EI、Scopus 收录

2009 Vol. 41, No. 1

Display Method:
Numerical study of the stability of hypersonic wake
Qing Shen, Dehua Zhu
In this paper, a numerical simulation on hypersonic wakeflow over a sphere is carried out to study the wake instability and itsevolving process with Mach number 6 and Reynolds number 1.71\times 10^6(based on the radius of sphere). Firstly, a stable solution of the flow is obtained;one primary separated zone and one second separated zone are presented inthe base flow while the flow is axisymmetric. Secondly, weakly unsteady flowhas been developed slowly in the base flow to induce the obvious structuralinstability and obtain the periodic solution with the dimensionless periodvalue of 12.0. Thirdly, the evolving process of the vortical structure andits singularity characteristic are also analyzed. The results indicate thatinstable phenomenon will exist in the base flow of hypersonic flow over asphere and the present simulation method has the ability to study thisinstability problem.
2009, 41(1): 1-7. doi: 10.6052/0459-1879-2009-1-2008-249
nfluence of exciting parameters on the vibration of single cavitation bubble driven by intensive sound
In this paper, the ultrasonic cavitation phenomenon of single bubble is numerically investigated with an improved initial bubble radius associated with the improved dynamic equation (Rayleigh-Plesse equation) considering the effect of compressible of the micro-bubble. The numerical results indicate that the bubble vibration is very sensitive to some exciting parameters. Enhancing the ultrasonic signal frequency can decrease the vibration-initial radius ratio of the bubble. On the contrary, the collapse of bubble will become more violent with higher acoustic pressure, but overhigh pressure will make it never happen. With the increase of bubbles' original radius, the collapse velocity of the bubble becomes higher; the vibration of cavitation bubble can keep on a steady process in a range of initial bubble radius. The cavitation is strongest at the initial radius of 1.6$\mu$m, and will weaken or never happen when initial bubble radius keep on increasing. Enhancing liquid surface tension or viscosity can weaken the collapse of bubble, but the collapse is hard to occur when they become higher. A comparison between the numerical results and the published experimental data shows that the liquid compressibility has a great influence on the determination of zone boundaries of the best cavitation, and makes the cavitation intensity of single bubble stronger.
2009, 41(1): 8-14. doi: 10.6052/0459-1879-2009-1-2007-461
Application of spalding formula in wall friction stress measurement on riblet surface
Huang Ming-ming
The mean velocity distributions in turbulent boundarylayers of smooth surface and two different riblet surfaces are measured withIFA300 constant-temperature anemometer in low-speed wind tunnel to validatethe full development of turbulent boundary laye. Friction velocity and wallfriction stress of turbulent boundary layer are obtained with theleast-squares fitted curve of the sublayer velocity distribution based onSpalding formula, and virtual origin position of riblet surface can beacquired accurately. Moreover, the momentum loss thicknesses of these threeexperimental models is also calculated. The results show that the dragreduction effect is obvious in turbulent flows due to the riblet surfacewith the comparsion of the momentum loss thicknesses and wall frictionstresses of smooth surface and riblet surface.
2009, 41(1): 15-20. doi: 10.6052/0459-1879-2009-1-2008-026
Experimental study on turbulent characteristics of a vertical jet in wave environment
This paper presents a turbulent jet vertically discharged into a flow fieldof regular waves in the intermediate between deep and shallow water usingthe particle image velocimetry (PIV) technique. Phase average method wasused to extract the turbulent velocity from the PIV measurements. Four waveswith different wave heights were applied in the experiments. The resultsfirstly showed that the turbulent characteristics of jets are significantlyaffected by the wave height. The convection term is more sensitive than theturbulent diffusion term for the wave height change. In addition, the ratioof the turbulent diffusion term to the convection term is about $1 / 8\sim 1/ 3$ to show that the effect of the turbulent diffusion can't be ignored intime-averaged N-S equations.
