Optimal prediction for moment models: crescendo diffusion and reordered equations

A direct numerical solution of the radiative transfer equation or any kinetic equation is typically expensive, since the radiative intensity depends on time, space and direction. An expansion in the direction variables yields an equivalent system of infinitely many moments. A fundamental problem is how to truncate the system. Various closures have been presented in the literature. We want to generally study the moment closure within the framework of optimal prediction, a strategy to approximate the mean solution of a large system by a smaller system, for radiation moment systems. We apply this strategy to radiative transfer and show that several closures can be re-derived within this framework, such as P _N , diffusion, and diffusion correction closures. In addition, the formalism gives rise to new parabolic systems, the reordered P _N equations, that are similar to the simplified P _N equations. Furthermore, we propose a modification to existing closures. Although simple and with no extra cost, this newly derived crescendo diffusion yields better approximations in numerical tests.     

Characterization of material inhomogeneity by stationary values of strain energy density

The strain energy density theory is applied to analyze the fracture instability of a mechanical system whose behavior is governed by the interaction of geometry, load and material inhomogeneity. This is accomplished by obtaining the location of the global and local relative minima of the strain energy density function dW/dV denoted, respectively, by [(dW/dV)_min]_g and [(dW/dV)_min]_¢l. The former refers to a fixed global coordinate system for the entire solid while the latter corresponds to local coordinate systems referred to each material point. An unique length parameter “ℓ” representing the distance between [(dW/dV)_min^max]_g and [(dW/dV)_min^max]_ℓ can thus be found and serves as a measure of the degree of system instability tending toward failure by fracture.Numerical results are obtained and displayed graphically for the case of a solid containing an inclusion of dissimilar material. The changes that take place in material inhomogeneity, loading type and physical dimensions of the solid and inclusion are reflected through ℓ. The method suggests the compatibility of ℓ for each member of a multi-component structure in order to avoid premature failure of a single member.     

Further Study on Pure Shear

It is known that the Cauchy stress tensor T is a pure shear when trT = 0. An elementary derivation is given for a coordinate system such that, when referred to this coordinate system, the diagonal elements of T vanish while the off-diagonal elements τ ₁, τ ₂, τ ₃, are the pure shears. The structure of τ _i (i = 1, 2, 3) depends on one non-dimensional parameter q = 54(detT)² / [tr(T ²)]³, 0 ≤ q ≤ 1. When q = 0, one of the three τ _i vanishes. A coordinate system can be chosen such that the remaining two have the same magnitude or one of the remaining two also vanishes. When q = 1, all three τ _i have the same magnitude. However, there is a one-parameter family of coordinate systems that gives the same three τ _i . For q ≠ 0 or 1, none of the three τ _i vanishes and the three τ _i in general have different magnitudes. Nevertheless, a coordinate system can be chosen such that two of the three τ _i have the same magnitude. ^★Professor Emeritus of University of Illinois at Chicago and Consulting Professor of Stanford University.     

Scattering of a plane monochromatic wave by a system of strips

Summary The diffraction of a plane wave by a general system of strips is treated by means of separation of variables. A new addition theorem for Mathieu functions is used to satisfy the boundary conditions on the strips. Numerical calculations are performed for the case of two strips lying in the same plane, the boundary conditions being grad _n =0, while the angle of incidence of the plane wave is arbitrary. The transmission coefficient for a system of two slits in a plane, perfectly conducting screen is calculated for a range of values of the parameters. An approximate expression relating the transmission coefficient for the system of two slits to the transmission coefficient for a system of a single slit is given. As the distance between the two slits is increased, the transmission coefficient for the system of two slitsvery rapidly becomes nearly identical with the transmission coefficient for a single slit.     

Unsteady state diffusion of vapour in a closed tube

Results of one-dimensional calculations for unsteady state diffusion of a vapour in a closed tube are compared to calculations (a) for unsteady state diffusion of a vapour in an open tube, and (b) for heat conduction in a finite slab and an infinite slab. In the cases of (b) the effect of the convective term in the transport equation is absent.The total pressure is calculated and differences in the two cases are explained.An experiment is described in which the pressure increase in a closed tube due to the diffusion of water vapour in dry air is measured as a function of time. The experimental results agree quite well with the theoretical predictions. The results of this investigation may be used for the experimental determination of the diffusion coefficient of a vapour in a gas and in pressure measurements in systems with an evaporating liquid.Nomenclature a thermal diffusivity - D diffusion coefficient - w mean particle velocity - g particle flux in the laboratory system - g* particle flux in a system moving with the mean particle velocity w - L length of the tube - n number density of molecules - n _vs saturation number density of vapour molecules - p pressure - p _vs saturation pressure of the vapour - t time - z distance from the liquid surface - N total quantity of evaporated liquid     

