Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering, 1982. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. / Includes bibliographical references. / by Vasantha Mohan Narendran. / M.S.

A large data set comprising normal-mode eigenfrequencies, quality factors and splitting functions, Earth's mass and moment of inertia, surface-wave phase anomalies and dispersion curves, body-wave arrivals and traveltime curves, as well as long-period waveforms is inverted to obtain the distribution of elastic properties, shear attenuation and density in the Earth's interior. We address three fundamental aspects of global seismology by reconciling and modeling data sets with several methodological improvements, such as accounting for radial and azimuthal anisotropy, development of better methods for crustal corrections, and devising novel regularization and parameterization schemes. In the first contribution, we incorporate normal-mode splitting functions with other seismological data sets to examine the variation of anisotropic shear-wave velocity in the Earth's mantle. Our preferred anisotropic model, S362ANI+M, has strong isotropic velocity anomalies in the transition zone while the anisotropy is restricted to the upper 300~km in the mantle. When radial anisotropy is allowed throughout the mantle, large-scale anisotropic patterns are observed in the lowermost mantle with v_SV > v_SH beneath Africa and South Pacific and v_SH > v_SV beneath several circum-Pacific regions. However, small improvements in fits to the data on adding anisotropy at depth leave the question open on whether large-scale radial anisotropy is required in the transition zone and in the lower mantle. We demonstrate the utility of mode-splitting data in reducing the tradeoffs between even-degree variations of isotropic velocity and anisotropy in the lowermost mantle. We then devise a methodology to detect seismological signatures of chemical heterogeneity using scaling relationships between shear velocity, density and compressional velocity in the Earth's mantle. Several features reported in earlier tomographic studies persist with the inclusion of new and larger data sets; anti-correlation between bulk-sound and shear velocities in the lowermost mantle as well as an increase in velocity scaling (nu=dlnv_S/dlnv_P) with depth in the lower mantle are found to be robust. Many spheroidal and toroidal modes are largely incompatible with perfect correlations between density and shear-velocity variations in the lowermost mantle. A way to fit concurrently the various data sets is by allowing independent density perturbations in the lowermost mantle. Our preferred joint model consists of denser-than-average anomalies (~1% peak-to-peak) at the base of the mantle roughly coincident with the low-velocity superplumes. The relative variation of shear velocity, density and compressional velocity in this study disfavors a purely thermal contribution to heterogeneity in the lowermost mantle. In the third contribution, we introduce an approach to construct a 1-D reference model that is consistent with crustal heterogeneities and various asphericities in the Earth's mantle. We demonstrate that the crust contributes substantially to fundamental-mode dispersion curves when the nonlinear effects of its thickness and velocity variations are taken into consideration. We apply appropriate crustal corrections and perform several iterations to converge to our preferred radial model NREM1D, which is anisotropic in the upper mantle and smooth across the 220-km discontinuity for all physical parameters. Radial anisotropy in the shallowest mantle, with a maximum at ~150~km depth, is required to fit global averages of fundamental-mode Rayleigh and Love wave dispersion (25--250s). NREM1D also predicts arrival times of major mantle and core phases in agreement (+/- 0.5s) with a recent isotropic velocity model that was optimized for earthquake location. The new reference Earth model NREM1D introduced here is easily extendable due to its modular construction as a linear combination of radial basis functions and can be used for earthquake location, spherical-earth normal mode calculations, and as a starting model in studies of lateral heterogeneity.

