Contrasting interannual and multidecadal NAO variability

Decadal and longer timescale variability in the winter North Atlantic Oscillation (NAO) has considerable impact on regional climate, yet it remains unclear what fraction of this variability is potentially predictable. This study takes a new approach to this question by demonstrating clear physical differences between NAO variability on interannual-decadal (<30 year) and multidecadal (>30 year) timescales. It is shown that on the shorter timescale the NAO is dominated by variations in the latitude of the North Atlantic jet and storm track, whereas on the longer timescale it represents changes in their strength instead. NAO variability on the two timescales is associated with different dynamical behaviour in terms of eddy-mean flow interaction, Rossby wave breaking and blocking. The two timescales also exhibit different regional impacts on temperature and precipitation and different relationships to sea surface temperatures. These results are derived from linear regression analysis of the Twentieth Century and NCEP-NCAR reanalyses and of a high-resolution HiGEM General Circulation Model control simulation, with additional analysis of a long sea level pressure reconstruction. Evidence is presented for an influence of the ocean circulation on the longer timescale variability of the NAO, which is particularly clear in the model data. As well as providing new evidence of potential predictability, these findings are shown to have implications for the reconstruction and interpretation of long climate records.

Time-periodic perturbation of a Lienard equation with an unbounded homoclinic loop

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Resonance and chaos .1. First-order interior resonances

Analytical models for studying the dynamical behaviour of objects near interior, mean motion resonances are reviewed in the context of the planar, circular, restricted three-body problem. The predicted widths of the resonances are compared with the results of numerical integrations using Poincare surfaces of section with a mass ratio of 10(-3) (similar to the Jupiter-Sun case). It is shown that for very low eccentricities the phase space between the 2:1 and 3:2 resonances is predominantly regular, contrary to simple theoretical predictions based on overlapping resonance. A numerical study of the 'evolution' of the stable equilibrium point of the 3:2 resonance as a function of the Jacobi constant shows how apocentric libration at the 2:1 resonance arises; there is evidence of a similar mechanism being responsible for the centre of the 4:3 resonance evolving towards 3:2 apocentric libration. This effect is due to perturbations from other resonances and demonstrates that resonances cannot be considered in isolation. on theoretical grounds the maximum libration width of first-order resonances should increase as the orbit of the perturbing secondary is approached. However, in reality the width decreases due to the chaotic effect of nearby resonances.

Resonance and chaos: I. First-order interior resonances

Analytical models for studying the dynamical behaviour of objects near interior, mean motion resonances are reviewed in the context of the planar, circular, restricted threebody problem. The predicted widths of the resonances are compared with the results of numerical integrations using Poincaré surfaces of section with a mass ratio of 10-3 (similar to the Jupiter-Sun case). It is shown that for very low eccentricities the phase space between the 2:1 and 3:2 resonances is predominantly regular, contrary to simple theoretical predictions based on overlapping resonance. A numerical study of the 'evolution' of the stable equilibrium point of the 3:2 resonance as a function of the Jacobi constant shows how apocentric libration at the 2:1 resonance arises; there is evidence of a similar mechanism being responsible for the centre of the 4:3 resonance evolving towards 3:2 apocentric libration. This effect is due to perturbations from other resonances and demonstrates that resonances cannot be considered in isolation. On theoretical grounds the maximum libration width of first-order resonances should increase as the orbit of the perturbing secondary is approached. However, in reality the width decreases due to the chaotic effect of nearby resonances.

Análise de estruturas flexíveis com aplicação de materiais viscoelásticos

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Influência das não linearidades geométricas em sistemas de suspensão com amortecimento assimétrico sob excitação harmônica

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Sistemi liquido cristallini complessi: simulazioni al calcolatore e studi ESR