Optimization of Conical Wings in Hypersonic Flow

A method of calculation is presented to determine conical wing shapes that minimize the coefficient of (wave) drag, C _D, for a fixed coefficient of lift, C _L, in steady, hypersonic flow. An optimization problem is considered for the compressive flow underneath wings at a small angle of attack δ and at a high free-stream Mach number M _∞ so that hypersonic small-disturbance (HSD) theory applies. A figure of merit, F=C _D/C _L ^3/2, is computed for each wing using a finite volume discretization of the HSD equations. A set of design variables that determine the shape of the wing is defined and adjusted iteratively to find a shape that minimizes F for a given value of the hypersonic similarity parameter, H= (M _∞δ)⁻², and planform area. Wings with both attached and detached bow shocks are considered. Optimal wings are found for flat delta wings and for a family of caret wings. In the flat-wing case, the optima have detached bow shocks while in the caret-wing case, the optimum has an attached bow shock. An improved drag-to-lift performance is found using the optimization procedure for curved wing shapes. Several optimal designs are found, all with attached bow shocks. Numerical experiments are performed and suggest that these optima are unique. Received 1 May 1998 and accepted 14 October 1998     

Evolution of a heavy rigid body rotation under the action of unsteady restoring and perturbation torques

The paper develops an approximate solution to the system of Euler’s equations with additional perturbation term for dynamically symmetric rotating rigid body. The perturbed motions of a rigid body, close to Lagrange’s case, under the action of restoring and perturbation torques that are slowly varying in time are investigated. We describe an averaging procedure for slow variables of a rigid body perturbed motion, similar to Lagrange top. Conditions for the possibility of averaging the equations of motion with respect to the nutation phase angle are presented. The averaging technique reduces the system order from 6 to 3 and does not contain fast oscillations. An example of motion of the body using linearly dissipative torques is worked out to demonstrate the use of general equations. The numerical integration of the averaged system of equations is conducted of the body motion. The graphical presentations of the solutions are represented and discussed. A new class of rotations of a dynamically symmetric rigid body about a fixed point with account for a nonstationary perturbation torque, as well as for a restoring torque that slowly varies with time, is studied. The main objective of this paper is to extend the previous results for problem of the dynamic motion of a symmetric rigid body subjected to perturbation and restoring torques. The proposed averaging method is implemented to receive the averaging system of equations of motion. The graphical representations of the solutions are presented and discussed. The attained results are a generalization of our former works where µ and M_i are independent of the slow time τ and M_i depend on the slow time only.

     

Localized Thickening of a Compressed Elastic Band

An infinite elastic band is compressed along its unbounded direction, giving rise to a continuous family of homogeneous configurations that is parameterized by the compression rate β < 1 (β = 1 when there is no compression). It is assumed that, for some critical value β ₀, the compression force as a function of β has a strict local extremum and that the linearized equation around the corresponding homogeneous configuration is strongly elliptic. Under these conditions, there are nearby localized deformations that are asymptotically homogeneous. When the compression force reaches a strict local maximum at β ₀, they describe localized thickening and they occur for values of β slightly smaller than β ₀. Since the material is supposed to be hyperelastic, homogeneous and isotropic, the localized deformations are not due to localized imperfections. The method follows the one developed by A. Mielke for an elastic band under traction: interpretation of the nonlinear elliptic system as an infinite dimensional dynamical system in which the unbounded direction plays the role of time, its reduction to a center manifold and the existence of a homoclinic solution to the reduced finite dimensional problem in [A. Mielke, Hamiltonian and Lagrangian fiows on center manifolds, Lecture Notes in Mathematics 1489. Springer, Berlin Heidelberg New York, 1991]. The main difference lies in the fact that Agmon's condition does not hold anymore and therefore the linearized problem cannot be analyzed as in Mielke's work.     

Modelling of rotational factor in notched bend specimen under general and local yield situation

The location of the plastic hinge axis in a three point SEN bend specimen is a highly controversial issue. An unambiguous and reliable estimation of rotational factor (r_p) is very essential for the accurate determination of CTOD data. In contrast to the numerous studies reported on the r_p determination in a cracked situation, limited information is available for a blunt notch situation, although many engineering structures do contain notchlike defects with finite root radius. An attempt is made to determine r_p for two situations, namely well below the general yield and around the general yield. The work is based on a theoretical estimation of the plastic zone size using the stress concentration factor and the elastic as well as the elastic-plastic stress distribution. A theoretical estimation of r_p in both the pseudo-elastic and the elastic-plastic situation is estimated through analytical modelling involving factors like plastic zone size, bend angle and notch opening displacement. The values of the rotational factor are found to increase from a small value to around 0.29 in a well below general yield situation to 0.53 to 0.54 in a general yield situation with continued loading. A wide discrepancy in the P/P_GY ratio separating the two situations, i.e. well below general yield and around general yield, is observed. Consideration of the elastic and the elasto-plastic stress distribution indicates a much smaller value of P/P_GY as compared to the ratio obtained from experimental load-displacement plots.     

