On unsteady unidirectional flows of a second grade fluid

Some properties of unsteady unidirectional flows of a fluid of second grade are considered for flows produced by the sudden application of a constant pressure gradient or by the impulsive motion of one or two boundaries. Exact analytical solutions for these flows are obtained and the results are compared with those of a Newtonian fluid. It is found that the stress at the initial time on the stationary boundary for flows generated by the impulsive motion of a boundary is infinite for a Newtonian fluid and is finite for a second grade fluid. Furthermore, it is shown that initially the stress on the stationary boundary, for flows started from rest by sudden application of a constant pressure gradient is zero for a Newtonian fluid and is not zero for a fluid of second grade. The required time to attain the asymptotic value of a second grade fluid is longer than that for a Newtonian fluid. It should be mentioned that the expressions for the flow properties, such as velocity, obtained by the Laplace transform method are exactly the same as the ones obtained for the Couette and Poiseuille flows and those which are constructed by the Fourier method. The solution of the governing equation for flows such as the flow over a plane wall and the Couette flow is in a series form which is slowly convergent for small values of time. To overcome the difficulty in the calculation of the value of the velocity for small values of time, a practical method is given. The other property of unsteady flows of a second grade fluid is that the no-slip boundary condition is sufficient for unsteady flows, but it is not sufficient for steady flows so that an additional condition is needed. In order to discuss the properties of unsteady unidirectional flows of a second grade fluid, some illustrative examples are given.     

Hausdorff dimensnon of set generated by exceptional oscillations of a class of N-parameter Gaussian processes

A class of N-parameter Gaussian processes are introduced, which are more general than the N-parameter Wiener process. The definition of the set generated by exceptional oscillations of a class of these processes is given, and then the Hausdorff dimension of this set is defined. The Hausdorff dimensions of these processes are studied and an exact representative for them is given, which is similar to that for the two-parameter Wiener process by Zacharie (2001). Moreover, the time set considered is a hyperrectangle which is more general than a hyper-square used by Zacharie (2001). For this more general case, a Fernique-type inequality is established and then using this inequality and the ments is required for showing the representative of the Hausdorff dimension by Zacharie (2001). This property is absent for the processes introduced here, so we have to find a different way.     

Crack shielding and material deterioration in damaged materials: an antiplane shear fracture problem

Summary A simple damage evolution model is proposed for a quasibrittle material in the framework of continuum damage mechanics. The model is used to obtain a closed form solution for a mode-III stationary crack under small scale damage conditions. It is found that the crack tip stress intensity factor is reduced, i.e., the crack is shielded by the damage. However, this shielding effect is completely offset by the material deterioration caused by the damage. It also holds for steady state crack growth. When the most effective shielding is reached for the stationary crack, the zone dominated by the stress intensity factor shrinks to the crack tip. The results for the antiplane shear problem should shed some light on the in- plane fracture problem. Received 4 August 1997; accepted for publication 7 October 1997     

Hausdorff dimension of set generated by exceptional oscillations of a class of N-parameter Gaussian processes

A class of N-parameter Gaussian processes are introduced, which are more general than the N-parameter Wiener process. The definition of the set generated by exceptional oscillations of a class of these processes is given, and then the Hausdorff dimension of this set is defined. The Hausdorff dimensions of these processes are studied and an exact representative for them is given, which is similar to that for the two-parameter Wiener process by Zacharie （2001）. Moreover, the time set considered is a hyperrectangle which is more general than a hyper-scluare used by Zacharie （2001）. For this more general case, a Fernique-type inequality is established and then using this inequality and the Slepian lemma, a Levy＇s continuity modulus theorem is shown. Independence of increments is required for showing the representative of the Hausdorff dimension by Zacharie （2001）. This property is absent for the processes introduced here, so we have to find a different way.     

