Moving Particle Semi-implicit Method Research Articles

Comparison of MPS and SPH methods for solving forced motion ship flooding problems

In order to assess the safety of damaged ships quantitatively, advanced numerical simulation methods are required for the prediction of ship dynamic behavior coupled with floodwater. For the prediction of fast and complicated free-surface flows in immediate flooding, particle methods are promising because largely deformed free-surface can be naturally and accurately followed without numerical diffusion. In this paper, systematic comparisons of the MPS (Moving Particle Semi-implicit) and SPH (Smoothed Particle Hydrodynamics) methods are made on ship flooding problems. Firstly, numerical simulations based on the MPS and SPH methods are performed on forced roll tests of a two-dimensional damaged compartment and are compared with a dedicated model experiment. Secondly, similar comparisons are made for a three-dimensional flooded compartment. Through these comparisons between MPS and SPH methods and to model-scale experiments, the capability of particle methods in simulating complicated flooding flows is investigated and the difference of prediction accuracy of the MPS and SPH methods is discussed. For the case of two-dimensional damaged compartment, it is also studied what effects has a single-phase approximation in comparison to simulations where air is modelled, and how the trapped air in damaged compartments influences the water flow evolution. Results from this study show that both particle methods are capable to simulate complex flooding flows in good quantitative agreement with respect to experiments and with similar execution times while there is no possibility to draw superiority of one method. However when dealing with air modeling, using a simplified approach limits computational times but induces quite large discrepancies compared to fully modeling the air phase.

Application of a Moving Particle Semi-Implicit Numerical Wave Flume (MPS-NWF) to model design waves

We model regular and irregular design waves of interest in coastal and ocean engineering, using a Numerical Wave Flume (NWF) based on a Moving Particle Semi-Implicit (MPS) method. A key feature of the model is the implementation of a new scheme for the artificial viscosity, which is original to the Smoothed Particle Hydrodynamics (SPH) method and is here formalised into the MPS context. In addition, we complement the NWF with a wave generator based on the dynamic boundary method approach, and a wave absorption mechanism using a numerical damping zone. We show that accurately calibrating the new terms has a strong effect in preventing particle clustering and tensile instability, and allows a precise simulation of wave phenomena of interest in ocean and coastal engineering. The NWF-MPS model is validated using the available analytical and experimental benchmarks, and is then applied to reproduce a violent wave impact on a vertical wall breakwater. Interestingly, the novel NWF-MPS requires a much lower number of particles to obtain a similar level of accuracy as compared to analogous SPH simulations run on similar machines.

On the free surface boundary of moving particle semi-implicit method for thermocapillary flow

The moving particle semi-implicit (MPS) method has great potential in dealing with free surface flow due to its Lagrangian nature. In most cases, the free surface boundary is simply served as the pressure boundary condition. In this paper, an improved MPS method is presented for thermocapillary driven free surface flow. A series of surface nodes explicitly represent the free surface boundary. The normal stress on the free surface provides the Dirichlet pressure boundary condition, while the velocity boundary condition, i.e., Marangoni stress, is enforced through the Taylor series expansion and least squares method. Meanwhile, a quasi-Lagrangian formulation is introduced to avoid particle clustering and the corresponding numerical instability by slightly modifying the advection velocity. The upwind scheme is employed for the convection term to obtain accurate and stable results. A novel constraint scheme with the divergence of provisional velocity is developed for the pressure gradient to enhance stability further. The consistency of the derived generalized boundary condition is firstly verified with a simple convergence test. Then, several numerical tests, including square patch rotation, lid-driven and square droplet oscillation, are simulated to show the improvements. Finally, thermocapillary driven flows in an open cavity without and with buoyancy effect are studied. Good agreements are obtained by comparing with reference simulations taken from literature. Heat transfer characteristics are further investigated for different dimensionless numbers, including the Rayleigh number and Marangoni number.<PE-FRONTEND>

Investigation of Numerical Conditions of Moving Particle Semi-implicit for Two-Dimensional Wedge Slamming

The sensitivity of moving particle semi-implicit (MPS) simulations to numerical parameters is investigated in this study. Although the verification and validation (V&V) are important to ensure accurate numerical results, the MPS has poor performance in convergences with a time step size. Therefore, users of the MPS need to tune numerical parameters to fit results into benchmarks. However, such tuning parameters are not always valid for other simulations. We propose a practical numerical condition for the MPS simulation of a two-dimensional wedge slamming problem (i.e., an MPS-slamming condition). The MPS-slamming condition is represented by an MPS-slamming number, which provides the optimum time step size once the MPS-slamming number, slamming velocity, deadrise angle of the wedge, and particle size are decided. The simulation study shows that the MPS results can be characterized by the proposed MPS-slamming condition, and the use of the same MPS-slamming number provides a similar flow.