The aim of this PhD thesis was to study at a microscopic level different liquid crystal (LC) systems, in order to determine their physical properties, resorting to two distinct methodologies, one involving computer simulations, and the other spectroscopic techniques, in particular electron spin resonance (ESR) spectroscopy. By means of the computer simulation approach we tried to demonstrate this tool effectiveness for calculating anisotropic static properties of a LC material, as well as for predicting its behaviour and features. This required the development and adoption of suitable molecular models based on a convenient intermolecular potentials reflecting the essential molecular features of the investigated system. In particular, concerning the simulation approach, we have set up models for discotic liquid crystal dimers and we have studied, by means of Monte Carlo simulations, their phase behaviour and self­-assembling properties, with respect to the simple monomer case. Each discotic dimer is described by two oblate Gay­Berne ellipsoids connected by a flexible spacer, modelled by a harmonic "spring" of three different lengths. In particular we investigated the effects of dimerization on the transition temperatures, as well as on the characteristics of molecular aggregation displayed and the relative orientational order. Moving to the experimental results, among the many experimental techniques that are typically employed to evaluate LC system distinctive features, ESR has proved to be a powerful tool in microscopic scale investigation of the properties, structure, order and dynamics of these materials. We have taken advantage of the high sensitivity of the ESR spin probe technique to investigate increasingly complex LC systems ranging from devices constituted by a polymer matrix in which LC molecules are confined in shape of nano- droplets, as well as biaxial liquid crystalline elastomers, and dimers whose monomeric units or lateral groups are constituted by rod-like mesogens (11BCB). Reflection-mode holographic-polymer dispersed liquid crystals (H-PDLCs) are devices in which LCs are confined into nanosized (50­-300 nm) droplets, arranged in layers which alternate with polymer layers, forming a diffraction grating. We have determined the configuration of the LC local director and we have derived a model of the nanodroplet organization inside the layers. Resorting also to additional information on the nanodroplet size and shape distribution provided by SEM images of the H-PDLC cross-section, the observed director configuration has been modeled as a bidimensional distribution of elongated nanodroplets whose long axis is, on the average, parallel to the layers and whose internal director configuration is a uniaxial quasi- monodomain aligned along the nanodroplet long axis. The results suggest that the molecular organization is dictated mainly by the confinement, explaining, at least in part, the need for switching voltages significantly higher and the observed faster turn-off times in H-PDLCs compared to standard PDLC devices. Liquid crystal elastomers consist in cross-linked polymers, in which mesogens represent the monomers constituting the main chain or the laterally attached side groups. They bring together three important aspects: orientational order in amorphous soft materials, responsive molecular shape and quenched topological constraints. In biaxial nematic liquid crystalline elastomers (BLCEs), two orthogonal directions, rather than the one of normal uniaxial nematic, can be controlled, greatly enhancing their potential value for applications as novel actuators. Two versions of a side-chain BLCEs were characterized: side­-on and end­-on. Many tests have been carried out on both types of LCE, the main features detected being the lack of a significant dynamical behaviour, together with a strong permanent alignment along the principal director, and the confirmation of the transition temperatures already determined by DSC measurements. The end­-on sample demonstrates a less hindered rotation of the side group mesogenic units and a greater freedom of alignment to the magnetic field, as already shown by previous NMR studies. Biaxial nematic ESR static spectra were also obtained on the basis of Molecular Dynamics generated biaxial configurations, to be compared to the experimentally determined ones, as a mean to establish a possible relation between biaxiality and the spectral features. This provides a concrete example of the advantages of combining the computer simulation and spectroscopic approaches. Finally, the dimer α,ω-bis(4'-cyanobiphenyl-4-yl)undecane (11BCB), synthesized in the "quest" for the biaxial nematic phase has been analysed. Its importance lies in the dimer significance as building blocks in the development of new materials to be employed in innovative technological applications, such as faster switching displays, resorting to the easier aligning ability of the secondary director in biaxial phases. A preliminary series of tests were performed revealing the population of mesogenic molecules as divided into two groups: one of elongated straightened conformers sharing a common director, and one of bent molecules, which display no order, being equally distributed in the three dimensions. Employing this model, the calculated values show a consistent trend, confirming at the same time the transition temperatures indicated by the DSC measurements, together with rotational diffusion tensor values that follow closely those of the constituting monomer 5CB.

Polymakromonomere als anisotropes Modellsystem zur Untersuchung kolloidaler Mehrkomponentensysteme

