Simon P. Walker

IL-13 production by allergen-stimulated T cells is increased in allergic disease and associated with IL-5 but not IFN-gamma expression.

Interleukin‐13 (IL‐13) shares many, but not all, of the properties of the prototypic T‐helper type 2 (Th2) cytokine IL‐4, but its role in allergen‐driven T‐cell responses remains poorly defined. We hypothesized that allergen stimulation of peripheral blood T cells from patients with atopic disease compared with non‐atopic controls results in elevated IL‐13 synthesis in the context of a ‘Th2‐type’ pattern. Freshly isolated peripheral blood mononuclear cells (PBMC) obtained from sensitized atopic patients with allergic disease, and non‐atopic control subjects, were cultured with the allergens Phleum pratense (Timothy grass pollen) or Dermatophagoides pteronyssinus (house dust mite) and the non‐allergenic recall antigen Mycobacterium tuberculosis purified protein derivative (PPD). Supernatant concentrations of IL‐13, along with IL‐5 and interferon‐γ (IFN‐γ) (Th2‐ and Th1‐type cytokines, respectively) were determined by enzyme‐linked immunosorbent assay (ELISA). Allergen‐induced IL‐13 and IL‐5 production by T cells from patients with allergic disease was markedly elevated (P=0·0075 and P=0·0004, respectively) compared with non‐atopic controls, whereas IFN‐γ production was not significantly different. In contrast to allergen, the prototypic Th1‐type antigen M. tuberculosis PPD induced an excess of IFN‐γ over IL‐13 and IL‐5 production, and absolute concentrations of cytokines were not affected by the presence or absence of atopic disease. Addition of exogenous recombinant IFN‐γ or IL‐12, cytokines known to inhibit Th2‐type responses, significantly inhibited allergen‐driven production of both IL‐13 and IL‐5, but not T‐cell proliferation, whereas exogenous IL‐4 did not significantly affect production of IL‐13 or IL‐5. We conclude that allergen‐specific T cells from atopic subjects secrete elevated quantities of IL‐13 compared with non‐atopic controls, in the context of a Th2‐type pattern of cytokine production.

Parallel computation of time-domain integral equation analyses of electromagnetic scattering and RCS

Electromagnetic scattering analyses for radar cross-section (RCS) prediction present very large computational demands. Processing on parallel machines will contribute to increasing the size of problem tractable. We present a parallel implementation of the time-domain integral equation method. The issues and options in its parallelization are identified, and domain decomposition strategies to suit these are implemented. Good parallelization is exhibited, with the most costly parts of the algorithm displaying essentially linear speedup. Demonstrations using the Cray T3D are given with, for example, results on the /spl sim/12 wavelength NASA almond obtained in /spl sim/30 min per illumination angle, using 256 processors.

Scattering analysis via time-domain integral equations: methods to reduce the scaling of cost with frequency

Methods are described by which very large reductions in the cost of time-domain integral-equation scattering computations can be achieved, associated with a reduction in the scaling of operation count with frequency from the fifth to the fourth, or even third, power. There is an accompanying reduction in storage scaling from the fourth to a third, or even second, power of frequency. The additional physical approximation inherent in the methods can be implemented, and the associated degradation in accuracy assessed and the accompanying cost savings estimated, without implementation of the method itself. This is done in this paper. As an example of the results, the head-on backscatter from a four-wavelength NASA almond could be obtained /spl sim/1 dB less accurately, at a cost more than an order of magnitude lower. This factor by which costs are reduced would increase with problem size. The issues and difficulties involved in implementation of both the fourth- and third-power approaches are identified and discussed.

The complex bi-conjugate gradient solver applied to large electromagnetic scattering problems, computational costs, and cost scalings

The complex bi-conjugate gradient iterative method is applied to an isoparametric boundary integral equation formulation for frequency-domain electromagnetic scattering problems. It is demonstrated to work well on large and geometrically complex examples, including a 20 wavelength slender dipole, the NASA almond, and a resonant cavity. On such problems, with asymmetric curvilinear irregular meshes and nontrivial geometries, the number of iterations required seems to increase rather more than linearly with body size, indicating an overall /spl sim/sixth power cost scaling. This scaling is essentially as for direct methods, but with costs still a small fraction of the direct approach. A method is proposed for the selection of a termination condition designed to avoid seeking the approximate answer too precisely; it typically permits a further halving of costs.