This dissertation consists of three chapters. Chapter 1, "Precision of Communication in Coordination Games of Regime Change,'' is in the field of macroeconomics and economics of information. I study a model of regime change in which the government can communicate with different levels of precision as a function of the underlying fundamentals. In the model, higher precision of communication corresponds to a lower dispersion of private information among market participants. I compare a policy of an uncommitted government, which chooses the precision of communication after it learns the realization of fundamentals, to a policy of a committed government, which commits to a state-dependent policy before it learns the realization of fundamentals. I find that an uncommitted government communicates imprecisely for weak fundamentals and precisely for strong fundamentals. In contrast, a committed government communicates precisely for weak fundamentals and imprecisely for strong fundamentals. Consequently, a committed government saves its regime more often than an uncommitted one. An uncommitted government can benefit from a rule that enforces constant precision of communication. Chapter 2, "Multiple Equilibria in Global Games with Varying Quality of Information,'' is a follow up chapter on Chapter 1. I show that global game models can have multiple equilibria if the quality of information available to agents varies with the state of economic fundamentals. First, I construct two examples that illustrate why the quality of information may vary and show that the corresponding information structures support several equilibria. Second, I construct an information structure that supports a given number of equilibria. Very different equilibria can exist simultaneously, even if agents' quality of information is arbitrarily high. The set of possible equilibria can be very similar to the set of equilibria under complete information, even when agents are uncertain about the state of fundamentals and beliefs of other agents. My results have practical implications for the disclosure of information by governments and for our ability to predict the outcome of currency attacks or debt runs based on economic fundamentals. Chapter 3, "Long-Run Price Elasticity of Trade and the Trade-Comovement Puzzle,'' is in the field of international macroeconomics and is coauthored with Lukasz Drozd and Jaromir Nosal. What role do international trade linkages play in transmitting shocks across borders? Analytically, we demonstrate that in a broad class of open economy macroeconomic models, shock transmission crucially depends on dynamic properties of trade elasticity---which is rarely modeled explicitly in business cycle theory. We illustrate the quantitative relevance of this point by exploring the well documented link between trade and comovement in the cross-section of countries, and by relating our theoretical findings to those in the literature. We find that dynamic elasticity does indeed affect the findings in a quantitatively significant way. Hence, our paper advocates for using dynamic elasticity models in contexts that evaluate international business cycle theory vis-a-vis data on cross-country variation of business cycle moments.

This dissertation investigates the distributed damage effect on Progressive Collapse of structures highlighted by applications on the nonlinear static and dynamic behavior of buildings, and contributes to the theoretical development of the Variability Response Function concept and its applicability extension in two-dimensional elasticity stochastic problems. Part I of this dissertation focuses on the recently emerging research field of Progressive Collapse of structures. The alternate load path method has so far dominated the field of progressive collapse of structures; in order to assess the resilience of structural systems, the concept of the removal of a key element is utilized as a means of damage introduction to the system. Recent studies have indicated that the complete column loss notion is unrealistic and unable to describe a real extreme loading event, e.g. a blast, that will introduce damage to more than one elements in its vicinity. This dissertation presents a new partial distributed damage method (PDDM) for steel moment frames, by utilizing powerful finite element computational tools that are able to capture loss of stability phenomena. Through the application of a damage index δj and the investigation of damage propagation, it is shown that the introduction of partial damage in the system can significantly modify the collapse mechanisms and overall affect the response of the structure. Subsequently, Part I elaborates on the distributed column damage effect on Progressive Collapse vulnerability in steel buildings exposed to an external blast event. Recent terrorist attacks on civil engineering infrastructure around the world have initiated extensive research on progressive collapse analysis of multi-story buildings subjected to blast loading. The widely accepted alternate load path method is a threat-independent method that is able to assess the response of a structure in case of extreme hazard loads, without the consideration of the actual loads occurring. Such simplification offers great advantages but at the same time fails to incorporate the role of a wider damaged area into the collapse modes of structures. To this end, the investigation of damage distribution on adjacent structural members induced by blast loads is considered critical for the evaluation of structural robustness against abnormal loads that may initiate progressive collapse. This dissertation presents detailed 3D nonlinear finite element dynamic analyses of steel frame buildings in order to examine the spatially distributed response and damage to frame members along the building exterior facing an external blast. A methodology to assess the progressive collapse vulnerability is also proposed, which includes four consecutive steps to simulate the loading event sequence. Three case studies of steel buildings with different structural systems serve as examples for the application of the proposed methodology. A high-rise (20-story) building is firstly subjected to a blast load scenario, while the complex 3D system results in the heavily impacted region are compared with individual column responses (SDOF) obtained from a simplified analytical approach consistent with current design recommendations. Parameters affecting the spatially distributed pressure and response quantities are identified, and the sensitivity of the damage results to the spatial variation of these parameters is established for the case of the 20-story building. Subsequently, two typical mid-rise (10-story) office steel buildings with identical floor plan layout but different lateral load resisting systems are examined; one including perimeter moment resisting frames (MRFs) and one including interior reinforced concrete (RC) rigid core. It is shown that MRFs offer a substantial increase in robustness against blast events, and the role of interior gravity columns identified as the `weakest links'\ of the structural framing is discussed. Part II of this dissertation focuses on the development of Variability Response Functions for apparent material properties in 2D elasticity stochastic problems. The material properties of a wide range of structural mechanics problems are often characterized by random spatial fluctuations. Calculation of apparent properties of such randomly heterogeneous materials is an important procedure, yet no general method besides Monte Carlo simulation exists for evaluating the stochastic variability of these apparent properties for structures smaller than the representative volume element (RVE). In this direction, the concept of Variability Response Function (VRF) has been proposed as a means to capture the effect of stochastic spectral characteristics of uncertain system parameters modeled by homogeneous stochastic fields on the uncertain response of structural systems, without the need for computationally expensive Monte Carlo simulations. Recent studies have formally proved the existence of VRF for apparent properties for statically determinate linear beams through elastic strain energy equivalence of the heterogeneous and equivalent homogeneous bodies, while a Monte-Carlo based methodology for the generalization of the VRF concept to statically indeterminate beams has been recently developed. In this dissertation, the VRF methodology of apparent properties is extended to two-dimensional elasticity stochastic problems discretized on a finite element domain, in order to analytically formulate a VRF that is independent of the marginal distribution and spectral density function of the underlying random heterogeneous material property field (it depends only on the boundary conditions and deterministic structural configuration). Representative examples that illustrate the approach include two-dimensional plane stress problems and underline the dependence of the VRFs on scale, shape and aspect ratio of the finite elements.