Das Ziel dieser Arbeit bestand in der Untersuchung des Phasenverhaltens kolloidaler Mehrkomponentensysteme. Konzentrierte und kompatible Mischungen aus linearen Polymerketten und Polymakromonomeren als kolloidanaloge Partikel wurden auf ihre Eignung hin ein kolloidales Glas zu bilden untersucht.Die Synthese der Polymakromonomere erfolgte durch radikalische Polymerisation von Vinylbenzyl-Makromer, was zu zylinderartigen Makromolekülen führt, die auch als molekulare „Flaschenbürsten“ angesehen werden können. Aufgrund der grossen Längen-Polydispersität der Hauptkette wurden die erhaltenen Polymakromonomere mittels Continous Polymer Fractionation (CPF) fraktioniert, was zu Modell-Partikeln von verschiedener Molekülgestalt führte: von kugelförmig über stäbchenförmig bis hin zu wurmartig.Weiterhin konnte gezeigt werden, dass die „haarige“ Oberfläche der Polymakromonomere zumindestens teilweise die entropisch bedingte Entmischung von Partikeln unterschiedlicher Molekülgestalt, wie Polymerknäuel und stäbchenartige, überwinden kann.Um das dynamische Verhalten zu untersuchen, wurden rheologische Messungen an den Schmelzen verschiedener Kolloid-Polymer-Mischungen ausgeführt. Es zeigte sich, dass Polymakromonomere mit grosser Hauptkettenlänge bzw. wurmartiger Gestalt einen signifikanten Verstärkungseffekt auf die Kolloid-Polymer-Mischung haben.

Contrasting interannual and multidecadal NAO variability

Decadal and longer timescale variability in the winter North Atlantic Oscillation (NAO) has considerable impact on regional climate, yet it remains unclear what fraction of this variability is potentially predictable. This study takes a new approach to this question by demonstrating clear physical differences between NAO variability on interannual-decadal (<30 year) and multidecadal (>30 year) timescales. It is shown that on the shorter timescale the NAO is dominated by variations in the latitude of the North Atlantic jet and storm track, whereas on the longer timescale it represents changes in their strength instead. NAO variability on the two timescales is associated with different dynamical behaviour in terms of eddy-mean flow interaction, Rossby wave breaking and blocking. The two timescales also exhibit different regional impacts on temperature and precipitation and different relationships to sea surface temperatures. These results are derived from linear regression analysis of the Twentieth Century and NCEP-NCAR reanalyses and of a high-resolution HiGEM General Circulation Model control simulation, with additional analysis of a long sea level pressure reconstruction. Evidence is presented for an influence of the ocean circulation on the longer timescale variability of the NAO, which is particularly clear in the model data. As well as providing new evidence of potential predictability, these findings are shown to have implications for the reconstruction and interpretation of long climate records.

Comparación de un nuevo método de aceleración de análisis dinámicos con métodos de reducción de modelos

