Singular Perturbations, Transversality, and Sil'nikov Saddle-Focus Homoclinic Orbits

We consider the singularly perturbed system $\dot x$ =εf(x,y,ε,λ), $\dot y$ =g(x,y,ε,λ). We assume that for small (ε,λ), (0,0) is a hyperbolic equilibrium on the normally hyperbolic centre manifold y=0 and that y ₀(t) is a homoclinic solution of $\dot y$ =g(0,y,0,0). Under an additional condition, we show that there is a curve in the (ε,λ) parameter space on which the perturbed system has a homoclinic orbit also. We investigate the transversality properties of this orbit and use our results to give examples of 4 dimensional systems with Sil'nikov saddle-focus homoclinic orbits.     

Parametric study of the collision modes of compound droplets in simple shear flow

Compound droplet collision has been found in various industrial and academic applications. Such colliding phenomena of two droplets in simple shear flow are numerically resolved by a front-tracking technique. Initially, each compound droplet, assumed cylindrical, contains one concentric inner droplet. They are separated at the lateral and vertical intervals denoted by Δx₀ and Δy₀. Because of the shear flow, the compound droplets interact with each other and exhibit three collision modes: passing-over, merging (i.e. coalescence) and reversing. These modes and their transition are affected by many parameters including the Reynolds Re and Capillary Ca numbers (based on the properties of the outer fluid), the viscosity ratios μ₁₃ and μ₂₃, the interfacial tension ratio σ₁₂ of the inner to outer interfaces, the ratio of the radii of the inner to outer droplets R₁₂ and the initial distance between them (in terms of Δx₀ and Δy₀). It is found that from a merging mode, decreasing Re from 2.51 to a value less than or equal to 1.0 induces a transition to a reversing mode, whereas, increasing Ca from 0.005 to a value greater than or equal to 0.04 leads to a transition to a passing-over mode. A transition from the merging to passing-over modes also appears when varying μ₂₃ in the range of 0.1–10.0 or varying R₁₂ in the range of 0.2–0.8. A transition from a passing-over mode to a reversing one is available when increasing Δx₀ or decreasing Δy₀. Three modes of collision all occur when μ₁₃ is varied in the range of 0.1–10.0. However, the variation of σ₁₂ does not induce any transition between different modes. Several phase diagrams in terms of Re versus Ca, or Δx₀ versus Δy₀ are also proposed to show the transitions between these modes.     

Theoretical and experimental studies in fracture mechanics of crack-like defects under biaxial loading

It is a common point of view in fracture mechanics that, for any geometry of the body with a crack and any boundary conditions for the loading acting in the body plane, the stress and displacement components near the crack tip can be approximated in the framework of the theory of elasticity by a one-parameter or one-term representation, i.e., strictly in terms of the stress intensity coefficients K _I and K _II for an arbitrary failure crack [1, 2]. The authors of [2] specified the Westergaard function of the singular solution for a central crack under the biaxial loading of a plate. This approximate two-component solution has satisfactory accuracy. It is clear from [2] that this method cannot be admitted as a general statement [1], although it has long been assumed to be correct. The cause is that one cannot reasonably justify neglecting the second term in the Williams representation of the stress components in the plane case in the form of eigenfunction series; the contribution of this term in the rectangular coordinate system x, y is independent of the distance from the crack tip. This method may result in a serious mistake, from both the qualitative and quantitative viewpoints, in the prediction of local stresses, displacements, and related variables that are of interest. Apparently, this can best be demonstrated by an example of biaxial loading of a plate with a crack [1]. The unfounded neglect of the second term (whose contribution is independent of the distance from the crack tip) in the series representing the stress components is the source of the above-mentioned difficulties. In this problem, the influence of the load applied in the direction parallel to the crack plane manifests itself only in the second term of the series [3]. Therefore, this term should be clearly determined and studied in detail in the case of technological welding defects (faulty fusions, incomplete fusions, undercuts, and slag inclusions) and crack-like defects (scratches and cuts) in the base metal. The influence of the stress σ _OX along the crack axis on the stress tensor σ _x , σ _y , τ _xy and on the displacements u _x and u _y is confirmed by experimental studies of cracks by the photoelasticity method [4].     