2009, 41(1): 21-27. doi: 10.6052/0459-1879-2009-1-2007-025
Investigations on external flow field of an exited detonation
Pulse Detonation Engine (PDE) tests in the PDE tubeH_2-O_2-N_2 have beenperformed with central ignition in the tube end. The tube has 1000mm longwith the inner diameterof 30mm, filled with the H_2-O_2-N_2 mixture.Moreover, numerical simulationsof pulse detonation process induced by flame in the tube withH_2-O_2-N_2 mixture andthe distribution of the internal and external flow field were performed. Adetailed chemical mechanism of H_2-O_2-N_2 system was investigatedwith 19 element reactions and9 species. The generation and development of the vortex ring and suspendedshock were exhibited with the sequential shadowgraphs of the external flowfield both in test and simulation. The numerical results were in goodagreement with the measured results. The experimental and numerical resultsshow that the vortex ring was induced by baroclinic effect and Helmholtzinstability in the vicinity of the thruster exit after the donation waveexhausted the tube. The intensity, shape and location of the suspended shockvaried when the shape of the vortex ring changed.
2009, 41(1): 28-34. doi: 10.6052/0459-1879-2009-1-2006-606
Numerical simulation of interstitial fluid based on a new view of starling's hypothesis of capillary wall
Guanghong Ding
Interstitial fluid is the surroundings that cells livein. Spatial distribution of protein in interstitial space is not uniform.This article is to numerically investigate the interstitial fluid ininterosseous membrane of lower limb based on a new view of Straling'shypothesis and to study the effect of unequal distribution of protein in theinterstitial fluid. The results showed these parallel capillaries could leadto directional flow of interstitial fluid under the equal and unequalprotein distribution conditions. The fiber matrix in the capillary wallopposed protein filtration induced the unequal distribution of protein. Thehigh pressure in the capillary near the artery prevented protein fromdiffusing to the capillary wall to induce the lower protein concentration.Furthermore, the unequal spatial distribution of protein in interstitialspace had a strong influence on the interstitial fluid flow. The numericalresults in the present work were in good agreement with the experimentaldata from Weinbaum.
2009, 41(1): 35-40. doi: 10.6052/0459-1879-2009-1-2007-534
A macro traffic flow model with three lanes and numerical study
Zheng Wu
With the development of traffic infrastructure, multilanes become common in city freeways. Because vehicles can change lanes inmulti lanes, new traffic problems have appeared. In the present paper, acontinuous traffic flow model for the unilateral three lanes systems and itsdifferent discretization schemes are developed based on the macro trafficflow theory associated with the lane changing of vehicles. Numerical studyon the propagation and development of traffic disturbance is carried outwith the different initial vehicle densities, disturbance intensions anddisturbance bounds. The results show that the direction and the speed of thedisturbance propagation are mainly induced by the initial vehicle density,and the disturbance bound may influence the disturbance wave amplitude, todetermine the impact of other lanes. Furthermore, the development of trafficdisturbance in multi-lanes systems with middle vehicle density shows to becomplex and non-linear, and the choice of numerical schemes with three-lanemay be more difficult than those with single lane.
2009, 41(1): 41-48. doi: 10.6052/0459-1879-2009-1-2007-172
Application of the fictitious domain method in fluid- structure interaction
Chunning Ji, Yuanzhan Wang, Jianfeng Wang
Fluid-Structure Interaction (FSI) is an engineeringproblem with many practical aspects such as the swash of seabed, thestagnation of contamination, the settlement of sand around coastalstructures, and the resonance of structures under internal/external currentexcitations. Many numerical schemes had been presented in the past twodecades. This paper deals with a finite element method to simulate theinteraction of a coupled incompressible fluid-rigid structure system. Basedon the Fictitious Domain (FD) method in the multi-phase flows field, a setof fully coupled FSI governing equations is presented. In this``monolithic'' approach, the structure is taken as ``fictitious'' fluid withzero strain rate and the whole computational domain is modeled by theNavier-Stokes equations. However, to keep the rigid body shape and behaviorsof the ``fictitious'' fluid, the Distributed Lagrange Multiplier (DLM)method is applied on this domain. The whole field, including fluid regionand structure region, is described by velocity and pressure, and the entireset of model equations is discretised with fixed Eulerian mesh. Three majoradvantages of the present formulation include: (i) The unitized governingequations both for fluid and structure help capturing the predominantphysics of interaction phenomena; (ii) The interfacial force/displacementbetween fluid and structure are internal actions for the overall system.Therefore, the stress/velocity consistency conditions on the fluid-structureinterface are automatically satisfied in this fully coupled model; (iii) Forthe using of fixed Eulerian mesh, it is not necessary to remesh thecomputational domain, and thus free from mesh distortions. Results fromnumerical simulations on particulate flow and FSI problems provideconvincing evidences for the model's high accuracy and the suitability forthe simulation of large-deformation/movement FSI problems.