Experimental and numerical investigations of dam break flow over dry and wet beds

Within the present work, the dam break flow over dry and wet beds are investigated experimentally, aiming to provide up-to-date data set and deeper insight into the morphology and hydrodynamics of the dam-break flood wave. The experiments are performed in a smooth prismatic channel with a rectangular cross-section. The wave profile and the downstream impact pressures are captured by a high-speed camera and several pressure sensors, respectively. The water depth is measured at five specified downstream locations through imaging processing. The laboratory observations are presented in terms of non-dimensional water level height, dynamic pressure variations on the impact wall, free-surface evolution, wave-front propagation, and high-resolution videos. Furthermore, two different numerical methods namely: (1) Volume-Of-Fluid (VOF) and (2) Moving Particle Semi Implicit (MPS) are also utilized to verify the repeatability of the experimental measurements. The comparisons show a satisfactory agreement between both models and the experiments. However, the results reveal that the Lagrangian model (MPS) slightly outperformed the standard VOF model in tracking highly deforming interfaces with severe topology changes. The dataset generated by small-scale laboratory investigations in the current work can provide a full panoramic view of dam-break flow which may be used as benchmark for validation of various CFD tools or understand the complex physics of dam-break wave and similar phenomena.

Numerical approach to pipe flow of fresh concrete based on MPS method

As a fundamental study of concrete pumping, this study developed a numerical method to simulate the flow and segregation of fresh concrete in pipes, based on the improved MPS (Moving Particle Semi-implicit) that has complete implicit algorithm, hereafter called I-MPS. The slip layer (LL) near the pipe inner wall was treated by the macroscopic approach, which estimates the slip resistance and the volumetric flow rate of LL from the apparent slip velocity of LL. Two constituent models were used to describe fresh concrete, called single-phase & mono-particle (SPMP) model and double-phase & multi-particle (DPMP) model, respectively. In the former, fresh concrete is considered as a single-phase granular fluid, but in the latter, fresh concrete is regarded as two-phase granular fluid with different particle shapes and sizes of coarse aggregate and matrix mortar. By comparing the numerical and theoretical pumping pressures of three concretes having slump from 13 cm to 21 cm, it was found that by using the macroscopic approach of LL and either of the constituent models, the numerical approach can predict properly the pumping pressure and velocity profile of fresh concrete in pipes. Moreover, if using the DPMP model, the segregation behavior of coarse aggregate particles during the pipe flow can also be simulated.

An improved moving particle semi-implicit method for interfacial flows

In this paper, the improved moving particle semi-implicit (IMPS) method is further developed into a multiphase method by introducing various multiphase models. Then, the developed multiphase MPS method is applied to a variety of 2-D/3-D simulations of interfacial flows, including Rayleigh-Taylor instability, bubble rising, dam-break flow, and internal solitary waves. The fair agreements between the results of MPS and other reference results demonstrate that the present multiphase MPS method is stable and accurate enough to capture the interface with a large deformation in different flow problems. The phenomenon of air entrapment induced by the dam-break flow is well reproduced in the multiphase simulation and its influence on impact pressure is discussed in detail. The internal solitary waves with different amplitudes are numerically generated and the wave heights show a good agreement with experimental data even after a long-distance propagation, indicating that the numerical diffusion of the present multiphase MPS method is relatively low and the energy conservation can be ensured.

A phantom and in vivo simulation of coronary flow to calculate fractional flow reserve using a mesh-free model.