Based on the expression specifying the deformation of a multi-component object, a component-based condensation method is developed. This further reduces the size of the matrix to be inverted from the total number of unknown displacements to the number of unknown displacements with changing boundary condition. To speed up the construction of matrices, a maximal matrix technology is proposed. By categorizing the changes in boundary conditions, three fast update strategies on matrix inverse are introduced. Based on the maximal matrix technology and the matrix inverse update strategy, eight easily-formed characteristic matrices are defined to enhance the computation speed further. / In this thesis, an algorithm is developed for simulating the deformation of multiple objects with different material properties using the boundary element method. By tessellating the surface of a geometric model into elements, classifying all the element nodes into different groups with different attributes, and partitioning the stiffness matrix into several sub-matrices according to these attributes, a compact expression about the unknown variables is deduced. In this expression, the dimension of the system matrix has been effectively reduced compared with the traditional method. This expression shows that the deformation of a multi-component object can be simulated in a way similar to that of a single-component object. / Research on the real-time deformation of elastic models has captured wide attention and gained considerable achievement in the past two decades. Most related works focus on developing efficient ways to simulate the behavior of a single-component elastic object. However, objects are usually made up of multiple components with different material properties in practice. It is thus essential to develop efficient techniques for modeling objects which are composed of more than one material. / To make the proposed accelerated algorithm more applicable, a method for simulating the deformation of multi-component models with non-matching interfaces is developed. By applying the interpolation and extrapolation methods, the displacement data can be transferred between non-conforming interfaces. With the application of the energy conservation principle, a relationship between the internal forces on different surfaces can also be established. Together with the force equilibrium conditions and displacement compatibility conditions over the common faces of objects, the deformation of models composed of multi-material components with non-matching interfaces can be simulated. During the application of the linear interpolation method, when the mesh densities on the interfaces of the neighboring components are not the same, unexpected phenomena arise in the simulation process because of this disparity. A traction super-imposition method is adopted to enforce the force constraints on the interface. Experiments showed that this approach produces the correct results. / Zhou, Aifang. / "August 2007." / Adviser: Keh Chuen Hui. / Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 1299. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (p. 144-155). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract in English and Chinese. / School code: 1307.

When deploying software engineering applications in the cloud there are two similar software components used. These are Virtual Machines and Containers. In recent years containers have seen an increase in popularity and usage, in part because of tools such as Docker and Kubernetes. Virtual Machines (VM) have also seen an increase in usage as more companies move to solutions in the cloud with services like Amazon Web Services, Google Compute Engine, Microsoft Azure and DigitalOcean. There are also some solutions using auto-scaling, a technique where VMs are commisioned and deployed to as load increases in order to increase application performace. As the application load decreases VMs are decommisioned to reduce costs. In this thesis we implement and evaluate auto-scaling policies that use both Virtual Machines and Containers. We compare four different policies, including two baseline policies. For the non-baseline policies we define a policy where we use a single Container for every Virtual Machine and a policy where we use several Containers per Virtual Machine. To compare the policies we deploy an image serving application and run workloads to test them. We find that the choice of deployment strategy and policy matters for response time and error rate. We also find that deploying applications as described in the methodis estimated to take roughly 2 to 3 minutes.

