This study established an erosion weakening model for the strength degradation of filling media, the rationality and validity of which were verified via indoor particle loss tests. To study the engineering-scale disaster evolution, fluid-solid coupling numerical simulations were conducted, with the water inrush in the Paomashan No.2 Tunnel used as the engineering case study. Simulations reproduced the catastrophic permeability failure, further validating the accuracy of the proposed model. Orthogonal numerical simulations with fault width, dip angle, constant water pressure, and tunnel excavation height as variables revealed the evolution laws of mud-sand volume and inrush morphology. The study also included a classification of the severity levels permeability failure-induced inrush in fault fracture zones and importance raking of influencing factors. This study deepens our understanding of the evolution mechanism of permeability failure in fault-filling media and provides important technical support for the prevention and control of such engineering disasters.
This comprehensive study presents a computational fluid dynamics analysis to evaluate the performance of high-performance shortboard surfboards with and without fins under various operating conditions subjected to rotational motion about their centre of gravity simulating realistic interaction with the flow. Computational simulations were conducted across angles of attack ranging from 0 to 15 degrees, the resulting lift, drag, and lift-to-drag ratios were compared. Performance evaluation of the different scenarios indicates that the presence of fins consistently enhances lift generation and slightly increases drag, with an overall improvement in hydrodynamic efficiency up to a critical angle. The lift-to-drag ratio peaked between 7 degrees and 9 degrees, suggesting an optimal range for manoeuvrability and performance. This study provides a reproducible CFD methodology and offers new insights into the role of fins in surfboard dynamics. The study is performed on a full-scale surfboard, showing close agreement with the scaled model and prior studies, confirming the accuracy of the numerical analysis.
This study presents benchmark simulations of marine propeller flow using the Generalized k-omega (GEKO) turbulence model, designed to consolidate various two-equation models in industrial flow scenarios. Tunable parameters in GEKO facilitate the agreement of simulations with experimental data while maintaining model calibration. Through computational simulations on the INSEAN E779A propeller, we analyse the effects of GEKO parameters Csep, Cmix and Ccurv on flow characteristics. Results show that Csep mildly influences boundary layers due to rotational effects, promoting attached flow. Wake investigations indicate that the wake axial velocity deficit decreases and the turbulence intensity increases with increasing Cmix. Ccurv suppresses turbulence inside the vortex cores, extending vortex lifespans downstream. Overall, GEKO parameters exhibit mild effects on marine propeller flows with moderate turbulence. Default GEKO parameters yielded different flow fields from the SST model. Specific tuning tailored to the problem allows fine calibration of the model which is desirable in RANS scenarios.
Dynamic mode decomposition method is deployed to investigate the heat transfer mechanism in a compressible turbulent shear layer and shockwave. To this end, highly resolved Large Eddy Simulations are performed to explore the effect of wall thermal conditions on the behavior of a reattaching free shear layer interacting with an oblique shock in compressible turbulent flows. Various different wall temperature conditions, such as cold adiabatic and hot wall, are considered. Dynamic mode decomposition is used to isolate and study the structures generated by the shear layer exposed in the boundary layer. Results reveal that the shear layer flapping is the most energetic mode. The hot wall gains the highest amplitude for the flapping frequency, and the vortical motions are most intense in the vicinity of the reattachment point of the heated wall. The vortex shedding due to the large-scale motion of the shear layer is associated with the second energetic mode. The cold wall not only has a higher amplitude of the shedding mode, but it also has a lower frequency compared to the adiabatic and hot walls. This work sheds light on the underlying physics of the nonlinear intercoupling of momentum and heat, hence providing guidelines for designing control systems for high speed flight vehicles and mitigating aircraft fatigue loading caused by intense wall pressure fluctuations and heat flux.
A novel approach is proposed for defining the amplitudes of modes extracted by dynamic mode decomposition based on the amplitudes and coefficients derived from proper orthogonal decomposition. To verify this methodology, unsteady flow fields are numerically simulated using an in-house flow solver and subsequently decomposed. By employing the amplitudes defined through our proposed framework, the dominant mode derived from dynamic mode decomposition can be readily identified. The frequency of this dominant mode aligns closely with the dominant frequency observed in the aerodynamic force spectrum. Furthermore, the coherent structures associated with the dominant mode exhibit clear and well-organised spatial patterns. These results demonstrate that the dominant mode can be effectively distinguished from other modes extracted by dynamic mode decomposition using the amplitudes derived from the proposed definition. The introduced methodology provides researchers with a valuable tool to efficiently identify dominant coherent structures in unsteady flow fields and elucidate flow mechanisms.