Moving particle semi-implicit (MPS) method is a mesh-free method to perform computational fluid dynamics (CFD). The purpose of this study was to calculate the simulated fractional flow reserve (sFFR) using a coronary stenosis model, and to validate the MPS-derived sFFR against invasive FFR using clinical coronary CT data. Coronary flow simulation included 21 stenosis models with stenosis ranging 30-70%. Patient coronary analysis was performed in 76 consecutive patients (100 vessels) who underwent coronary CT angiography and subsequent invasive FFR between November 2016 and March 2020. Accuracy of sFFR and CT angiography for diagnosis of invasive FFR ≤ 0.80 was compared. Quantitative morphological stenosis data of CT angiography were also obtained. Area stenosis showed a good correlation to sFFR (R2 = 0.996, p < 0.001) in coronary stenosis models. In the patient study, the mean FFR value was 0.82 ± 0.10, and 37 out of 100 vessels showed FFR ≤ 0.80. FFR and sFFR values showed a good correlation (R2 = 0.59, p < 0.001) with a slight underestimation of sFFR as compared with FFR (mean difference - 0.015 ± 0.096, p = 0.12). The sensitivity, specificity, positive predictive value, and negative predictive value of sFFR to predict FFR ≤ 0.80 was 86%, 89%, 82%, 92%, respectively. The accuracy to predict FFR ≤ 0.80 using sFFR was greater than using diameter stenosis and minimum lumen area (88% vs. 74%, p = 0.008). CFD using the MPS method showed feasible results validated against invasive FFR. The accuracy to predict significant stenosis was higher than morphological stenosis.

Investigation on the effects of vertical baffles on liquid sloshing based on a particle method

This study adopts the numerical simulations of Moving Particle Semi-Implicit Methods (MPS), which are meshless methods based on Lagrange particles. Using Lagrange particle has an advantage that it can avoid numerical dissipation problems without directly discretizing the convection term in the governing equation. First of all, a numerical model of a liquid sloshing tank without baffles is used to confirm the effectiveness of the MPS by comparing the numerical results with the experimental data of Kang and Li. And the pressure curves obtained with MPS method were in good agreement with the experimental findings, which confirmed its effectiveness. On that basis, simulations of liquid sloshing movements with one baffle, two symmetrical baffles, and three baffles are performed, respectively. The results indicate that the addition of vertical baffles in the tanks effectively enhanced the ability to reduce liquid sloshing.

Development of a Numerical Model for Metal Vapor Dynamics in Low Pressure Arcs based on Moving Particle Semi-implicit Method and Influence of Shield Wall Radius on Vapor Behavior

Development of a Numerical Model for Metal Vapor Dynamics in Low Pressure Arcs based on Moving Particle Semi-implicit Method and Influence of Shield Wall Radius on Vapor Behavior

Numerical simulation for coastal morphodynamics using DEM-MPS method

Numerical simulations were performed for morphodynamics with free water surface fluctuation using the DEM-MPS method, which employs the discrete element method (DEM) and the enhanced moving particle semi-implicit (MPS) method for the solid and liquid phases, respectively. The reproducibility of the DEM-MPS method to a swash ripple formation process of a mixed grain bed is demonstrated. There, the bed is fully covered with water, and its void is filled with pore water. It is suggested that the ripple formation with coarse-grained particles activates the momentum exchange in the surface layer of the movable bed. In addition, a numerical simulation for a solitary wave induced by a landslide by DEM solid particles was carried out. Here, the applicability of the model in the situation where the pore water surface exists in a moving and deforming solid-sediment layer is examined. The simulation results show that it is essential to apply the apparent volume to a fluid particle inside a solid-sediment layer to predict the water surface evolution accurately.

A Simulation of Soil Dumping Using Moving Particle Semi-Implicit Method

Kim, K.S. and Lee, J.H., 2021. A simulation of soil dumping using moving particle semi-implicit method. In: Lee, J.L.; Suh, K.-S.; Lee, B.; Shin, S., and Lee, J. (eds.), Crisis and Integrated Management for Coastal and Marine Safety. Journal of Coastal Research, Special Issue No. 114, pp. 549–553. Coconut Creek (Florida), ISSN 0749-0208. Soil dumping is one of most important problems on the coastal and ocean construction project. In general, the computational fluid dynamics method used grid system to simulate soil particle under flow with concentration. It can be useful for fast calculation; however, it is restricted to demonstrate its movement. The Moving Particle Semi-implicit (MPS) method follows fully Lagrangian approach, thus it can simulate soil particle behavior due to tracking each particle. In this study, the newly developed MPS method modified for solid particle was used to simulate soil dumping. For the solid particle interaction, drag force and friction force term were implemented instead of viscosity term in the governing equation. The friction force is due to friction between solid particles, and the drag force can be calculated by relative velocity around center particle within effective range. All the solid particle interaction models were validated by corresponding physical phenomena. The numerical results were compared to the corresponding experimental results and they show good agreement in the tendency of behavior.