An equivalent polynomial approximation technique in nonlinear structural dynamics

Summary A new technique is proposed to obtain an approximate probability density for the response of a general nonlinear system under Gaussian white noise excitations. In this new technique, the original nonlinear system is replaced by another equivalent nonlinear system, structured by the polynomial formula, for which the exact solution of stationary probability density function is obtainable. Since the equivalent nonlinear system structured in this paper originates directly from certain classes of real nonlinear mechanical systems, the technique is applied to some very challenging nonlinear systems in order to show its power and efficiency. The calculated results show that applying the technique presented here can yield exact stationary solutions for the nonlinear oscillators. This is obtained by using an energy-dependent system, and for a nonlinearity of a more complex type. A more accurate approximate solution is then available, and is compared with the approximation. Application of the technique is illustrated by examples.     

A simplified model for cavity growth along a grain-boundary by coupled surface and grain-boundary diffusion

Following a model for the sintering of a row of grains by Sun et al. (1996), a simplified model is developed for cavity growth along a grain-boundary by surface and grain-boundary diffusion. The cavity surface is approximated by two arcs of equal radius truncated by the grain-boundary. The arcs evolve by changing the radius and the intersection angle they make with the grain-boundary. A variational principle for the coupled diffusion problem is used to obtain the rate equations for the two degrees of freedom which are numerically integrated to follow the cavity growth. The simplified model can be reduced to the well established equilibrium cavity growth model for the limiting case of fast surface diffusion. A validity map for the model is constructed by comparing the approximate solutions with full numerical solutions over a wide range of values of relative diffusivity, initial dihedral angles and applied stresses. It is found that the simplified solution can be used under most of the practical conditions. The model described here is two dimensional although the approach can be easily extended to axisymmetric cases.     

A mixed FE-BE method for coupled vibration of gravity dam-reservoir system with variable water depth

To solve the coupled vibration of a gravity dam-reservoir system with variable water depth by using a hybrid element method, the fluid region with variable water depth needs to be discretized by FE meshes. However, such a method asks for a great computational cost owing to the excessive unknowns, especially when the fluid region with variable water depth is relatively large. To overcome the shortcoming, a refined boundary element method is proposed to analyze the fluid field, in which only the discretization for the boundary of the variable depth region is required. But as a basis of this approach, it is necessary to construct a new Green's function corresponding to an infinite strip region. The problem is solved as the first step in this paper by employing Fridman's operator function theory, and then a mixed FE-BE formulation for analyzing the free vibration of the gravity damreservoir system is derived by means of the coupling conditions on the dam-reservoir interface. Finally, a numerical example is provided to illustrate a great improvement of the method developed herein over the hybrid element method. The project supported by the National Key Research Plan of China.     

Three-dimensional continuum analysis for a unidirectional composite with a broken fiber

The three-dimensional problem of a periodic unidirectional composite with a penny-shaped crack traversing one of the fibers is analyzed by the continuum equations of elasticity. The solution of the crack problem is represented by a superposition of weighted unit normal displacement jump solutions, everyone of which forms a Green’s function. The Green’s functions for the unbounded periodic composite are obtained by the combined use of the representative cell method and the higher-order theory. The representative cell method, based on the triple discrete Fourier transform, allows the reduction of the problem of an infinite domain to a problem of a finite one in the transform space. This problem is solved by the higher-order theory according to which the transformed displacement vector is expressed by a second order expansion in terms of local coordinates, in conjunction with the equilibrium equations and the relevant boundary conditions. The actual elastic field is obtained by a numerical evaluation of the inverse transform. The accuracy of the suggested approach is verified by a comparison with the exact analytical solution for a penny-shaped crack embedded in a homogeneous medium. Results for a unidirectional composite with a broken fiber are given for various fiber volume fractions and fiber-to-matrix stiffness ratios. It is shown that for certain parameter combinations the use of the average stress in the fiber, as it is employed in the framework of the shear lag approach, for the prediction of composite’s strength, leads to an over estimation. To this end, the concept of “point stress concentration factor” is introduced to characterize the strength of the composite with a broken fiber. Several generalizations of the proposed approach are offered.     

Conserved quantity of elastic waves in multi-layered media: 2D SH case — Normalized Energy Density

We introduce a new conserved quantity, Normalized Energy Density (NED), alternative to the conventional definition of energy for a layered structure in a 2D SH problem. NED is defined by the average of power of a half transfer function multiplied by the impedance, and the conservation across the material interface is analytically proved for a two-layered case. For three, four, and ten-layered cases, the conservation is examined by applying the Monte Carlo simulation method, and then NED is supposed to be conserved through the layers.     