La dinámica estructural estudia la respuesta de una estructura ante cargas o fenómenos variables en el tiempo. En muchos casos, estos fenómenos requieren realizar análisis paramétricos de la estructura considerando una gran cantidad de configuraciones de diseño o modificaciones de la estructura. Estos cambios, ya sean en fases iniciales de diseño o en fases posteriores de rediseño, alteran las propiedades físicas de la estructura y por tanto del modelo empleado para su análisis, cuyo comportamiento dinámico se modifica en consecuencia. Un caso de estudio de este tipo de modificaciones es la supervisión de la integridad estructural, que trata de identificar la presencia de daño estructural y prever el comportamiento de la estructura tras ese daño, como puede ser la variación del comportamiento dinámico de la estructura debida a una delaminación, la aparición o crecimiento de grieta, la debida a la pérdida de pala sufrida por el motor de un avión en vuelo, o la respuesta dinámica de construcciones civiles como puentes o edificios frente a cargas sísmicas. Si a la complejidad de los análisis dinámicos requeridos en el caso de grandes estructuras se añade la variación de determinados parámetros en busca de una respuesta dinámica determinada o para simular la presencia de daños, resulta necesario la búsqueda de medios de simplificación o aceleración del conjunto de análisis que de otra forma parecen inabordables tanto desde el punto de vista del tiempo de computación, como de la capacidad requerida de almacenamiento y manejo de grandes volúmenes de archivos de datos. En la presente tesis doctoral se han revisado los métodos de reducción de elementos .nitos más habituales para análisis dinámicos de grandes estructuras. Se han comparado los resultados de casos de estudio de los métodos más aptos, para el tipo de estructuras y modificaciones descritas, con los resultados de aplicación de un método de reducción reciente. Entre los primeros están el método de condensación estática de Guyan extendido al caso con amortiguamiento no proporcional y posteriores implementaciones de condensaciones dinámicas en diferentes espacios vectoriales. El método de reducción recientemente presentado se denomina en esta tesis DACMAM (Dynamic Analysis in Complex Modal space Acceleration Method), y consiste en el análisis simplificado que proporciona una solución para la respuesta dinámica de una estructura, calculada en el espacio modal complejo y que admite modificaciones estructurales. El método DACMAM permite seleccionar un número reducido de grados de libertad significativos para la dinámica del fenómeno que se quiere estudiar como son los puntos de aplicación de la carga, localizaciones de los cambios estructurales o puntos donde se quiera conocer la respuesta, de forma que al implementar las modificaciones estructurales, se ejecutan los análisis necesarios sólo de dichos grados de libertad sin pérdida de precisión. El método permite considerar alteraciones de masa, rigidez, amortiguamiento y la adición de nuevos grados de libertad. Teniendo en cuenta la dimensión del conjunto de ecuaciones a resolver, la parametrización de los análisis no sólo resulta posible, sino que es también manejable y controlable gracias a la sencilla implementación del procedimiento para los códigos habituales de cálculo mediante elementos .nitos. En el presente trabajo se muestra la bondad y eficiencia del método en comparación con algunos de los métodos de reducción de grandes modelos estructurales, verificando las diferencias entre sí de los resultados obtenidos y respecto a la respuesta real de la estructura, y comprobando los medios empleados en ellos tanto en tiempo de ejecución como en tamaño de ficheros electrónicos. La influencia de los diversos factores que se tienen en cuenta permite identificar los límites y capacidades de aplicación del método y su exhaustiva comparación con los otros procedimientos. ABSTRACT Structural dynamics studies the response of a structure under loads or phenomena which vary over time. In many cases, these phenomena require the use of parametric analyses taking into consideration several design configurations or modifications of the structure. This is a typical need in an engineering o¢ ce, no matter the structural design is in early or final stages. These changes modify the physical properties of the structure, and therefore, the finite element model to analyse it. A case study, that exempli.es this circumstance, is the structural health monitoring to predict the variation of the dynamical behaviour after damage, such as a delaminated structure, a crack onset or growth, an aircraft that suffers a blade loss event or civil structures (buildings or bridges) under seismic loads. Not only large structures require complex analyses to appropriately acquire an accurate solution, but also the variation of certain parameters. There is a need to simplify the analytical process, in order to bring CPU time, data .les, management of solutions to a reasonable size. In the current doctoral thesis, the most common finite element reduction methods for large structures are reviewed. Results of case studies are compared between a recently proposed method, herein named DACMAM (Dynamic Analysis in Complex Modal space Acceleration Method), and different condensation methods, namely static or Guyan condensation and dynamic condensation in different vectorial spaces. All these methods are suitable for considering non-classical damping. The reduction method DACMAM consist of a structural modification in the complex modal domain which provides a dynamic response solution for the reduced models. This process allows the selection of a few degrees of freedom that are relevant for the dynamic response of the system. These d.o.f. are the load application points, relevant structural points or points in which it is important to know the response. Consequently, an analysis with structural modifications implies only the calculation of the dynamic response of the selected degrees of freedom added, but with no loss of information. Therefore, mass, stiffness or damping modifications are easily considered as well as new degrees of freedom. Taking into account the size of the equations to be solved, the parameterization of the dynamic solutions is not only possible, but also manageable and controllable due to the easy implementation of the procedure in the standard finite element solvers. In this thesis, the proposed reduction method for large structural models is compared with other published model order reduction methods. The comparison shows and underlines the efficiency of the new method, and veri.es the differences in the response when compared with the response of the full model. The CPU time, the data files and the scope of the parameterization are also addressed.

Prediction of the growth of wear-type rail corrugation

The growth behaviour of the vibrational wear phenomenon known as rail corrugation is investigated analytically and numerically using mathematical models. A simplified feedback model for wear-type rail corrugation that includes a wheel pass time delay is developed with an aim to analytically distil the most critical interaction occurring between the wheel/rail structural dynamics, rolling contact mechanics and rail wear. To this end, a stability analysis on the complete system is performed to determine the growth of wear-type rail corrugations over multiple wheelset passages. This analysis indicates that although the dynamical behaviour of the system is stable for each wheel passage, over multiple wheelset passages, the growth of wear-type corrugations is shown to be the result of instability due to feedback interaction between the three primary components of the model. The corrugations are shown analytically to grow for all realistic railway parameters. From this analysis an analytical expression for the exponential growth rate of corrugations in terms of known parameters is developed. This convenient expression is used to perform a sensitivity analysis to identify critical parameters that most affect corrugation growth. The analytical predictions are shown to compare well with results from a benchmarked time-domain finite element model. (C) 2004 Elsevier B.V. All rights reserved.

