A modified SPH method to model the coalescence of colliding non-Newtonian liquid droplets

This paper develops a modified smoothed particle hydrodynamics (SPH) method to model the coalescence of colliding non-Newtonian liquid droplets. In the present SPH, a van der Waals (vdW) equation of state is particularly used to represent the gas-to-liquid phase transition similar to that of a real fluid. To remove the unphysical behavior of the particle clustering, also known as tensile instability, an optimized particle shifting technique is implemented in the simulations. To validate the numerical method, the formation of a Newtonian vdW droplet is first tested, and it clearly demonstrates that the tensile instability can be effectively removed. The method is then extended to simulate the head-on binary collision of vdW liquid droplets. Both Newtonian and non-Newtonian fluid flows are considered. The effect of Reynolds number on the coalescence process of droplets is analyzed. It is observed that the time up to the completion of the first oscillation period does not always increase as the Reynolds number increases. Results for the off-center binary collision of non-Newtonian vdW liquid droplets are lastly presented. All the results enrich the simulations of the droplet dynamics and deepen understandings of flow physics. Also, the present SPH is able to model the coalescence of colliding non-Newtonian liquid droplets without tensile instability.     

梯级溃坝洪水洪峰增强机制

我国在多条河流上修建了大量梯级水库, 梯级坝溃决诱发洪水大大超过单坝溃决洪水洪峰, 因此亟需加深对梯级坝溃决洪水洪峰增强机制的认识. 本文建立了梯级坝溃决洪水演进过程的一维浅水动力学模型, 发展了一套能捕捉激波、干湿边界和保平衡结构的数值求解方法, 通过大量算例, 系统研究了梯级坝溃决洪水演进过程的质量转化和能量转化机制. 研究结果表明, 梯级溃决中, 上游溃决诱发的洪水大大增大下游水库的质量和动量, 形成一个带动量的水塔, 同时在尾部残留一个动量较大的射流, 不断补充下游坝体溃决后水塔的质量和动量, 持续维持洪峰高度. 根据该射流-水塔机制, 建立了梯级坝溃决洪水演进过程对应的射流-水塔单坝溃决洪水过程等效模型, 该等效模型基本反映了梯级坝溃决诱发洪水的洪峰过程, 并成功预测了多个坝间距为百公里量级的梯级坝溃决洪水洪峰高程和流量, 可望为流域防洪和梯级坝设计提供理论依据.      

Momentum and heat transfer of a special case of the unsteady stagnation-point flow

This paper investigates the unsteady stagnation-point flow and heat transfer over a moving plate with mass transfer,which is also an exact solution to the unsteady Navier-Stokes(NS)equations.The boundary layer energy equation is solved with the closed form solutions for prescribed wall temperature and prescribed wall heat flux conditions.The wall temperature and heat flux have power dependence on both time and spatial distance.The solution domain,the velocity distribution,the flow field,and the temperature distribution in the fluids are studied for different controlling parameters.These parameters include the Prandtl number,the mass transfer parameter at the wall,the wall moving parameter,the time power index,and the spatial power index.It is found that two solution branches exist for certain combinations of the controlling parameters for the flow and heat transfer problems.The heat transfer solutions are given by the confluent hypergeometric function of the first kind,which can be simplified into the incomplete gamma functions for special conditions.The wall heat flux and temperature profiles show very complicated variation behaviors.The wall heat flux can have multiple poles under certain given controlling parameters,and the temperature can have significant oscillations with overshoot and negative values in the boundary layers.The relationship between the number of poles in the wall heat flux and the number of zero-crossing points is identified.The difference in the results of the prescribed wall temperature case and the prescribed wall heat flux case is analyzed.Results given in this paper provide a rare closed form analytical solution to the entire unsteady NS equations,which can be used as a benchmark problem for numerical code validation.     

Assembling hydrodynamic cloaks to conceal complex objects from drag

Transformation hydrodynamics and the corresponding metamaterials have been proposed as a means to exclude the drag force acting on an object. Here, we report a strategy to deploy the hydrodynamic cloaks in a more practical manner by assembling different-shaped cloaking parts. Our strategy is to first model a square-shaped cloak and a carpet cloak and then combine them to conceal a more complex-shaped space in the three-dimensional hydrodynamic flow. With the derivation of transformation hydrodynamics, the coordinate transformations for each hydrodynamic cloaking are demonstrated with the calculated viscosity tensors. The pressure and velocity fields of the square, triangular (carpet), and exemplary three-dimensional house-shaped cloaks are numerically simulated, thus showing a cloaking effect and reduced drag. This study suggests an efficient way of cloaking complex architectures from fluid-dynamic forces.     