Dynamic mode decomposition based analysis of flow over a sinusoidally pitching airfoil

Dynamic mode decomposition (DMD) has proven to be a valuable tool for the analysis of complex flow-fields but the application of this technique to flows with moving boundaries is not straightforward. This is due to the difficulty in accounting in the DMD formulation, for a body of non-zero thickness moving through the field of interest. This work presents a method for decomposing the flow on or near a moving boundary by a change of reference frame, followed by a correction to the computed modes that is determined by the frequency spectrum of the motion. The correction serves to recover the modes of the underlying flow dynamics, while removing the effect of change in reference frame. This method is applied to flow over sinusoidally pitching airfoils, and the DMD analysis is used to derive useful insights regarding flow-induced pitch oscillations of these airfoils.     

A bending quasi-front generated by an instantaneous impulse loading at the edge of a semi-infinite pre-stressed incompressible elastic plate

A refined membrane-like theory is used to describe bending of a semi-infinite pre-stressed incompressible elastic plate subjected to an instantaneous impulse loading at the edge. A far-field solution for the quasi-front is obtained by using the method of matched asymptotic expansions. A leading-order hyperbolic membrane equation is used for an outer problem, whereas a refined (singularly perturbed) membrane equation of an inner problem describes a boundary layer, which smoothes a discontinuity predicted by the outer problem at the wave front. The inner problem is then reduced to one-dimensional by an appropriate choice of inner coordinates, motivated by the wave front geometry. Using the inherent symmetry of the outer problem, a solution for the quasi-front is derived that is valid in a vicinity of the tip of the wave front. Pre-stress is shown to affect geometry and type of the generated quasi-front; in addition to a classical receding quasi-front the pre-stressed plate can support propagation of an advancing quasi-front. Possible responses may even feature both types of quasi-front at the same time, which is illustrated by numerical examples. The case of a so-called narrow quasi-front, associated with a possible degeneration of contribution of singular perturbation terms to the governing equation, is studied qualitatively.     

An image treatment of elastostatic transmission from an interface layer

A basic theorem for representing the Airy stress function for two perfectly bonded semi-infinite planes in terms of the corresponding Airy function for the unbounded homogeneous plane is applied in a systematic stepwise fashion to generate the corresponding Airy stress function for a three-phase composite comprising two semi-infinite planes separated by a thick layer. The loading of the three-phase composite is arbitrary, and may be in or near the interface layer. The basic theorem is first illustrated by applying it to an elastic medium which is bounded by two unloaded straight edges which intersect at an angle π/n, where n is a positive integer. This example illustrates a case of a finite system of images, while the plane-layered medium problem leads to an infinite series of images.     

Numerical analysis of turbulent flow separation in a rectangular duct with a sharp 180‐degree turn by algebraic Reynolds stress model

Turbulent flow in a rectangular duct with a sharp 180‐degree turn is difficult to predict numerically because the flow behavior is influenced by several types of forces, including centrifugal force, pressure‐driven force, and shear stress generated by anisotropic turbulence. In particular, this type of flow is characterized by a large‐scale separated flow, and it is difficult to predict the reattachment point of a separated flow. Numerical analysis has been performed for a turbulent flow in a rectangular duct with a sharp 180‐degree turn using the algebraic Reynolds stress model. A boundary‐fitted coordinate system is introduced as a method for coordinate transformation to set the boundary conditions next to complicated shapes. The calculated results are compared with the experimental data, as measured by a laser‐Doppler anemometer, in order to examine the validity of the proposed numerical method and turbulent model. In addition, the possibility of improving the wall function method in the separated flow region is examined by replacing the log‐law velocity profile for a smooth wall with that for a rough wall. The analysis results indicated that the proposed algebraic Reynolds stress model can be used to reasonably predict the turbulent flow in a rectangular duct with a sharp 180‐degree turn. In particular, the calculated reattachment point of a separated flow, which is difficult to predict in a turbulent flow, agrees well with the experimental results. In addition, the calculation results suggest that the wall function method using the log‐law velocity profile for a rough wall over a separated flow region has some potential for improving the prediction accuracy. Copyright © 2007 John Wiley & Sons, Ltd.     

