The asymptotic behaviour of an elastically supported infinite string and an elastic isotropic half plane (in frames of specific asymptotic model) under a moving point load are studied. The main results of this work are uniform asymptotic formulae and the asymptotic profile for the string and the exact solution and uniform asymptotic formulae for a half plane. The crucial assumption for both structures is that the acceleration is sufficiently small. In order to describe asymptotically the oscillations of an infinite string auxiliary canonical functions are introduced, asymptotically analyzed and tabulated. Using these functions uniform asymptotic formulae for the string under constant accelerating and decelerating point loads are obtained. Approximate formulae for the displacement in the vicinity of the point load and the singularity area behind the shock wave using the steady speed asymptotic expansion with additional contributions from stationary points where appropriate are derived. It is shown how to generalise uniform asymptotic results to the arbitrary acceleration case. As an example these results are applied for the case of sinusoidal load speed. It is shown that the canonical functions can successfully be used in the arbitrary acceleration case as well. The graphical comparative analysis of numerical solu- tion and approximations is provided for different moving load speed intervals and values of the parameters. Vibrations of an elastic half plane are studied within the framework of the asymp- totic model suggested by J. Kaplunov et al. in 2006. Boundary conditions for the main problem are obtained as a solution for the problem of a string on the surface of a half plane subject to uniformly accelerated moving load. The exact solution over the interior of the half plane is derived with respect to boundary conditions. Steady speed and Rayleigh wave speed asymptotic expansions are obtained. In the neighborhood of the Rayleigh speed the uniform asymptotic formulae are derived. Some of their interesting properties are discovered and briefly studied. The graphical comparative analysis of the exact solution and approximations is provided for different moving load speed intervals and values of the parameters.

This paper is devoted to the computation of statistical characteristics of the response of discrete vibration systems with a random external excitation. The excitation can act at multiple points and is modeled by a time-shifted random process and its derivatives up to the second order. Statistical characteristics of the response are given by expansions as to the correlation length of a weakly correlated random process which is used in the excitation model. As the main result analytic expressions of some integrals involved in the expansion terms are derived.

The novelty and originality of the work consists in the fact that in order to obtain asymptotic expansions with optimal values of the remainder terms in the zone of large deviations, along with the cumulant method the classical method of characteristic functions has to be used. In addition, when solving the problems stated in the work, other than the well known results in the problems of limit theorems of the probability theory and mathematical statistics, we have to estimate constants. Technically it is frequently rather a complicated task. The results obtained in the work have good opportunities to be applied in probability theory, mathematical statistics, econometric, etc. That is illustrated in the last section of the work in which theorems of large deviations are proved in the summation of weighted random variables with weights as well as discounted limit theorems.

In this thesis we study the asymptotic behaviour of the solution of a discrete time renewal equation depending on a small perturbation parameter. In particular, we construct asymptotic expansions for the solution of the renewal equation and related quantities. The results are applied to studies of quasi-stationary phenomena for regenerative processes and asymptotics of ruin probabilities for a discrete time analogue of the Cramér-Lundberg risk model.

The increasing concentration of CO_(2) has been linked to global warming and changes in climate. Geologic sequestration of CO_(2) in deep saline aquifers is a proposed greenhouse gas mitigation technology with potential to significantly reduce atmospheric emissions of CO_(2). Feasibility assessments of proposed sequestration sites require realistic and computationally efficient models to simulate the subsurface pressure response and monitor the injection process, and quantify the risks of leakage if there is any. This study investigates the possibility of obtaining closed form expressions for spatial distribution of CO_(2) injected in brine aquifers and gas reservoirs. Four new semi-analytical solutions for CO_(2) injection in brine aquifers and gas reservoirs are derived in this dissertation. Both infinite and closed domains are considered in the study. The first solution is an analysis of CO_(2) injection into an initially brine-filled infinite aquifer, exploiting self–similarity and matched asymptotic expansion. The second is an expanding to the first solution to account for CO_(2) injection into closed domains. The third and fourth solutions are analyzing the CO_(2) injection in infinite and closed gas reservoirs. The third and fourth solutions are derived using Laplace transform. The brine aquifer solutions accounted for both Darcyian and non-Darcyian flow, while, the gas reservoir solutions considered the gas compressibility variations with pressure changes. Existing analytical solutions assume injection under constant rate at the wellbore. This assumption is problematic because injection under constant rate is hard to maintain, especially for gases. The modeled injection processes in all aforementioned solutions are carried out under constant pressure injection at the wellbore (i.e. Dirichlet boundary condition). One major difficulty in developing an analytical or semi-analytical solution involving injection of CO_(2) under constant pressure is that the flux of CO_(2) at the wellbore is not known. The way to get around this obstacle is to solve for the pressure wave first as a function of flux, and then solve for the flux numerically, which is subsequently plugged back into the pressure formula to get a closed form solution of the pressure. While there is no simple equation for wellbore flux, our numerical solutions show that the evolution of flux is very close to a logarithmic decay with time. This is true for a large range of the reservoir and CO_(2) properties. The solution is not a formation specific, and thus is more general in nature than formation-specific empirical relationships. Additionally, the solution then can be used as the basis for designing and interpreting pressure tests to monitor the progress of CO_(2) injection process. Finally, the infinite domain solution is suitable to aquifers/reservoirs with large spatial extent and low permeability, while the closed domain solution is applicable to small aquifers/reservoirs with high permeability.