离子浓度及表面结构对岩石孔隙内水流动特性的影响

酸性环境引发的岩石孔隙表面溶解增加了孔隙内水溶液的盐离子浓度,破坏了孔隙的表面结构.本文采用分子动力学模拟的方法研究了纳米级岩石孔隙内水溶液的流动特性,分析了盐离子浓度和孔隙表面结构对水流速度分布的影响及原因.研究结果表明:纳米级岩石孔隙内的水溶液流动符合泊肃叶流动特性,流速呈"抛物线"分布;随盐离子浓度增加,水溶液内部氢键网络变得更为致密,水黏度随其呈线性增长;水溶液中离子浓度越大,孔隙表面对水流动的阻力越大,最大流速越小,速度分布的"抛物线"曲率半径越大;岩石孔隙表面结构的破坏改变了流动表面的粗糙程度,增加了孔隙表面对H2O分子的吸引力.随表面结构破坏程度的增大,水溶液在近壁区域的密度增大,流速降低;当表面破坏程度达到50%时,水溶液在近壁区域出现了明显的负边界滑移现象.     

Time-averaged flow characteristics and mixing properties of double-concentric jets

We examined the flow behaviors and mixing characteristics of double-concentric jets using laser-assisted smoke flow visualization method to analyze typical flow patterns and binary boundary detection technique to investigate jet spread width. Time-averaged velocity vectors, streamline patterns, velocity distributions, turbulence properties, and vorticity contours were analyzed using Particle Image Velocimetry (PIV). Topological flow patterns were analyzed to interpret the vortical flow structures. Mixing properties were investigated using a tracer-gas concentration detection method. Four characteristic modes were observed: annular flow dominated mode, transition mode, central jet dominated mode-low shear, and central jet dominated mode-high shear. The jets’ mixing properties were enhanced by two major phenomena: the merging of annular flow and central jet at the centerline and the large turbulence fluctuations produced in the flow field. The merging of the jets induced stagnation points on the central axis in the annular flow dominated mode, which caused reverse flow on the central axis and drastic turbulence fluctuations of the near field region. When the central jet penetrated the recirculation region in the other three modes, the stagnation points on the central axis and the reverse flow vanished. Therefore, the mixing behaviors were prominently enhanced in the annular flow dominated mode.     

More efficient time integration for Fourier pseudospectral DNS of incompressible turbulence

Time integration of Fourier pseudospectral DNS is usually performed using the classical fourth-order accurate Runge-Kutta method or other second- or third-order methods, with a fixed step size. We investigate the use of higher-order Runge-Kutta pairs and automatic step size control based on local error estimation. We find that the fifth-order accurate Runge-Kutta pair of Bogacki and Shampine gives much greater accuracy at a significantly reduced computational cost. Specifically, we demonstrate speedups of 2× to 10× for the same accuracy. Numerical tests (including the Taylor-Green vortex, Rayleigh-Taylor instability, and homogeneous isotropic turbulence) confirm the reliability and efficiency of the method. We also show that adaptive time stepping provides a significant computational advantage for some problems (like the development of a Rayleigh-Taylor instability) without compromising accuracy.     

Investigation on the effect of density ratio on the convergence behavior of partitioned method for fluid–structure interaction simulation

In order to investigate the effect of density ratio of fluid and solid on the convergence behavior of partitioned FSI algorithm, three strong-coupling partitioned algorithms (fixed-point method with a constant under-relaxation parameter, Aitken’s method and Quasi-Newton inverse least squares (QN-ILS) method) have been considered in the context of finite element method. We have employed the incompressible Navier–Stokes equations for a Newtonian fluid domain and the total Lagrangian formulation for a non-linear motion of solid domain. Linear-elastic (hyper-elastic) model has been employed for solid material with small (large) deformation. A pulsatile inlet-flow interacting with a 2D circular channel of linear-elastic material and a pressure wave propagation in a 3D flexible vessel have been simulated. Both linear-elastic and hyper-elastic (Mooney–Rivlin) models have been adopted for the 3D flexible vessel. From the present numerical experiments, we have found that QN-ILS outperforms the others leading to a robust convergence regardless of the density ratio for both linear-elastic and hyper-elastic models. On the other hand, the performances of the fixed-point method with a constant under-relaxation parameter and the Aitken’s method depend strongly on the density ratio, relaxation parameter selected for coupling iteration, and degree of deformation. Although the QN-ILS of this work is still slower than a monolithic method for serial computation, it has an advantage of easier parallelization due to the modularity of the partitioned FSI algorithm.     

Downstream evolution of junction flow three-component velocity fluctuations through the lens of the diagnostic plot

Experimental measurements of the streamwise and transverse velocity components have been acquired in three spanwise/wall-normal planes in the wakes of both a streamlined ‘wing’ and a bluff ‘wing’ junction. The ‘extended’ diagnostic plot introduced by Alfredsson et al. (2011) (see figure 3 therein) is used as a benchmark to locally evaluate the departure of turbulent wing-body junction flow wakes from ‘equilibrium’ boundary layers. Both obstacles produce a secondary flow of Prandtl’s first kind, which disrupts the equilibrium implied by the universality of the extended diagnostic plot. The plane wake of the obstacle itself (away from the junction) also disrupts this equilibrium. It is found that with downstream development the boundary layer eventually recovers to the base zero-pressure-gradient ‘equilibrium’, and that this recovery process emanates from the near-wall region. The transverse velocity components are also examined in “extended diagnostic” form, revealing that the wall-normal fluctuations recover to the zero-pressure-gradient case near the wall more rapidly than the wall-parallel components.