Computing Systems in Engineering最新文献

Thermal-mechanical analyses of soils are complicated by the porous nature of the material. Macroscopically soils exhibit an anisotropic, inelastic, hardening (and softening), time- and temperature-dependent behavior. To further complicate matters, the thermal properties of soils are not as well known as those for other materials such as metals. Several important problems have recently arisen that necessitate the realistic prediction of thermal-mechanical behavior of soils. The analysis of such problems requires a general but practical methodology that accounts for not only material non-linearities and thermo-mechanical coupling, but also for time-dependent characteristics of soils.

This objective has been realized by extending a generalized bounding surface model, originally developed for isothermal analyses of saturated cohesive soils, to consider temperature effects. Besides accounting for thermal-mechanical coupling and for the inelastic nature of soils, this model includes time-dependent behavior. This latter aspect appears to be a novel proposition in thermo-mechanical modeling of soils.

The emphasis of the present paper is on the numerical implementation and solution of the aforementioned model for thermo-mechanical analysis of saturated cohesive soils. It is shown that the consideration of time effects in the analysis introduces no additional complexity into potential algorithms used in the solution of coupled three-field problems.

This paper deals with the development and application of reliable, creative and efficient computational tools for the structural optimisation of variable thickness axisymmetric and prismatic shells and folded plates using computer-aided analysis and design procedure. The problem of finding optimal forms and thickness variations for such structures is solved by integrating computer aided geometry modelling tools, automatic mesh generation, structural analysis, sensitivity evaluation and mathematical programming methods. The shape and thickness variation of the structures are defined using parametric cubic splines and the structural analysis is carried out with either finite element or finite strip methods in which Mindlin-Reissner assumptions are adopted. In static situations, the composition of the strain energy is monitored during the optimisation process to obtain insight into the energy distribution for the optimum structures. This allows us to demonstrate that, in the majority of cases, the optimum shells are membrane energy dominated as might be expected. For the vibrating structures, the mode shapes of the initial and optimum solutions are presented. A set of carefully defined, unambiguous benchmark examples is presented and studied with independent verification to test the various features of the structural optimisation process.

Assumed Gaussian and β probability density functions (PDFs) for temperature are used with series expansions of the reaction-rate coefficients to compute the mean reaction-rate coefficients in a turbulent, reacting flow. The series-expansion/assumed PDF approach does not require any numerical integration, which substantially reduces computational cost with little loss of accuracy. An assumed multivariate β PDF for species is investigated for use in modeling the interaction of species fluctuations and chemical combustion. The multivariate β PDF is initially evaluated through a parametric study. Results of the parametric study indicate that species fluctuations can increase or decrease the magnitude of the species production term, depending on the type of reaction. The models are then tested on a two-dimensional high-speed turbulent reacting hydrogen-air mixing layer. For the conditions tested the numerical simulations indicate that the net effect of species fluctuations is to reduce the mean species production rate.

Multi-dimensional flows have been calculated with complex chemical kinetics for hydrocarbon fuels. The flows include both premixed and diffusion flames, and there is a wide variation of both thermodynamic and chemical reaction conditions. The numerical methods have included an implicit block solver of the ADI type for the transport equations, and the low Mach number limit has been taken for the calculation of the pressure field. The applications consisted of a steady premixed methane flame about a hot sphere, and the time-dependent ignition of a methanol droplet in air. The methane calculation has been compared to a detailed stagnation point calculation, and the results compare favorably. The methanol droplet flow includes the gas, liquid, and interface conditions, and the calculation has been performed under diesel engine thermodynamic conditions. The methane flame was calculated with only temperature change error control and this form of error control was tested first on zero and one-dimensional models with the same chemistry. With this type of error control the time dependent calculation was stable, and steady state was approached rapidly. These calculations of complex chemical kinetics required substantial computer time, but the computational times are not large when compared to the potential of massively parallel computers. The methods used in the paper can be extended to parallel machines in a straight forward manner.

This paper presents an accurate and practical technique for coupling shell element models to three-dimensional continuum finite element models. The compatibility between these two types of formulations is enforced by degenerating a continuum element through kinematic constraints compatible with shell deformations. Two formulations of two-dimensional/three-dimensional transition elements are presented. The first and simplest formulation is based on the Mindlin-Reissner plate assumptions, and is found to perform well in a variety of problems involving the analysis of geometrically linear/non-linear laminated structures. The second formulation is based on a higher-order shell theory that allows stretching in the through-the-thickness direction. This additional freedom virtually eliminates the interlaminar normal stress boundary layer that can form in lower-order transition elements. Finally, the coupling of two-dimensional to three-dimensional subdomains is enriched with the use of an interface element, which can be used in conjunction with either transition formulation. The interface element improves the efficiency of the solid-to-shell transition modeling scheme by allowing the independent selection of optimal mesh sizes in the shell and the three-dimensional regions of the model.

Since direct numerical simulation of the Navier-Stokes plus combustion chemistry equations will not be practical in the foreseeable future, models are required for the parameter range of practical interest, i.e. high Reynolds Numbers and a wide range of Damkohler Numbers. Models based on the notion of a flamelet are not appropriate when the turbulence intensity is much greater than the laminar flame speed, but a stochastic model based on the joint PDF of velocity and composition is promising. If the velocity field and inhomogeneities in physical space are ignored in the joint PDF equation, the “Partially Stirred Reactor” or PaSR model is obtained. The PaSR model has recently been studied in detail. Full chemical schemes are computationally tractable. Because the composition PDF has a large number of dimensions (e.g. N_s > 20 for methane), finite-element/volume techniques are not viable, but particle-tracking Monte-Carlo algorithms work well. An enabling feature of the PaSR is that, with the IEM scalar mixing sub-model, it is well suited to parallel computers. The PaSR can describe the effect of turbulence (coupled to a full kinetic scheme) on combustion, including the behavior of emissions such as NO_x and CO, of minor species such as free radicals, and the ignition-extinction bifurcation.

This paper demonstrates the use of automatic differentiation in solving finite element problems with random geometry. In the area of biomechanics, the shape and size of the domain is often known only approximately. Stochastic finite element analysis can be used to compute the variability in the structural response as a result of variability in the shape of the structural domain. Automatic differentiation can be used to compute the shape sensitivites accurately and effortlessly. Unlike randomness in material properties, the response variability can be the same as or greater than the variability in the input. When both the Young's modulus and geometry are random, it is likely that randomness in geometry will dominate randomness in Young's modulus.