2009, 41(1): 49-59. doi: 10.6052/0459-1879-2009-1-2007-549
Energy non-local model and new strain gradient theory
Dake Yi, T.C. Wang
A new strain gradient theory based on energy non-localmodel is proposed in the paper, and applied to predict the micro-indentationhardness of polycrystalline copper and the thin film--substrate system.First, an energy non-local model is presented to derive the constitutiveequation of a new strain gradient theory. Second, the finite element formulaof the new strain gradient theory is obtained with the case of theprediction of micro-indentation hardness. It is found that the predictedresults by the new strain gradient theory agree well with the experimentaldata, while the hardness values predicted by classic plasticity theory areremarkably lower than the experimental data.
2009, 41(1): 60-66. doi: 10.6052/0459-1879-2009-1-2008-025
The elasto-plastic damage constitutive relations of orthotropic materials
Based on the elasto-plastic mechanics and continuumdamage theory, a yield criterion that relates to the spherical tensor ofstress is proposed to describe the mixed hardening of damaged orthotropicmaterials, and the dimensionless form of which is isomorphic with the Misescriterion of isotropic materials. Furthermore, the incrementalelasto-plastic damage constitutive equations are established. As an exampleof its application, the elasto-plastic buckling problem of the orthotropicrectangular thin plates with local damage are investigated by using theGalerkin method and iteration method, and the effects of geometricparameters and local damage on the critical buckling stress are discussed.
2009, 41(1): 67-75. doi: 10.6052/0459-1879-2009-1-2006-585
The asymptotic elastic-viscoplastic field at mode II dynamic propagating crack-tip
Zhenqing Wang
Investigations on crack-tip stress-strain fields indifferent media are of engineering significance. The development of crack isdetermined by the mechanical state of high stress zone near crack tip due todynamic fracture failure as the major structural catastrophic form. Thestructure of crack tip field is complex and it is hard to get the solutionof asymptotic fields with the viscous effect. In the present paper, amechanical model is established in order to investigate the viscous effectin dynamic growing crack-tip field of mode II crack. Under the assumptionthat the viscosity coefficient is in inverse proportion to power law of theplastic equivalent strain rate, the elasticity, plasticity and viscosity ofmaterial at crack-tip can be matched reasonably by asymptotic analysis, andthe boundary conditions can be served as the supplementary conditions tosolve the growing crack-tip field. An elastic-viscoplastic asymptoticanalysis is carried out on the moving crack-tip fields in perfect plasticmaterials under plane-strain condition. A continuous solution withoutdiscontinuities is obtained And the variations of numerical solution formode II crack are discussed. It is shown that both stress and strain possessexponential singularity, and the tip field contains no elastic unloadingzone for mode II crack. By introducing Airy stress function, the governingequation of mode II quasi-static crack tip has been obtained and analyzednumerically. When the crack moving speed approaches zero, the quasi-staticsolution is recovered to show that the quasi-static solution is a specialcase of a dynamic one.