Development of adaptive multi-resolution MPS method for multiphase flow simulation

In this study, an adaptive multi-resolution moving particle semi-implicit (MPS) method is developed aiming for efficient and accurate multiphase flow simulation. The least-square MPS(LSMPS) scheme is adopted for high-order accurate operator discretization. An adaptive multi-resolution framework is developed to realize adaptive particle splitting and merging near the interface of multiphase flow. LSMPS also provides high-order interpolation of physical quantities such as velocity and pressure during particle splitting and merging. Accuracy and efficiency of the proposed method is demonstrated through convergence study and numerical simulation of static two-fluid pool, Rayleigh–Taylor instability and single rising bubble.

Modified moving particle semi-implicit method for 3D print process simulations of short carbon fiber/polyamide-6 composites

In 3D printing of short fiber-reinforced thermoplastic composites using the fused filament fabrication method, numerical simulation of the printing process is required to predict voids between print paths and to estimate the fusion of composites. In this study, the moving particle semi-implicit method for liquid simulation was extended to short fiber-reinforced thermoplastic composites considering the latent heat of melting and solidification with composite particles. Numerical simulation of the 3D printing process of short carbon fiber/PA6 composites was conducted, and the results were compared with those of experiments; the simulated print path cross-sectional configuration agreed well with the experimental results.

Solving 2-D Slamming Problems by an Improved Higher-Order Moving Particle Semi-Implicit Method

Abstract A higher-order moving particle semi-implicit (MPS) method was further developed to solve 2-D water entry problems. To overcome the inconsistency in the original MPS methods, a pressure gradient correction was implemented to guarantee the first-order consistency of the gradient. The corrective matrix was modified by using the derivative of the kernel function. A particle shifting technique was also applied to improve the numerical stability. Validation studies were carried out for water entry of a rigid wedge with the tilting angles of 0, 10, and 20, and two rigid ship sections. Convergence studies were conducted on domain size, particle spacing, and time step. A particle convergence index method was proposed to evaluate numerical uncertainties in the improved MPS method. Uncertainties in numerical solutions due to spatial discretization were calculated. The predicted impact pressures and forces by the present method are in good agreement with experimental data. Introduction Slamming can cause local structural deformation and damages. It is important to solve highly nonlinear water-entry problems involving breaking free surfaces. Many studies on the impact of pressures/ loads during water entry or slamming have been performed using experiments, analytical solutions, and numerical simulations. Extensive drop tests have been conducted. For example, Bisplinghoff and Doherty (1952) carried out a series of experiments for free-fall wedges. Because ship motions in waves can lead to asymmetric water entry, wedges with different tilt angles were tested by Barjasteh et al. (2016). A more detailed review of studies based on theoretical, potential-flow, and computational fluid dynamics (CFD) methods is presented in the work of Peng and Qiu (2018).

A multiphase MPS method coupling fluid–solid interaction/phase-change models with application to debris remelting in reactor lower plenum

In a postulated severe accident, the in-vessel debris remelting may result in severe local hot spots, probably threatening the reactor pressure vessel (RPV). The moving particle semi-implicit (MPS) method has great potential for simulating such complicated phenomena. To simulate debris remelting, the models for fluid–solid interactions and solid–solid collisions are coupled into an improved MPS method. A new speedup algorithm that can consider the flow and heat-transfer simulations separately is proposed to handle the multi-time-scale problem in debris remelting. The influence of debris size and decay heat power on the occurrence of local hot spots and RPV breach patterns during debris melting was investigated. It is found that (1) large debris blocks tended to result in severe local hot spots probably causing a lower breach point and (2) high decay power might cause the migration of local hot spots probably producing limited ablation to the vicinity of the RPV wall.