Heat transfer in the flow of a viscoelastic fluid over a stretching surface

An analysis is made of heat transfer in the boundary layer of a viscoelastic fluid flowing over a stretching surface. The velocity of the surface varies linearly with the distance x from a fixed point and the surface is held at a uniform temperature T _w higher than the temperature T _∞ of the ambient fluid. An exact analytical solution for the temperature distribution is found by solving the energy equation after taking into account strain energy stored in the fluid (due to its elastic property) and viscous dissipation. It is shown that the temperature profiles are nonsimilar in marked contrast with the case when these profiles are found to be similar in the absence of viscous dissipation and strain energy. It is also found that temperature at a point increases due to the combined influence of these two effects in comparison with its corresponding value in the absence of these two effects. A novel result of this analysis is that for small values of x, heat flows from the surface to the fluid while for moderate and large values of x, heat flows from the fluid to the surface even when T _w >T _∞. Temperature distribution and the surface heat flux are determined for various values of the Prandtl number P, the elastic parameter K ₁ and the viscous dissipation parameter a. Numerical solutions are also obtained through a fourth-order accurate compact finite difference scheme. Received on 14 October 1997     

The isotope effect of gaseous hydrogen under shock compression

The shock compression method has been used to measure the Hugoniot data and shock temperature for gaseous hydrogen samples, covering the pressure range of 55-140 MPa and the temperature range of 3400-4500 K and with the initial conditions of P ₀ = 0.6 MPa, 1.2 MPa and T ₀ at room temperature. Spectral radiance histories emitted from shocked D ₂ and H ₂ + D ₂ (equimolar mixture) are monitored by a pyrometer system with seven wavelength channels. Theoretical calculations based on the Saha model with Debye-Hückel correction for the shock compression behavior of shocked gaseous samples are in good agreement with the measured Hugoniot data, but show slightly higher values for the shock temperature when comparing with experiments. An isotope effect relevant to these shocked hydrogen species has been found in the linear shock velocity vs particle velocity relation, in which the correlation factor between these hydrogen isotopes or hydrogen mixtures is simply of initial density dependence.Received: 8 December 2002, Accepted: 8 May 2003, Published online: 2 September 2003PACS: 62.50 + p, 31.30.GS, 51.90. + r     

Stability of elastic systems under a stochastic parametric excitation

An efficient method to investigate the stability of elastic systems subjected to the parametric force in the form of a random stationary colored noise is suggested. The method is based on the simulation of stochastic processes, numerical solution of differential equations, describing the perturbed motion of the system, and the calculation of top Liapunov exponents. The method results in the estimation of the almost sure stability and the stability with respect to statistical moments of different orders. Since the closed system of equations for moments of desired quantities y _j (t) cannot be obtained, the statistical data processing is applied. The estimation of moments at the instant t _n is obtained by statistical average of derived from the solution of equations for the large number of realizations. This approach allows us to evaluate the influence of different characteristics of random stationary loads on top Liapunov exponents and on the stability of system. The important point is that results found for filtered processes, are principally different from those corresponding to stochastic processes in the form of Gaussian white noises.     

The behavior of cracked cross-ply composite laminates under shear loading

An interlaminar-shear-stress analysis developed earlier by Tsai et al. (1990, Micro-cracking-Induced Damage in Composites) for a [φ_m/θ_n], bi-directional composite laminate is used to solve the case of a cross-ply [0_m/90_n]_x laminate with the 90° layer only or both layers cracked under pure shear loading. Strains, forces and laminate shear modulus reduction due to matrix cracking were obtained. Experimental results for shear modulus as a function of crack densities were obtained by a simple shear test and they agree very well with the theoretical prediction.     

Some Optimization Problems for Vibroprotective Systems

Some analytically solvable problems of optimal protection against vibration and shocks are considered in linear and nonlinear formulations. In the linear formulation, they can be solved by using up-to-date methods of systems theory (optimization in Hardy H ₂ H spaces, solution of the algebraic Riccati equation, synthesis of optimal systems under uncertainty, etc.). An expression is presented for the optimal nonlinear characteristic (tangensoid) of a damper with a limited free running under broadband stochastic vibrations. Relations linking the free running of the damper, the vibroload of the damped device, and the vibration intensity are derived for the optimal nonlinear system     

Instability of solitary waves for generalized Boussinesq equations

An equation of Boussinesq-type of the formu _tt -u _xx +(f(u)+u_xx)_xx=0 is considered. It is shown that a traveling wave may be stable or unstable, depending on the range of the wave's speed of propagation and on the nonlinearity. Sharp conditions to that effect are given.This research is supported in part by NSF Grant DMS 90-23864.     