La sismique et la sismologie sont des moyens puissants pour comprendre la croûte terrestre.Ces deux méthodes reposent notamment sur une compréhension approfondie de la propagation des ondes sismiques dans des milieux sédimentaires saturés en fluides.Ce travail a pour but de comprendre les effets statique et dynamique du fluide sur la réponse élastique de roches clastiques saturées.Deux points spécifiques de l'interaction fluide-roche sont étudiées: (i) l'intéraction physico-chimique, le « shear weakening », affectant la réponse élastique de la roche; et (ii) l'interaction mécanique, le « frequency effect », induisant une dépendance des propriétés élastiques à la fréquence de mesure.Deux types de roches sont étudiés: les grès et les calcaires.Ces échantillons de roche sont sélectionnés pour leurs propriétés isotropes et leur forte concentration en un minéral dominant: le quartz pour les grès et la calcite pour les carbonates.Le phénomène de « shear weakening » est d’abord étudié pour de très faibles saturations en eau afin de tester l’effet de l'adsorption.Aucun affaiblissement n’est mesuré dans les carbonates, au contraire un affaiblissement élastique global est observé dans certains grès : Les modules de cisaillement et d’incompressibilité sont également affectés.L'effet ne semble pas provenir d'une différence intrinsèque entre les minéraux de quartz et de calcite, mais d’une différence microstructurale entre roches. Un modèle micromécanique est développé, montrant que les deux paramètres clef sont le caractère granulaire et le degré de cimentation de la roche.Le même résultat est obtenu pour les compressibilités mesurées lors des saturations totales en eau.Ces deux études montrent que l'adsorption est la cause du « shear weakening », et implique un affaiblissement élastique global dans les roches granulaires peu cimentées (gréseuses et probablement carbonatées).L'effet de fréquence est étudié dans des grès de Fontainebleau et de Berea. Deux méthodes sont étudiées, toutes deux basées sur le principe de "stress-strain" (i.e. contrainte-déformation): l'oscillation "isotrope" (de la pression de confinement) et "déviatorique" (de la contrainte déviatorique).Ces deux modes d'oscillations sont tout d’abord calibrés à l’aide de plusieurs standards (e.g. aluminium, verre, gypse, plexiglass).Les échantillons de roche, saturés par des fluides de différentes viscosités, sont ensuite mesurés avec ces deux modes d'oscillation.Pour le premier mode d'oscillation, dit "isotrope", ce travail a permis de (i) mettre en évidence trois régimes élastiques distincts;et (ii) mesurer à la fois la conséquence (i.e. dispersion et atténuation du module d'incompressibilité) et la cause (i.e. écoulement fluide global) de la transition en fréquence entre état drainé et état non-drainé.Pour le second mode d'oscillation, dit "déviatorique", le module de Young et le coefficient de Poisson sont mesurés sur une gamme de fréquence apparente de [10-3;105] Hz.Pour un échantillon de grès de Fontainebleau, les deux transitions élastiques sont observées. Les mesures sont cohérentes avec les théories existantes.Un modèle 1D, prenant en compte les conditions de bord du système, est finalement développé. Ce modèle donne des résultats cohérents, et explique l'effet du volume mort sur les propriétés mesurées dans le cas d'une oscillation « isotrope ». / Seismics or Seismology are powerful tools to investigate Earth's crust. However, both rely on seismic waves that travelled through fluid-saturated sedimentary layers. This work, mainly experimental, aims at understanding the static and dynamic effects of the saturating fluid on the elastic response of clastic rocks.In this framework, two specific studies are emphasized:(i) the rock-fluid physico-chemical interaction, often addressed as the "shear weakening" effect, thought to affect the rock overall elastic response; and (ii) the rock-fluid mechanical interaction, addressed as "frequency effect", thought to induce a dependence of elastic properties to the measuring frequency.Two main rock types are investigated: Sandstone and Limestone. All rock samples are chosen to be isotropic and composed of a dominant mineral content, i.e. quartz for sandstone and calcite for limestone.

Errata inserted / Bibliography: leaves 170-175 / iii, 175 leaves : ill ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, 1987

No description available.

This thesis presents the design, construction, control and evaluation of a novel force controlled actuator. Traditional force controlled actuators are designed from the premise that "Stiffer is better''. This approach gives a high bandwidth system, prone to problems of contact instability, noise, and low power density. The actuator presented in this thesis is designed from the premise that "Stiffness isn't everything". The actuator, which incorporates a series elastic element, trades off achievable bandwidth for gains in stable, low noise force control, and protection against shock loads. This thesis reviews related work in robot force control, presents theoretical descriptions of the control and expected performance from a series elastic actuator, and describes the design of a test actuator constructed to gather performance data. Finally the performance of the system is evaluated by comparing the performance data to theoretical predictions.