Superconvergence of recovered gradients of discrete time/piecewise linear Galerkin approximations for linear and nonlinear parabolic problems

Superconvergent error estimates in l₂(H¹) and l∞(H¹) norms are derived for recovered gradients of finite difference in time/piecewise linear Galerkin approximations in space for linear and quasinonlinear parabolic problems in two space dimensions. The analysis extends previous results for elliptic problems to the parabolic context, and covers problems in regions with nonsmooth boundaries under certain assumptions on the regularity of the solutions. © 1994 John Wiley & Sons, Inc.     

Structural Transformations in Fe/Zr Bi- and Multilayers Induced by Ion-Beam Mixing

The Ar-ion-beam mixing of Fe/Zr bi- and multilayers is studied by conversion electron Mössbauer spectroscopy as a function of ion dose ranging from 1×10¹⁴ to 1×10¹⁷ at./cm². The mixing leads to amorphization of the Fe/Zr system. It was shown that the mixing process in bilayers depends strongly on the thickness of the Zr substrate. The amorphization is much more effective and occurs at lower ion doses in multilayers as compared to corresponding bilayers.     

Luminescence of self-trapped excitons in calcium fluoride under pulsed electron irradiation

Time-resolved pulsed spectroscopy was used to measure the luminescence spectra of calcium fluoride. Characteristic features of the luminescence of self-trapped excitons are discussed. It is shown that various configurations of self-trapped excitons incorporating hole nuclei of a more complex structure, may be formed in CaF₂ crystals. Fiz. Tverd. Tela (St. Petersburg) 40, 1226–1227 (July 1998)     

On the optimization of approximate integration by Monte Carlo methods

We solve the problem of optimization of monte Carlo methods for approximate integration over an arbitrary absolutely continuous measure. We propose a convenient model of Monte Carlo methods which uses the notion of transition probability.     

1.2 Anorganische analyse

Ohne Zusammenfassung     

Improvement of least-squares finite difference method on simulating slider air bearings in head disk systems

Least-squares finite difference (LSFD) method, one of mesh-free methods, is used to solve slider air bearings problem through discritizing the generalized Reynolds equation into nonlinear systems of algebraic equations. Two approximation schemes for the linearization of these equations are presented and compared. And, some new techniques to search supporting points for the reference node in the mesh-free method were proposed and explored. Therefore, these improvements eliminate some potential limitation of the LSFD method previously published and further facilitate its employment in complex slider models. Advanced step slider as an example of negative pressure sliders is simulated and verified using the improved LSFD mesh-free method in head disk systems.     

Coercivity analysis in R17Fe83−χBχ magnets

We studied the coercivity in magnets of composition R₁₇Fe_83−χB_χ (R = Nd, Pr and χ = 8, 30), using measurements of the coercive field H_c, its angular dependence, and the magnetic viscosity coefficient S_v, for temperatures between 4.2 and 500 K. The results are discussed in relation to a model which does not specifically consider the detailed mechanisms involved in magnetization reversal, but which provides information about the magnetic properties in the activation volume v where magnetization reversal is initiated. It is concluded that the ordering temperature in v tends to be slightly smaller than in the bulk and that the room temperature anisotropy in v is not strongly reduced with respect to the bulk value. Finally, a direct evaluation of the dipolar interactions is in good agreement with results obtained from H_c(T).     

Taylor-Galerkin algorithms for viscoelastic flow: Application to a model problem

Coupled and decoupled Taylor-Galerkin algorithms are considered for viscoelastic flow and a model problem—transient startup Poiseuille flow in a channel under a fixed pressure gradient. All algorithms reproduce the steady-state solutions and are stable at high elasticity numbers (E). For a fixed mesh, the coupled and decoupled versions (TGC and TGD) give exceptional time-accuracy at low elasticity numbers [to within O(1%) at E = 1] and reasonable accuracy at high elasticity numbers [to within O(10%) at E = 10, 100]. By definition, the decoupled false-transient scheme (TGF), which uses different time scales for velocity and stress time stepping, provides a poor transient history. Where the main requirement is to compute a steady-state algorithm efficiency is crucial. The TGF scheme attains a steady state between six to eight times faster than does the TGC scheme, and the latter is over twice as fast as the TGD form. © 1994 John Wiley & Sons, Inc.