In this thesis the author presents a history of non-convergent series which, in the past, played an important role in mathematics. Euler's formula, Stirling's series and Poincare's theory are examined to show the development of asymptotic series, a subdivision of divergent series.

Indiana University-Purdue University Indianapolis (IUPUI) / We consider the behavior of zeros of random polynomials of the from \begin{equation*} P_{n,m}(z) := \eta_0\varphi_m^{(m)}(z) + \eta_1 \varphi_{m+1}^{(m)}(z) + \cdots + \eta_n \varphi_{n+m}^{(m)}(z) \end{equation*} as \( n\to\infty \), where \( m \) is a non-negative integer (most of the work deal with the case \( m =0 \) ), \( \{\eta_n\}_{n=0}^\infty \) is a sequence of i.i.d. Gaussian random variables, and \( \{\varphi_n(z)\}_{n=0}^\infty \) is a sequence of orthonormal polynomials on the unit circle \( \mathbb T \) for some Borel measure \( \mu \) on \( \mathbb T \) with infinitely many points in its support. Most of the work is done by manipulating the density function for the expected number of zeros of a random polynomial, which we call the intensity function.

京都大学 / 新制・課程博士 / 博士(理学) / 甲第23675号 / 理博第4765号 / 新制||理||1683(附属図書館) / 京都大学大学院理学研究科数学・数理解析専攻 / (主査)教授 坂上 貴之, 教授 泉 正己, 教授 國府 寛司 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

La propagation d'une fissure chargée par un écoulement de fluide visqueux est un phénomène complexe où la compréhension des phénomènes mécaniques mis en jeu en pointe de fissures reste encore partielle. C'est le cas de la zone de décollement entre le solide et le fluide qui apparaît pour un certain choix de débit d'injection, de viscosité du fluide et de ténacité du matériau. Cette thèse propose une modélisation simplifiée de ce problème d'interaction fortement couplé. Dans un premier chapitre, on étudie un modèle simplifié unidimensionnel de film élastique collé sur un substrat rigide et on considère une injection de fluide visqueux entre le film et le substrat. On suppose que la propagation de la fissure est régie par la loi de Griffith. On néglige l'existence du retard possible entre le fluide et le solide et on choisit une loi de comportement non-linéaire pour le fluide visqueux. A partir d'une analyse asymptotique pour une faible viscosité, on établit une solution approchée du problème. On montre que le champ pression est singulier en pointe de fissure et on montre l'influence du débit d'injection sur la cinétique du trajet de fissuration. Dans le deuxième chapitre on propose de prendre en compte l'existence de la zone de décollement en modifiant la formulation du modèle et en le réécrivant sous la forme d'un problème d'optimisation en temps discret où les zones de décollement font partie des inconnues du problème. On valide la formulation proposée sur l'exemple analytique de l'écrasement d'une goutte par une barre rigide. On montre ensuite que cette formulation et l'algorithme lié à son implémentation sont capables de gérer l'évolution de l'écrasement de plusieurs gouttes de forme quelconque en capturant correctement les phase d'étalement des gouttes ainsi que de leur coalescence. On étend ensuite cette formulation au cas de l'écrasement d'une goutte par un film élastique. Dans le dernier chapitre, on examine la validité de l'hypothèse de lubrification utilisée en fracturation hydraulique. A l'aide de la méthode de développement asymptotique, on construit une équation de Reynolds régularisée avec des termes de gradient supérieur tenant compte de la variation spatiale de la hauteur des parois. On compare alors le comportement des champs de pression donnés par les équations de Reynolds classique et régularisée sur des exemples d'écoulement entre des conduits de formes multiples. / The crack evolution under a viscous fluid action is a complex phenomenon where the understanding of the mechanical phenomena near the crack tip is still largely limited. This is the case for the lag between the solid and the fluid front propagation which appears for some configurations of injection rate, fluid viscosity and material toughness. This thesis proposes a simplified model for this strongly coupled interaction problem.The first chapter studies a simplified one-dimensional model of a elastic film bonded to a rigid substrate. We consider a viscous fluid injection between the film and the substrate. The crack propagation is assumed to follow the Griffith's law. The existence of the lag is neglected and a non-linear behavior law is chosen for the viscous fluid. Using an asymptotic analysis, an approximate solution is established for the low viscosity case. It is shown that the pressure field diverges at the crack tip and that the kinetics of the crack is influenced by the injection rate. The second chapter proposes to take into account the existence of the lag by modifying the model formulation and rewriting it as a discrete time optimisation problem where the delamination zones are part of the unknowns of the problem. This formulation is validated for the analytical example of a drop crushed by a rigid bar. It is shown that this formulation and its implementation can manage the evolution of several drops of any shape and correctly captures the drops spreading and coalescence. This formulation is then extended to the case of a drop crushed by an elastic film. In the last chapter, the validity of the lubrication hypothesis is examinated. Using an asymptotic analysis, a regularized Reynolds equation is constructed with higher gradient terms taking into account the spatial variation of the walls height. A comparison between the pressure fields behaviour given by the classical and the regularized Reynolds equation is shown for different conducts.