Axisymmetric stagnation flow obliquely impinging on a moving circular cylinder with uniform transpiration

Laminar stagnation flow, axisymmetrically yet obliquely impinging on a moving circular cylinder, is formulated as an exact solution of the Navier–Stokes equations. Axial velocity is time‐dependent, whereas the surface transpiration is uniform and steady. The impinging free stream is steady with a strain rate k?. The governing parameters are the stagnation‐flow Reynolds number Re=k?a²/2ν, and the dimensionless transpiration S=U₀/k?a. An exact solution is obtained by reducing the Navier–Stokes equations to a system of differential equations governed by Reynolds number and the dimensionless wall transpiration rate, S. The system of Boundary Value Problems is then solved by the shooting method and by deploying a finite difference scheme as a semi‐similar solution. The results are presented for velocity similarity functions, axial shear stress and stream functions for a variety of cases. Shear stresses in all cases increase with the increase in Reynolds number and suction rate. The effect of different parameters on the deflection of viscous stagnation circle is also determined. Copyright © 2010 John Wiley & Sons, Ltd.     

MODAL SYNTHESIS METHOD FOR NORM COMPUTATION OF H∞ DECENTRALIZED CONTROL SYSTEMS （ I ）

When using H∞ techniques to design decentralized controllers for large systems,the whole system is divided into subsystems, which are analysed using H∞ control theorybefore being recombined. An analogy was established with substructural analysis instructural mechanics, in which H∞ decentralized control theory corresponds to substructuralmodal synthesis theory so that the optimal H∞ norm of the whole system corresponds to thefundamental vibration frequency of the whole structure. Hence, modal synthesismethodology and the extended Wittrick-Williams algorithm were transplanted from structuralmechanics to compute the optimal H∞ norm of the control system. The orthogonality and theexpansion theorem of eigenfunctions of the subsystems H∞ control are presented in part(I) of the paper. The modal synthesis method for computation of the optimal H∞ norm ofdecentralized control systems and numerical examples are presented in part (Ⅱ).     

Neural networks based self-learning PID control of electronic throttle

An electronic throttle is a low-power DC servo drive which positions the throttle plate. Its application in modern automotive engines leads to improvements in vehicle drivability, fuel economy, and emissions. In this paper, a neural networks based self-learning proportional-integral-derivative (PID) controller is presented for electronic throttle. In the proposed self-learning PID controller, the controller parameters, K _P , K _I , and K _D are treated as neural networks weights and they are adjusted using a neural networks algorithm. The self-learning algorithm is operated iteratively and is developed using the Lyapunov method. Hence, the convergence of the learning algorithm is guaranteed. The neural networks based self-learning PID controller for electronic throttle is verified by computer simulations.     

A comparative study of crack propagation models for PMMA and PVC

An analysis of examining the validity of a unified approach proposed earlier by the authors for the fatigue crack propagation (FCP) of engineering materials to include PMMA and PVC is described. The proposed formulation has been shown capable of characterizing a diversified range of materials with a master FCP diagram and expressed as da/dN = A(ΔG)^m/(G_c − G_max).An experimental program is undertaken to measure fatigue growth rate with the standard compact tension specimen. The FCP results are for the first instance analysed for each material using the unified formulation. The validity of the formulation for producing a master FCP diagram is verified when the fatigue crack growth rates of the materials are successfully characterized in one master diagram, yielding an excellent coefficient of correlation of 0.993. No such success is attained using a number of conventional FCP laws considered most acceptable to characterize polymeric materials.     

Long-term strength of metals in complex stress state (a survey)

An analytic survey of experimental data and theoretical approaches characterizing the long-term strength of metals in complex stress state is given. In Sections 2 and 3, the results of plane stress tests (with opposite and equal signs of the nonzero principal stresses, respectively) are analyzed. In Section 4, the results of inhomogeneous stress tests (thick-walled tubes under the action of internal pressures and tensile forces) are considered. All known experimental data (35 test series) are analyzed by a criterion approach. An equivalent stress σ _e is introduced as a characteristic of the stress state. Attention is mainly paid to the dependence of σ _e on the principal stresses. Statistical methods are used to obtain an expression for σ _e, which can be used to study various types of the complex stress state. It is shown that for the long-term strength criterion one can use the power or power-fractional dependence of the time to rupture on the equivalent stress. The methods proposed to describe the test results give a good correspondence between the experimental and theoretical values of the time to rupture. In Section 5, the possibilities of complicating the expressions for σ _e by using additional material constants are considered.