An experimental method for determining dynamic fracture toughness

The boundary and loading conditions in many dynamic fracture test methods are frequently not well defined and, therefore, introduce a degree of uncertainty in the modeling of the experiment to extract the dynamic fracture resistance for a rapidly propagating crack. A new dynamic fracture test method is presented that overcomes many of these difficulties. In this test, a precracked, three-point bend specimen is loaded by a transmitter bar that is impacted by a striker bar fired from a gas gun. Different levels of energy can be imparted to the specimen by varying the speed and length of the striker to induce different crack growth rates in the material. The specimen is instrumented with a crack ladder gage, crack-opening displacement gage and strain gages to develop requisite data to determine toughness. Typical data for AISI 4340 steel specimen are presented. A simple quasi-dynamic analysis model for deducing the fracture toughness for a running crack from these data is presented, and the results are compared with independent measurements.     

A stability analysis of an oscillating body located in fluid flow using automatic differentiation

This paper presents a stability analysis of an oscillating body subjected to fluid forces located in a transient incompressible viscous flow. If the body is supported by elastic springs, oscillation will begin. If the characteristic period of the body and the excited oscillating period due to fluid forces match each other, resonance can occur. Stability analysis is therefore needed to determine the nonlinear behavior of the body. This paper presents an analysis of the changing stability of bodies by the numerical computation. To implement the computation, the motion of fluid around a body is expressed by the Navier–Stokes equation described in the arbitrary Lagrangian–Eulerian form. The fluid influence on the body is discretized by the finite element method based on a mixed interpolation by the bubble function in space. The motion of the body is assumed to be expressed by the equations of motion. To evaluate stability, stability function is defined by the total energy of the oscillating body. The stability is judged according to a stability index, obtained by the use of the automatic differentiation (AD) of the stability function. AD is a derivative computation method that gives high accuracy. By the use of AD, the second‐order derivative matrix, which is needed to compute the stability index, can be obtained exactly. For the numerical studies, analyses of one degree of freedom and two degrees of freedom (2DOF) for a circular cylinder and 2DOF for a rectangular cylinder are carried out. A combination of a cylinder and supporting elastic spring can produce stable, neutral and unstable states. It is shown that the stability of the cylinder can be determined by the stability index. This paper shows new possibilities for stability analysis of bodies located in a fluid flow. Copyright © 2008 John Wiley & Sons, Ltd.     

超大型结构固有频率实验模拟方法

对超大型结构的固有频率分析,采用常见的实验力学方法,不是受到实验技术条件的限制就是受到研究经费的制约。文章介绍了一种实验模拟方法,它由相似理论作指导,根据动力学平衡微分方程导出频率的相似准则“Kf”,从而实现了用小模型获取超大结构固有频率的转换方法,这种方法既简易又经济,适用于一般工程问题的应用,为超大型结构的动态特性研究提供了一个分析途径。作为一个实例,文章讨论了某钢厂大型水泥平台的固有频率值,所得结果与有限元计算进行了比较,数据十分吻合,说明模拟转换法是可行的。     

Hydrodynamic instability of extensional flow

The stability analysis for sheet stretching recently presented by Minoshima and White for Newtonian fluids is repeated for non-Newtonian fluids. For that purpose a constitutive law for nearly extensional flow is derived which, apart from a restriction to short memory of the fluid, is generally valid. Using this constitutive law the result found by Minoshima and White is generalized. Application of the general result to a Maxwell-type fluid shows that the elasticity of the fluid has a stabilizing influence.     

The transverse and torsional vibrations of A.C. motors in the starting process

According to the electromagnetic field theory, a set of differential equations is derived for coupling of the transient process for an electrical machine and transverse and torsional vibrations for the rotor by a unified method. A non-dimensional constant coefficient state equation is obtained by transformation. Computation is done. Rules of the change are obtained for transverse and torsional vibration, unilateral magnetic pull, electrical torque, and so on in the starting process. The theoretical results are shown to be in good agreement with the experimental results. The mathematical model verified by the experiment is used to analyse the effect of electromagnetic and mechanical parameters on electromagnetic force and vibration.