This study presents a numerical investigation of blood flow through a porous thrombus located in a 6 mm diameter venous segment. The goal is to examine the effects of clot-induced obstruction under varying flow rates and rheological assumptions. The clot is modelled as a rigid porous medium, and the Brinkman equations are used to simulate flow through the thrombus region. Simulations are conducted in 3D using COMSOL Multiphysics, incorporating both Newtonian and non-Newtonian (Carreau model) descriptions of blood viscosity. Results highlight key post-thromb us hemodynamic features, including flow acceleration along the clot and persistent downstream recirculation zones. These altered flow regions are associated with increased risk of secondary thrombosis and embolic events. The comparison between viscosity models reveals differences in shear-dependent flow behaviour, including vortex persistence and asymmetric velocity distributions. Overall, this work improves understanding of clot-related flow disturbances and may support future numerical investigations aimed at improving thrombosis risk assessment.
This study proposes a stepwise shape optimisation method for a dragonfly-inspired corrugated wing by combining the adjoint variable method with deep reinforcement learning. The adjoint method provides a locally optimised shape using FreeFEM++, and the reinforcement learning, i.e. the Double Deep Q Network (DDQN), is used to obtain further improved solutions. Lift is defined as the objective function in the shape optimisation analysis based on the adjoint variable method, and a surrogate model, created using supervised learning, is applied to reduce the computational cost of the calculation of the lift in the DDQN. The results show that the lift force increases as a result of shape updating using the adjoint variable method and is further enhanced by DDQN. Taking the initial shape as 100%, the lift increases to 142.14% with the adjoint method and to 150.78% with DDQN, indicating an additional improvement of 8.64% over the adjoint-based result.
A scale-resolving methodology accounting for energy backscatter from subgrid to resolved scales is implemented within a second-order accurate compressible flow solver. The objective is to show the improved predictive capability that the artificial forcing associated with backscatter mechanisms can provide in low-order accurate CFD codes. The discretisation of the convective terms is designed to ensure favourable dissipation and dispersion properties while requiring minimal modifications to the baseline flow solver, which was originally developed for RANS simulations. To enhance numerical robustness, a blending strategy combining non-dissipative and strongly dissipative numerical schemes is adopted. The methodology is calibrated and validated using the canonical test case of decaying isotropic homogeneous turbulence. Its capability to mitigate the grey-area phenomenon is then assessed through the simulation of the mixing co-flow between a zero-incidence airfoil wake and a zero-pressure-gradient turbulent boundary layer. The results indicate that reliable scale-resolving simulations can be achieved with low-order discretisations when energy backscatter effects are properly accounted for.
This study introduces a computationally efficient, second-order precise algorithm for investigating thermal power-law fluids in stick-slip flows under various thermal boundary conditions. The study includes two main contributions. First, it introduces a novel numerical algorithm based on a modified artificial compressibility formulation that employs the Taylor-Galerkin finite element method. To demonstrate its efficacy, we compare this algorithm with an equivalently accurate scheme, the Taylor Galerkin/pressure correction algorithm, highlighting the superior convergence characteristics of the proposed algorithm, which requires nearly half the number of time steps. Second, investigation highlights the critical influence of disregarding the temperature dependence of viscous fluid on key parameters, such as viscous dissipation and heat transfer rate. Ignoring this dependence can lead to substantial underestimations of these parameters, thus affecting the simulation's accuracy. Additionally, we examine the influence of dimensionless parameters on average and local Nusselt numbers, finding remarkable agreement with previous findings.Highlights Proposed a high-accuracy algorithm for non-Newtonian thermal flows based on the extension of artificial compressibility equations.Combined of two-step Lax-Wendroff scheme and Galerkin finite element methods in the proposed algorithm.Compared the proposed algorithm with the Taylor-Galerkin pressure correction algorithm.Results demonstrate the accuracy and computational efficiency of the proposed algorithm at low Mach numbers.Investigated the consequences of ignoring the temperature dependence of a fluid's viscosity on pressure drop, viscous dissipation, and heat transfer under different thermal boundary conditions and stick-slip scenarios.