Entirely Lagrangian meshfree computational methods for hydroelastic fluid-structure interactions in ocean engineering—Reliability, adaptivity and generality

The paper presents a concise review on latest advances related to development of entirely Lagrangian meshfree computational methods for hydroelastic fluid-structure interactions (FSI) in ocean engineering and highlights several corresponding key issues. This review and highlight of key issues are made with respect to three perspectives, namely, i) reliability, ii) adaptivity and iii) generality. In specific, for reliability, we focus on stability, accuracy, proper choice of governing equations and consistent fluid-structure coupling. As for adaptivity, adaptive refinement of fluid/structure subdomains through development of multi-resolution hydroelastic FSI solvers would be considered and related developments of this aspect would be outlined. Regarding generality of FSI solvers corresponding to extension to 3D, arbitrary choice of constitutive equations, and reproduction of FSI comprising composite structures would be reviewed. The considered entirely Lagrangian FSI solvers correspond to projection-based MPS (Moving Particle Semi-implicit) or ISPH (Incompressible Smoothed Particle Hydrodynamics) fluid models coupled with Newtonian or Hamiltonian MPS/SPH structure models. Finally, future perspectives for continued developments of reliable, adaptive and general hydroelastic FSI solvers in the context of Lagrangian meshfree methods would be outlined.

A multiphase MPS method for bubbly flows with complex interfaces

The bubbly flows have always been a key research object in the field of multiphase simulations. However, when the bubbly flows become violent to cause the large deformations of bubbles, complex interfaces may be generated and bring great challenges to the traditional mesh-based methods. To overcome this challenge, one of the most commonly used mesh-free particle methods - moving particles semi-implicit (MPS) method, is adopted in this paper for the numerical simulations of bubbly flows with complex interfaces. Firstly, a multiphase MPS method is developed by combining the single-phase improved MPS (IMPS) method proposed in our previous study with different multiphase models, including the improved density smoothing scheme, interparticle viscosity model, continuous acceleration model, surface tension model, incompressible-compressible model and multiphase collision model. Then, the deformations of 2D/3D square droplets are simulated to validate the surface tension model, and the multiphase MPS method is verified through the benchmark cases of 2D/3D single bubble rising with large density and viscosity ratios (up to 1000 and 100, respectively). Finally, a series of bubbly flows with increasing complexity are performed, including the bubble rising and breaking at a high Reynolds number, the co-axial coalescence of two bubbles with different radiuses, and the violent interactions between a set of bubbles. The MPS results keep fair agreements with other numerical and experimental results, through which the advantages of the present multiphase MPS method in bubbly flows with complex interfaces can be well demonstrated.

Improvement and application of weakly compressible moving particle semi-implicit method with kernel-smoothing algorithm

The moving particle semi-implicit method (MPS) is a well-known Lagrange method that offers advantageous in addressing complex fluid problems, but particle distribution is an area that requires refinement. For this study, a particle smoothing algorithm was developed and incorporated into the weakly compressible MPS (sWC-MPS). From the definition and derivation of basic MPS operators, uniform particle distribution is critical to numerical accuracy. Within the framework of sWC-MPS, numerical operators were modified by implementing coordinate transformation and smoothing algorithm. Modifying numerical operators significantly improved particle clustering, smoothed pressure distributions, and reduced pressure oscillations. To validate the numerical feasibility of the method, several cases were numerically simulated to compare sWC-MPS to the weakly compressible MPS (WC-MPS): a pre-defined two-dimensional (2-D) analytical function, Poiseuille's flow, Taylor Green vortex, and dam break. The results showed a reduction of errors caused by irregular particle distribution with lower particle clustering and smaller pressure oscillation. In addition, a larger Courant number, which represents a larger time step, was tested. The results showed that the new sWC-MPS algorithm achieves numerical accuracy even using a larger Courant number, indicating improved computational efficiency.

Numerical analysis of melt migration and solidification behavior in LBR severe accident with MPS method

In Lead-based reactor (LBR) severe accident, the meltdown and migration inside the reactor core will lead to fuel fragment concentration, which may further cause re-criticality and even core disintegration. Accurately predicting the migration and solidification behavior of melt in LBR severe accidents is of prime importance for safety analysis of LBR. In this study, the Moving Particle Semi-implicit (MPS) method is validated and used to simulate the migration and solidification behavior. Two main surface tension models are validated and compared. Meanwhile, the MPS method is validated by the l-plate solidification test. Based on the improved MPS method, the migration and solidification behavior of melt in LBR severe accident was studied furthermore. In the Pb–Bi coolant, the melt flows upward due to density difference. The migration and solidification behavior are greatly affected by the surface tension and viscous resistance varying with enthalpy. The whole movement process can be divided into three stages depending on the change in velocity. The heat transfer of core melt is determined jointly by two heat transfer modes: flow heat transfer and solid conductivity. Generally, the research results indicate that the MPS method has unique advantage in studying the migration and solidification behavior in LBR severe accident.