Full-field displacement and strain rosettes by moiré interferometry

A cross-line phase-type specimen grating was interrogated by moiré interferometry techniques to produce full-field fringe patterns of three displacement components:U _x ,U _y , andU ₄₅. ForU ₄₅, sensitivity was 1/1700 mm/fringe (1/43,000 in./fr). Closely packed fringes with much information content were obtained. Fringe patterns of strain components? _x ,? _y and? ₄₅ were produced by mechanical shearing (or optical differentiation) of the displacement patterns; a shearing distance of only 0.6 mm (0.025 in.) was used. The rosette method yields complete strain information from these three components of normal strain—which were derived from three direct derivatives of displacement. Consequently, the need for cross-derivatives of displacement, which are highly sensitive to accidental rigid-body rotations, is circumvented.     

Stress and finite-strain analysis of a circular ring under diametral compression

Using a moiré, large-strain analysis method, a complete solution is shown in this paper of the fields of strain and stress for a circular ring subjected to diametral compression between two flat platens. The isotheticsu andv, obtained using 1000-lines-per-inch gratings, were differentiated photographically by the shifting technique (moiré-of-moiré) to determine ?u/?x, ?v/?y, ?u/?y and ?v/?x. Using the exact finite strain-displacement relationship, the Eulerian strains ? _x ^E , ? _y ^E and γ _xy ^E were computed. From these, the principal Eulerian strains were obtained. These results were verified with the isochromatics obtained from a large-deformation photoelasticity analysis. The ring was made of a polyurethane rubber which exhibits a linear relationship between natural strain and a newly introduced concept of “natural stress”. The Eulerian strains were converted to natural strains, and from these natural stresses were computed using the newly developed concept. Results are presented graphically for the whole field of the ring.     

State vector approach to analysis of multilayered magneto-electro-elastic plates

The state vector equations for three dimensional, orthotropic and linearly magneto-electro-elastic media are derived from the governing equations by eliminating σ_x, σ_y·τ_xy, B_x, B_y, D_x and D_y. An efficient method is presented for analysis of multilayered magneto-electro-elastic plates. The methodology is based on the mixed formulation, in which basic unknowns are formed by collecting not only displacements, electrical potential and magnetic potential but also some of stresses, electrical displacements, and magnetic induction. As special case, simply supported and multilayered rectangular plate is analyzed under the surface loading. Numerical results are presented graphically. The procedure of numerical calculation shows that the formulation presented here is simple and direct.     

Time-domain finite difference calculations for interaction of an ultrasonic wave with a surface-breaking crack

Four two-dimensional configurations are considered in this paper. The first two concern a homogeneous slab (0⩽y⩽H, −∞<x<∞), with a surface-breaking crack (x=0, 0⩽y⩽a), and without such a crack. The other two configurations concern semi-infinite slabs of different mechanical properties which are in welded contact over x=0, 0⩽y⩽H. One of these has a surface-breaking crack in the interface (x=0, 0⩽y⩽a), and the other has perfect contact over the whole interface. Results are presented for diffraction and corner reflection of an ultrasonic displacement pulse. Time-domain calculations have been carried out bu the use of the finite difference method. The results are presented as full-field snapshots of the displacement fields at specified times, and as time histories of the particle velocity at the midpoint of the transducer-specimen interface at x=−H, y=H.     

Impact energy absorption by a rate-dependent locking target

The impact by an elastic cylindrical piston on a thin plate-like target resting on a rigid foundation is considered. The relationship between force F acting on the target and displacement x is given by F=kx+q dx/dt provided dx/dt≥0 and 0≤x<d (k, q and d≥0). When x=d locking occurs, and F can assume any value ≥kd without increase in x. The displacement is assumed to be completely irreversible. The motion of the impactor is assumed to be governed by the elementary wave equation and, since the target is thin, wave motion in the target is neglected. The energy W=εFdx and its components W _k=kεx dx (the energy absorbed in a corresponding quasistatic process) and W _q=qε(dx/dt)² dt (the excessive energy because of the rate-dependence) are determined in terms of the impact energy as functions of non-dimensional parameters representing k, q and d. With the aid of diagrams, it is shown under what circumstances locking occurs, and under what circumstances W _k or W _q, or both, are large.     

Resonant non-linear vibrations in continuous systems—I. Undamped case

A method is presented for obtaining periodic solutions to forced oscillations of non-linear systems governed by equations of the form u_ss?u_yy?εf(u,u,_y,u_yy…,s) = 0. The method is presented by application to the equation u_ss?u_yy?εu²_yu_yy= 0 which governs the vibrations of a soil layer that is free on the top surface and is forced harmonically at the bedrock. It is shown that unlike the ODE case (Duffing equation), the PDE requires an infinite number of periodicity conditions to correctly characterize the resonant region and these conditions lead to an infinite number of branches in the dispersion spectrum. Calculations indicate that these branches tend to an envelope curve. The uniform approach presented by Millmann and Keller is discussed in order to determine in what sense it can be viewed as an effective approximation for the fundamental mode.     