2009, 41(1): 76-83. doi: 10.6052/0459-1879-2009-1-2007-345
Perforation modes of metal plates struck by a blunt rigid projectile
Xiaowei Chen
The present paper analyzes the possible modes of shearplugging and adiabatic shear plugging in the perforation of metal platesstruck by a blunt rigid projectile. The modified ballistic limit andresidual velocity with adiabatic shear plugging are further formulated.Further experimental analyses are conducted on the perforations of Weldox Esteel plates, in order to discuss the effects of plate thickness andmaterial strength/hardness on the terminal ballistic performance. Moreexperimental evidences confirm the jump of residual velocity at ballisticlimit induced by the structural response of plate. With increasing thethickness of plate and the material strength, failure modes of plate maytransform from shear plugging to adiabatic shear plugging.
2009, 41(1): 84-90. doi: 10.6052/0459-1879-2009-1-2007-572
Finite-strain consolidation finite-element method subject to non-conservative body force
Numerical analysis of finite-strain consolidation is investigatedby finite element method with non-conservative body force. Firstly, amodified static equilibrium equation for finite-strain consolidation isproposed during the consolidation process of saturated soils using TotalLagrangian description with non-conservative body force. Secondly, thecorresponding incremental finite-element (FE) )formulations of finite-strainconsolidation are presented with the substitution of the non-conservativebody force for a fellow body force. Both the equilibrium equation and thenew FE formulations can be reduced to the traditional conservative forms byneglect of the volume strain of soils, respectively. A case of finite-strainself-weight consolidation was presented to study the significance ofnon-conservative body force. The numerical results show that traditionalmethod underestimates the consolidation degree of foundation, butoverestimates the consolidation settlement and excess pore water pressure.The present finite-strain consolidation analysis based on non-conservativebody force is available to precisely predict the consolidation behavior offoundation.
2009, 41(1): 91-97. doi: 10.6052/0459-1879-2009-1-2008-224
Topology optimization of continuum structures subjected to self weight loads
Xiong Zhang
The topology optimization of continuum structures withself weight loads is investigated in this paper. Such difficulties asnon-monotonous behavior of the compliance, inactive volume constraint of theoptimal topology and the parasitic effect of low density regions are oftenarisen in topology optimization problems with self weight loads. Solutionstrategies based on the RAMP material model and the mean sensitivityfiltering technique are proposed to get rid of the difficulties and MMA codeis used to solve the optimization problem. The influence of penalty factorsof the RAMP model on the optimal topology results is investigated. Researchresults show that the material resource constraint turn to active when thepenalty factor is set to an appropriate value (for example, 20.0). Inaddition, compared with the sensitivity filtering suggested by Sigmund,discrete 0/1 design can be obtained by the mean sensitivity filteringtechnique.
2009, 41(1): 98-104. doi: 10.6052/0459-1879-2009-1-2007-285
Modeling and simulation of multi-body systems with multi-friction and fixed bilateral constraint
Qi Wang
Modeling and simulating the dynamics of the multibodysystems with bilateral constraints and dry friction are important inmechanical system and robotics. For smooth bilateral constraints, it is easyto solve the dynamical equations numerically. The dynamic equations of themultibody systems with the friction of constraint are the discontinuousdifferential-algebraic equations (DAE) and the equations cannot be expressedas being linear with respect to the generalized accelerations and theLagrange multipliers directly. In the present paper, modeling of planarmulti-rigid-body system with multi-friction and fixed-bilateral constraintsis proposed. It is assumed that the system has sliding joints with Coulomb'sdry friction and smooth hinge joints, while the sliding joints move alongthe fixed-slots. Firstly, the motion equations of the system are derivedfrom Lagrange's equations of the first kind in Cartesian coordinate system,and constraint equations are expressed by local approach. A one-to-one mapbetween the normal constraint forces and the Lagrange multipliers isestablished to analysis and compute the friction forces. Secondly, using theconstraint equations and the principle of virtual work, the generalizedforces of the friction forces are derived in the matrix form. The absolutevalue of Lagrange multiplier $| \lambda |$ in the motion equationsis given as $\lambda {\rm sgn} ( \lambda )$ by sign function. Therefore,the sign function, ${\rm sgn}(\lambda )$, ${\rm sgn}(\dot {s})$ and${\rm sgn}(\ddot {s})$,included in the motion equations, correspond to Lagrange multipliers, thevelocity and tangential acceleration of the slider, respectively. Thirdly,the DAE are transformed into ordinary differential equations (ODE) by meansof the augmentation approach. An improved trial-and-error method is proposedaccording to the characteristics of the piecewise smooth of thesystems, which can improve the efficiency of computation. Finally, anexample of one degree of freedom mechanism is given by improvedtrial-and-error method and R-K method.