A comprehensive understanding of liquid metal flow dynamics is essential for advancing magnetohydrodynamic (MHD) devices. This study investigates the stability characteristics of MHD mixed convective flow in a vertical pipe under a transverse magnetic field. The fluid motion is driven by buoyancy forces and an imposed external pressure gradient. Key dimensionless parameters, including the Prandtl number (0.01 to 1) and Hartmann number (0 to 50), are examined to understand their influence on flow instability. The instability behavior is primarily governed by the Hartmann, Reynolds, Rayleigh, and Prandtl numbers. The results demonstrate that increasing the magnetic field strength enhances flow stability by amplifying the Lorentz force, which effectively suppresses disturbance growth. In contrast, a higher Prandtl number generally reduces the stability of the base flow, except for a slight stabilizing effect observed within a specific range of Pr. These findings provide critical insights into the stabilization mechanisms in MHD flows, which are crucial for the design of next-generation MHD systems. Under limiting conditions, the numerical results show good agreement with existing benchmark solutions, validating the accuracy of the present analysis. This work contributes to a deeper understanding of MHD flow behavior in confined magnetic fields.
In this study, an analysis of near-field noise reduction using the active jet method is proposed, and the underlying mechanisms of this noise reduction are explored. The results of the flow field analysis reveal that the boundary layer at the leading edge of the cavity and around the pantograph rod is stabilised following the application of the active jet, and the scale of the vortices is also significantly reduced. POD analysis indicates that the energy of the near-field noise in the pantograph area is reduced by up to 71.76% along the symmetry plane, with the degree of noise energy reduction increasing as the spatial height decreases. These findings demonstrate that the application of the active jet at the leading edge of the cavity is effective in reducing the near-field noise of the pantograph. This research provides a theoretical foundation for further studies on aerodynamic noise reduction in high-speed trains.
Accurate simulation of evaporation is vital for engineering applications, yet traditional methods face limitations. Molecular Dynamics (MD) offers high accuracy but is computationally intensive, while the S-model Boltzmann Kinetic Equation (SBKE) efficiently models rarefied gas flows but lacks a robust surface boundary condition. To overcome these challenges, we propose a novel hybrid MD-SBKE method for nano - and micro-scale evaporation. MD simulates the liquid phase and interface, while SBKE handles the vapour region, bridging microscopic and mesoscopic scales. This approach addresses the critical challenge of mass transfer between regimes during evaporation. Validated through numerical experiments, our one-dimensional hybrid model simulates argon nanoscale thin-film evaporation in parallel plate geometry. Results demonstrate improved accuracy over pure SBKE and significant computational efficiency gains compared to pure MD, enabling reliable evaporation modelling at reduced cost.
This study develops and validates an ultrasound-jet reactor that overcomes mixing and mass-transfer limits in multiphase systems. Hydrodynamic-acoustic cavitation, imposed by 20 kHz/15 & micro;m forcing at a Venturi throat, lowers the nucleation barrier and phase-locks bubble growth. Simulations and experiments show conversion of intermittent clouds into a stable, high-density plume (water-vapour fraction 21.17%; peak TKE 50.22 m(2)/s(2)) with a broadband bubble spectrum (0.001-0.064 mm; >60% at 1-8 & micro;m), enlarging interfacial area. Response-surface optimisation identifies 0.595 MPa inlet pressure, 44.387 degrees contraction angle, and 52.393 mm standoff as near-optimal. Carbonation tests report a reaction yield of 86.25% at 60 min versus 54.62% for a conventional jet, reducing time by >60%. The work establishes a process-structure-performance map enabling precise, scalable, and greener nano-ZnO synthesis.
Numerical computation in fluid mechanics is widely used in aerospace, marine engineering, electronics, and automotive aerodynamics. Traditional methods discretize the Navier-Stokes equations, but complex scenarios require exponentially more grids, increasing computational cost. Although deep learning has been applied to flow reconstruction and coarse-grid correction, few models mimic simulation workflows. This paper proposes DONTF (DeepONet-Transformer), which learns flow motion patterns and performs rolling predictions similar to conventional solvers, forecasting the next time step from current values. Numerical experiments confirm its reliability and adaptability under varying boundary conditions. Building on this, a hybrid solver - DONTF framework is developed to accelerate computations: DONTF rapidly predicts multiple time steps, after which the solver resumes calculation. This alternating, iterative process maintains physical consistency while greatly improving efficiency. The approach combines the accuracy of traditional solvers with the speed of deep learning, offering a robust solution for complex, time-dependent flow simulations.
The objective of this work is to develop an efficient numerical tool capable of modelling free-surface flows with transport of pollutants. The mathematical model that governs the phenomenon consists of the shallow water equations coupled to the advection-diffusion equation. The system of partial differential equations is solved by the semi-implicit Roe scheme, verifying the C-property. The numerical model was first validated on academic tests and then successfully applied to the Nador lagoon. We analyae the transport-diffusion of pollution emitted from the shore to the lagoon under boundary conditions reproducing the movement of high and low tides and under the effect of wind.
