermally induced cavity-mode mismatch and cavity-length drift, which limit further power scaling. In this work, we establish a thermo–optical chain-linked model that quantitatively describes the coupled evolution of thermal deposition, equivalent thermal lensing, intracavity-mode matching and output power in high-power DRLs, providing a direct physical link between pump power and cavity-length compensation. Based on this model, a peak–valley co-location criterion is proposed to determine the optimal cavity-length compensation point. Both theoretical and experimental results show that the output power exhibits a single-peak dependence on cavity-length offset, while the power stability follows a U-shaped distribution, and their extrema coincide at an optimal compensation value ΔL*. By operating at ΔL*, the maximum Stokes output power increases from 27 to 32 W, and the root mean square power fluctuation decreases from 4.8% to 3.2%. These results demonstrate that thermal-lens-induced performance degradation can be effectively mitigated by cavity-length compensation using a single control parameter, providing a practical design rule for simultaneous enhancement of output power and stability in high-power Raman laser systems.

The hypersonic flow over 30$^{\circ }$–50$^{\circ }$ double-cone configurations with three nose bluntness levels was experimentally investigated at Mach 6. High-speed schlieren photography, pressure sensors and pressure-sensitive paint were used to examine both global flow patterns and unsteady dynamics at a transitional Reynolds number. The experimental results indicate that the size of the separation region at the cone junction increases with increasing nose bluntness. Type V shock–shock interactions were observed in all three configurations, while the shock wave structures in the region below the triple point exhibited two patterns: Mach shock wave reflection in the sharp and small-blunt-nose cases, and regular shock wave reflection in the large-blunt-nose case. Spectral analysis of high-speed schlieren sequences revealed two types of unsteadiness across all cases: low-frequency shock oscillations and high-frequency unsteady structures along the boundary of supersonic jet on the second cone. For the low-frequency unsteadiness, shock oscillations displayed a broadband nature in the sharp and small-blunt-nose configurations, while a dominant frequency of approximately 2 kHz was observed in the large-blunt-nose case, characterised by shock motion and bubble breathing – an observation not experimentally reported before. Additionally, spectral analysis of wall pressure contours indicated that the low-frequency unsteadiness was primarily characterised by axisymmetric modes for all configurations. Global stability analysis and resolvent analysis further demonstrated noise-amplifier behaviour in all configurations, and the dominant low-frequency unsteadiness in the large-blunt-nose case is attributed to modal resonance induced by environmental noise.

An experimental study was conducted to investigate the characteristics of unsteady oblique shock trains in a constant-area rectangular duct under an asymmetric incoming boundary layer. High-speed schlieren techniques and high-frequency pressure measurements were utilised in this study. The results indicate that the oblique shock train mainstream leans significantly towards the thin-boundary-layer side. Under downstream periodic excitation, the shock train moves periodically, and its shape changes during the movement. This phenomenon occurs to match the downstream pressure by altering the relative Mach number in front of the shock train, with the average pressure rise slope along the thick-boundary-layer side changing periodically. Additionally, unlike a normal shock train, the pressure rise distribution along the thick-boundary-layer side is nearly linear, and the correlation coefficients between the transducers on this side and the most downstream transducer are higher than those on the thin-boundary-layer side. Due to differences in flow structure and pressure rise distribution, the existing amplitude prediction model proposed by Xiong et al. (J. Fluid Mech. vol. 846, 2018, pp. 240–262) for the unsteady normal shock train is no longer applicable to the unsteady oblique shock train. Therefore, a new prediction model is derived and verified by experiments. Moreover, it is found that using only the downstream pressure transducer information on the thick-boundary-layer side can effectively predict the amplitude of the shock-train motion. Combined with the prediction model, a novel method is proposed to estimate the amplitude of the shock-train motion conveniently.

We examine how Confucian culture operates as an informal institution by fostering relational contracts that substitute for formal legal frameworks in shaping corporate behavior. Using data on historical Confucian academies near firms’ headquarters in China, we find that greater cultural exposure is associated with higher investment in stakeholder relationships—measured by social contribution, stakeholder protection, courtesy expenses, patenting, and trade credit. These effects persist after controlling for human capital and alternative cultural influences, and weaken in regions with stronger formal institutions. Our findings highlight the enduring role of culture in supporting trust-based governance when formal contracting is limited.