On the similarity solutions of wave propagation for a general class of non-linear dissipative materials

Deductive similarity analysis is employed to study one-dimensional wave propagation in rate dependent materials whose constitutive laws are special cases of Maxwellian materials (σ_t = φ(ε, σ)ε_t + ψ(ε, σ), ε = strain, σ = stress). The general problem is shown not to have a similar solution although many special cases have the independent similar variable (x ? c)/(t ? d)^e. These cases are studied and tabulated. Analytic similar solutions are presented for several cases and a discussion of permissable boundary conditions is given.     

Displacement and Strain Measurement by Circular and Radial Gratings Moiré Method

Direct digital moiré method of circular and radial gratings with its phase shifting technique is proposed and applied in a mixed-mode fracture problem of large deformation materials. Circular and radial curves are used as both reference and specimen gratings in this digital moiré method for deformation measurement. And phase shifting technique is automatically implemented directly on circular and radial fringes, which facilitates obtaining displacement fields in polar coordinate system, radial u _r and circumferential u _θ , and strain fields ε _r, ε _θ , ε _rθ are calculated subsequently. Application of the direct digital moiré method of circular and radial gratings in a mixed-mode fracture problem with large deformation illustrates the process of this method, and also demonstrates its feasibility and validity for large deformation materials.     

The rheology of strongly-flocculated suspensions

The rheology of strongly-flocculated dispersions of colloidal particles has been investigated at particle concentrations where a continuous network is formed rather than a collection of discrete flocs. Such networks are shown to possess a true yield stress in both shear and in uniaxial compression (as realised in a centrifuge). Properties measured as a function of particle concentration and particle size include the yield stresses in shear (σ_y) and compression (P_y); the limiting and strain-dependent, instantaneous shear moduli G_O and G(γ); the elastic recovery at finite strains, and the rate of centrifugally-driven compaction. The yield stresses and moduli appear to show a power-law dependence on particle concentration with G_O and P_y, having the same power-law index and σ_y a somewhat lower one. The data are in part consistent with predictions based on the idea that the networks have a heterogeneous structure comprising a collection of interconnected fractal aggregates. The behaviour as a function of particle size and concentration is however not completely scaleable as might be expected on this basis. Thus, whereas the shear yield stress could be scaled to remove its dependence on particle radius a and volume fraction φ (over the measured range 0.25 μm ⩽ a ⩽ 3.4 μm; 0.05 ⩽ φ ⩽ 0.25) as could the strain dependent modulus (0.25 ⩽ a ⩽ 1.3 μm; 0.08 ⩽ 0.25), the particle-size and concentration dependence of P_y and G_O could only be scaled for particles with radii between 0.16 and 0.5 μm, smaller and larger particles having different and much higher power-law index in respect of their concentration dependencies. In the case of the smaller particles the failure of the scaling is thought to be due to an anomaly since these particles distort significantly under the influence of the strong van der Waals forces and this causes the aggregates to be more compact then they otherwise would be. The reasons for the failure at larger sizes is not clear.     

Proving conditional attractivity of a closed set with a family of liapunov functions

Considering a closed set M of some x-space and a solution x(t), y(t) of a differential system x = X(x, y, t), y = Y(x, y, t), we give sufficient conditions in order that x(t) approaches M. We use several auxiliary functions and employ Salvadori's method of a one parameter family of Liapunov functions. An application is given to the two-body problem in the presence of some friction forces and when the reference frame is non-inertial.     

A Theory of General Solutions of Plane Problems in Two-Dimensional Octagonal Quasicrystals

A theory of general solutions of plane problems is developed for the coupled equations in plane elasticity of two-dimensional octagonal quasicrystals. In virtue of the operator method, the general solutions of the antiplane and inplane problems are given constructively with two displacement functions. The introduced displacement functions have to satisfy higher order partial differential equations, and therefore it is difficult to obtain rigorous analytic solutions directly and is not applicable in most cases. In this case, a decomposition and superposition procedure is employed to replace the higher order displacement functions with some lower order displacement functions, and accordingly the general solutions are further simplified in terms of these functions. In consideration of different cases of characteristic roots, the general solution of the antiplane problem involves two cases, and the general solution of the inplane problem takes three cases, but all are in simple forms that are convenient to be applied. Furthermore, it is noted that the general solutions obtained here are complete in x ₃-convex domains.      