La premième partie a pour but l’établissement d’un développement asymptotique pour la solution du problème de Stokes avec une petite perturbation du domaine. Dans ce travail, nous avons appliqué la théorie du potentiel. On a écrit les solutions du problème non-perturbé et du problème perturbé sous forme des opérateurs intégraux. En calculant la différence, et en utilisant des propriétés liées aux noyaux des opérateurs on a établi un développement asymptotiquede la solution.L’objectif principal de la deuxième partie de ce rapport est de déterminer les termes d’ordre élevé de l’expansion asymptotique des valeurs propres et fonctions propres pour l’opérateur de Stokes dues aux changements d’interface de l’inclusion. Dans la troisième partie, nous proposons une méthode pour l’évaluation des integrales singulières provenant de la mise en oeuvre de la méthode des éléments finis de frontière en électromagnetisme. La méthodeque nous adoptons consiste en une réduction récursive de la dimension du domained’intégration et aboutit à une représentation de l’intégrale sous la forme d’une combinaison linéaire d’intégrales mono-dimensionnelles dont l’intégrand est régulier et qui peuvent s’évaluer numériquement mais aussi explicitement. Pour la discrétisation du domaine, destriangles plans sont utilisés ; par conséquent, nous évaluons des intégrales sur le produit de deux triangles. La technique que nous avons développée nécessite de distinguer entre diverses configurations géométriques. / This thesis contains three main parts. The first part concerns the derivation of an asymptotic expansion for the solution of Stokes resolvent problem with a small perturbation of the domain. Firstly, we verify the continuity of the solution with respect to the small perturbation via the stability of the density function. Secondly, we derive the asymptotic expansion ofthe solution, after deriving the expansion of the density function. The procedure is based on potential theory for Stokes problem in connection with boundary integral equation method, and geometric properties of the perturbed boundary. The main objective of the second part on this report, is to present a schematic way to derive high-order asymptotic expansions for both eigenvalues and eigenfunctions for the Stokes operator caused by small perturbationsof the boundary. Also, we rigorously derive an asymptotic formula which is in some sense dual to the leading-order term in the asymptotic expansion of the perturbations in the Stokes eigenvalues due to interface changes of the inclusion. The implementation of the boundary element method requires the evaluation of integrals with a singular integrand. A reliable andaccurate calculation of these integrals can in some cases be crucial and difficult. In the third part of this report we propose a method of evaluation of singular integrals based on recursive reductions of the dimension of the integration domain. It leads to a representation of the integralas a linear combination of one-dimensional integrals whose integrand is regular and that can be evaluated numerically and even explicitly. The Maxwell equation is used as a model equation, but these results can be used for the Laplace and the Helmholtz equations in 3-D.For the discretization of the domain we use planar triangles, so we evaluate integrals over the product of two triangles. The technique we have developped requires to distinguish between several geometric configurations.