2009, 41(1): 105-112. doi: 10.6052/0459-1879-2009-1-2008-309
On the contact law and structure responses of sandwich beam subject to low-velocity impact
Jiaxi Zhou, Zichen Deng
The paper presented the contact law and structuralresponse of sandwich beam subject to low-velocity impact. Firstly, the upperface sheet of the sandwich beam is assumed as a beam supported on theelastic-plastic foundation. The quasi-static contact law is derived from theequilibrium equation of the beam associated with large deformation and axialtensile stiffness of the upper face sheet, and validated with the structureresponses of the sandwich beam fixed on the rigid plate subject tolow-velocity impact. Secondly, the structure responses of simply supportedsandwich beam with low-velocity impact are investigated by energy-balancemodel, discrete model and complete model to obtain the impact duration, thehistory of contact force and the transverse deflection. In addition, theeffects of the mass ratio and the initial velocity on the structuralresponses and the applicability of different models are investigated. Thetheoretical and numerical results demonstrate that the contact law proposedby present paper is efficient. All the models in the paper can be used toacquire the structural response when the mass ratio is large, and only thecomplete model can be accurately predict the response when the mass ratio issmall or the initial velocity is large, and the impact duration predicted bycomplete model is not influenced by the initial velocity.
2009, 41(1): 113-121. doi: 10.6052/0459-1879-2009-1-2007-427
Dual reciprocity hybrid boundary node method for elastodynamics
Yu Miao, Yuanhan Wang
Combined the hybrid boundary node method (HBNM) and thedual reciprocity principle, a truly boundary-type meshless method, dualreciprocity hybrid boundary node method (DR-HBNM), is presented forelastodynamics. HBNM is a truly boundary-type meshless method, which has theadvantages of meshless method and the BEM. The HBNM is used to solve thehomogeneous equations, and the DRM is employed to solve the inhomogeneousterms using radial basis functions (RBF). The domain integrals for theinhomogeneous terms in the govern equations of elastodynamics can be takeninto boundary integrals using the reciprocity principle. The solution isdivided into the particular part and the general part. No meshes are neededeither for the interpolation purposes or for integration process. Onlydiscrete nodes are constructed on the boundary of a domain and several nodesin the domain are needed just for the RBF interpolation. The serious``boundary layer effect'' in the HBNM is circumvented by the adaptiveintegration scheme. The present method has many advantages, such as simplepostprocess and higher accuracy. The numerical exampleshave demonstrated the accuracy and convergence of the present scheme.
2009, 41(1): 122-128. doi: 10.6052/0459-1879-2009-1-2008-054
Bifurcation, chaotic dynamics and control of piezoelectric laminated composite beam
Z. G. Yao
The paper investigated the nonlinear dynamics, bifurcation and chaoticdynamics of simply supported laminated piezoelectric composite beam withaxial and transverse loading. The nonlinear equations of motions of the beamare first derived based on the general von Karman-type equations and theReddy third-order shear deformation plate theory. The partial differentialequations are discretized to the ordinary differential equations using theGalerkin approach. The averaged equations are obtained with the method ofmultiple scales. Numerical simulation is carried out to investigate theperiodic and chaotic motions of the beam. The results show that the chaoticresponses of the beam are sensitive to the piezoelectric excitations.Furthermore, the oscillation of the piezoelectric laminated composite beamcan be controlled with the piezoelectric excitation.
2009, 41(1): 129-140. doi: 10.6052/0459-1879-2009-1-2008-250