Developments in time-domain integral-equation modeling at Imperial College

This paper reviews the status of time-domain integral-equation modeling. It concentrates on work concerning the development of the method at Imperial College, done since the review of Gomez Martin et al. (1992). A curvilinear, isoparametric approach, incorporating Gaussian quadrature, with special treatment of singular and hypersingular integrals, is described. The approach is implicit, providing stable predictions without recourse to the averaging schemes found necessary with explicit treatments. The issues of parallelization, hybridization to time-domain optical methods, and application to dielectric and lossy targets are addressed.

Pellet-clad mechanical interaction: pellet-clad bond failure and strain relief

The effects of pellet-clad mechanical interaction would be expected to be particularly severe in the presence of bonding between the fuel and the cladding. However, such bonding is observed far more frequently than is corresponding cladding damage. It has recently been shown that the radial stress in the bond during power changes is very large and tensile, and thus likely to cause failure of the bond. In this paper the likely azimuthal extent of this de-bonding is considered, and the relief of hoop stress which this offers is assessed. It is shown that the magnitude of this relief is such as to provide an explanation of the low cladding failure rate observed.

Applications of differential forms to boundary integral equations

In this paper we discuss the application of differential forms to integral equations arising in the study of electromagnetic wave propagation. The usual Stratton-Chu integral equations are derived in terms of differential forms and corresponding Galerkin formulations are constructed. All numerical schemes require the specification of basis functions and the use of differential forms provides a very general method for the construction of arbitrary order basis functions on curvilinear geometries. It is noted that the lowest order approximations on flat geometries reduce to forms essential equivalent to the standard Rao-Wilton-Glisson functions. The effect on accuracy is investigated for electric field integral equation and magnetic field integral equation formulations for a range of bases. Hierarchical classes of functions are also developed, as are transition elements useful in p-adaptive schemes where variable orders of approximation are sought.

TERMINATION CRITERIA IN ITERATIVE SOLUTION OF LARGE SCATTERING PROBLEMS USING INTEGRAL EQUATION METHODS

Iterative methods are used increasingly for solution of the extremely large matrix equations generated by integral equation analysis of multi-wavelength frequency domain scattering. Although much cheaper than direct methods, the matrix solution remains the dominant cost, and is very costly. The criterion adopted for termination of the iteration can have a marked effect on this cost. We show that for large scattering problems a robust and rational prior choice of termination criterion can be made, based only on discretization. This allows confident use of a much larger termination residual than those commonly used, with consequent cost reduction.

An Evaluation of the RPI Model for the Prediction of the Wall Heat Flux Partitioning in Subcooled Boiling Flows

Subcooled boiling flows are encountered in most nuclear reactor configurations. Wall heat flux partitioning models form an integral part of the subcooled boiling formulations in CFD codes. These models attempt to describe the flow of heat from the wall into the fluid by dividing it according to several mechanisms of heat transfer. This work presents a one-dimensional evaluation of the wall heat flux partitioning model of Kurul and Podowski, also referred to commonly as the RPI model, which is used in the state-of-the-art codes of today. This model was assessed against the measurements of Okawa et al. for a vertically upward subcooled boiling flow of water at near atmospheric pressure. Although the predictions showed good agreement with the measured wall temperatures, significant discrepancies were observed in the predictions of the constituent sub-models that comprised the overall model. Prospects for improvement are discussed.Copyright

The rewetting of PWR fuel cladding during post-LOCA reflood: a proposed physical explanation for the micro-scale high-frequency sputtering observed

Following a loss of coolant accident in a light water reactor, reflooding of the core to quench overheating fuel is urgent. The process of quenching hot metal with liquid water has been much studied, in large part because of its great importance in this context. There is good experimental evidence such as quenching occurs via multiple micro-scale wettings of the surface, followed by vapour-driven ejection of liquid water. It is only after some tens of such wettings that permanent wetting seems to take place. In this article, we investigate this process, and put forward a physical mechanism that seems to explain this cyclical behaviour. The mechanism involves the need for some time to elapse following wetting of a hot body for heat to be conducted from within the depths of the body into the liquid water, such as to form a layer of water hot and deep enough to accommodate a critically sized bubble and thus cause ejection of the water film. Simple one-dimensional calculations using this model produce results that are in broad accord with experimental observations.