Accelerating precursory activity within a class of earthquake analogue automata

A statistical fractal automaton model is described which displays two modes of dynamical behaviour. The first mode, termed recurrent criticality, is characterised by quasi-periodic, characteristic events that are preceded by accelerating precursory activity. The second mode is more reminiscent of SOC automata in which large events are not preceded by an acceleration in activity. Extending upon previous studies of statistical fractal automata, a redistribution law is introduced which incorporates two model parameters: a dissipation factor and a stress transfer ratio. Results from a parameter space investigation indicate that a straight line through parameter space marks a transition from recurrent criticality to unpredictable dynamics. Recurrent criticality only occurs for models within one corner of the parameter space. The location of the transition displays a simple dependence upon the fractal correlation dimension of the cell strength distribution. Analysis of stress field evolution indicates that recurrent criticality occurs in models with significant long-range stress correlations. A constant rate of activity is associated with a decorrelated stress field.

Description and training of neural network dynamics

Attractor properties of a popular discrete-time neural network model are illustrated through numerical simulations. The most complex dynamics is found to occur within particular ranges of parameters controlling the symmetry and magnitude of the weight matrix. A small network model is observed to produce fixed points, limit cycles, mode-locking, the Ruelle-Takens route to chaos, and the period-doubling route to chaos. Training algorithms for tuning this dynamical behaviour are discussed. Training can be an easy or difficult task, depending whether the problem requires the use of temporal information distributed over long time intervals. Such problems require training algorithms which can handle hidden nodes. The most prominent of these algorithms, back propagation through time, solves the temporal credit assignment problem in a way which can work only if the relevant information is distributed locally in time. The Moving Targets algorithm works for the more general case, but is computationally intensive, and prone to local minima.

A structural method for the design of fault tolerant distributed control systems

Distributed digital control systems provide alternatives to conventional, centralised digital control systems. Typically, a modern distributed control system will comprise a multi-processor or network of processors, a communications network, an associated set of sensors and actuators, and the systems and applications software. This thesis addresses the problem of how to design robust decentralised control systems, such as those used to control event-driven, real-time processes in time-critical environments. Emphasis is placed on studying the dynamical behaviour of a system and identifying ways of partitioning the system so that it may be controlled in a distributed manner. A structural partitioning technique is adopted which makes use of natural physical sub-processes in the system, which are then mapped into the software processes to control the system. However, communications are required between the processes because of the disjoint nature of the distributed (i.e. partitioned) state of the physical system. The structural partitioning technique, and recent developments in the theory of potential controllability and observability of a system, are the basis for the design of controllers. In particular, the method is used to derive a decentralised estimate of the state vector for a continuous-time system. The work is also extended to derive a distributed estimate for a discrete-time system. Emphasis is also given to the role of communications in the distributed control of processes and to the partitioning technique necessary to design distributed and decentralised systems with resilient structures. A method is presented for the systematic identification of necessary communications for distributed control. It is also shwon that the structural partitions can be used directly in the design of software fault tolerant concurrent controllers. In particular, the structural partition can be used to identify the boundary of the conversation which can be used to protect a specific part of the system. In addition, for certain classes of system, the partitions can be used to identify processes which may be dynamically reconfigured in the event of a fault. These methods should be of use in the design of robust distributed systems.

Non-Linear oil firm force coefficients for a journal bearing operating under aligned and misaligned conditions

It is well established that hydrodynamic journal bearings are responsible for self-excited vibrations and have the effect of lowering the critical speeds of rotor systems. The forces within the oil film wedge, generated by the vibrating journal, may be represented by displacement and velocity coefficient~ thus allowing the dynamical behaviour of the rotor to be analysed both for stability purposes and for anticipating the response to unbalance. However, information describing these coefficients is sparse, misleading, and very often not applicable to industrial type bearings. Results of a combined analytical and experimental investigation into the hydrodynamic oil film coefficients operating in the laminar region are therefore presented, the analysis being applied to a 120 degree partial journal bearing having a 5.0 in diameter journal and a LID ratio of 1.0. The theoretical analysis shows that for this type of popular bearing, the eight linearized coefficients do not accurately describe the behaviour of the vibrating journal based on the theory of small perturbations, due to them being masked by the presence of nonlinearity. A method is developed using the second order terms of Taylor expansion whereby design charts are provided which predict the twentyeight force coefficients for both aligned, and for varying amounts of journal misalignment. The resulting non-linear equations of motion are solved using a modified Newton-Raphson method whereby the whirl trajectories are obtained, thus providing a physical appreciation of the bearing characteristics under dynamically loaded conditions.