A unified treatment of strain gradient plasticity

A theoretical framework is presented that has potential to cover a large range of strain gradient plasticity effects in isotropic materials. Both incremental plasticity and viscoplasticity models are presented. Many of the alternative models that have been presented in the literature are included as special cases. Based on the expression for plastic dissipation, it is in accordance with Gurtin (J. Mech. Phys. Solids 48 (2000) 989; Int. J. Plast. 19 (2003) 47) argued that the plastic flow direction is governed by a microstress q_ij and not the deviatoric Cauchy stress σ_ij′ that has been assumed by many others. The structure of the governing equations is of second order in the displacements and the plastic strains which makes it comparatively easy to implement in a finite element programme. In addition, a framework for the formulation of consistent boundary conditions is presented. It is shown that there is a close connection between surface energy of an interface and boundary conditions in terms of plastic strains and moment stresses. This should make it possible to study boundary layer effects at the interface between grains or phases. Consistent boundary conditions for an expanding elastic-plastic boundary are as well formulated. As examples, biaxial tension of a thin film on a thick substrate, torsion of a thin wire and a spherical void under remote hydrostatic tension are investigated.     

Resonant non-linear vibrations of continuous systems—II Damped and transient behavior

The method presented in Part I is extended to cover the damped and transient behavior of nonlinear systems described by equations of the form u_ss?u_yy?εf(u,u_y, u_yys,…,s) = 0. The method is presented by application to the equation u_ss?u_yy?εu_yys?εu²_yu_yy= 0. Similar to the undamped case it is again shown that the PDE requires an infinite number of periodicity conditions to correctly characterize the resonant region. However, damping eliminates some of the branches of the amplitude-frequency spectrum of the undamped case. In fact, for ε = 0.25 all but the outermost branch disappear. A method of multiple time scales is presented for the study of the transient behavior and the stability of the branches for steady vibrations. The stability analysis yields an interior stable point in the amplitude-frequency spectrum which has no analog in the Duffing equation. Finally via the multiple scale procedure in the spirit of the early work of Zabusky and Kruskal one obtains forced Burgers and Korteweg-de Vries equations on a finite interval.     

Multiaxial yield behaviour of Al replicated foam

Multiaxial experiments are performed on replicated aluminium foam using a custom-built apparatus. The foam structure is isotropic, and features open monomodal pores in average diameter. Plane stress (σ₁, σ₂, σ₃=0) and axisymmetric (σ₁, σ₂=σ₃) yield envelopes are measured using cubical specimens, supplemented by tests on hollow cylindrical and uniaxial samples. In addition to the three stress components at 0.2% offset strain, the computer-controlled testing apparatus also measures the three instantaneous displacement vectors.Results show that the shape of the yield surface is independent of the relative density of the foam in the explored range (13-28%). Strain increment vectors lie, within error, roughly normal to the line traced through data points in stress space. Replicated foams feature asymmetric yield behaviour between tension and compression. The data additionally show an influence on the yield surface of the third stress tensor invariant (i.e., of the Lode angle). Simple general expressions for the yield surface are fitted to the data, leading to conclude that their behaviour is slightly better captured by parabolic rather than elliptic expressions dependent on all three stress invariants.     

The asymptotic expansions of singularly perturbed boundary value problems

In this paper we study the singularly perturbed boundary value problem:εy″=f(t,y,ε),y(0)=ξ(ε),y(1)=η(ε).where εis a positive small parameter.In the conditions:f_(?)(0,y,0)≥m_0 ,f_(?)(l,y,0)≥m_0 and f_(?)f(t,y,ε)≥0 ,we prove the existences,and uniformly valid asymptotic expansions of solutions for the given boundary value problems,and hence we improve the existing results.     

Use of equivalent body forces for acoustic emission from a crack in a plate

A method to determine acoustic emission of surface waves from a crack near the free edge of a plate, is presented, in terms of the function f(t), which defines the time dependence of the crack opening process, the crack opening volume per unit thickness of the plate, and the elastic constants of the plate. The determination of the time-varying displacement is based on the use of equivalent body forces, which are shown to be two double forces. The acoustic emission of the crack, or the equivalent radiation from the double forces, has been obtained by a novel use of the elastodynamic reciprocity theorem. It is of interest that the normal surface-wave displacement at a position x₀ of the free edge comes out as depending on df/dt evaluated at x₀ for t > x₀/c_R, where c_R is the velocity of surface waves on the free edge.