No description available.

The design of the freight train truck has gone relatively unchanged over the past 150 years. There has been relatively little change to the fundamental railway truck design because of the challenges of implementing a cost effective and reliable modification to designs that have proven effective in decades of operation. A common U. S. railway truck consists of two sideframes, a bolster, two spring nests, and four friction wedges. The two sideframes sit on the axels. The bolster rides on springs on top of the sideframes. The friction wedges also ride on springs on top of the sideframe, and are positioned between the bolster and sideframe, acting as a damping mechanism. Better understanding the dynamic behavior and forces on the bodies are critical in reducing unnecessary wear on the components, along with potential negative behavior such as loss of productivity and increase in operating costs. This thesis will investigate the dynamic behavior of the truck under warping conditions using a stand-alone model created in Virtual.Lab. This research covers two main areas. First, the full-truck model will be developed and its simulation results will be compared to test data from the Transportation Technology Center, Inc. (TTCI). Data was provided from warp testing performed at the TTCI facilities in the spring of 2008. Once validated, the model will be used to gain a better understanding of the forces and moments that are propagated through the system, and of the dynamics of all bodies. Due to costs and physical constraints, not every bogie component can be instrumented during test, so the computer model will be able to provide valuable information not easily obtained otherwise. Second, full-truck models using different contact geometry between the wedges, sideframes, and bolster will be compared. A model with extremely worn sideframes will allow for investigation into the effects of wear on the damping abilities and warp stiffness of the truck. Another model using split wedges will be compared with the previous model to investigate into the behavior differences in the truck using different types of wedges. By understanding the impact of different geometries on the overall performance of the truck, better decisions on design and maintenance can be made in the future. After creating the models, we found that the full-truck model created in LMS® Virtual.Lab compared well with the test data collected by TTCI. In the comparison with NUCARS® we determined that the stand-alone model, which incorporates the wedges as bodies, captures the warp dynamics of the truck better than NUCARS®, which models the wedges as connections. By creating a model with severely worn sideframes, we were able to determine that the truck loses its abilities to damp bounce in the system as well as to prevent warping when the components become sufficiently worn. The split-wedge model behaved similarly to the standard full-truck model for bounce inputs, but had a significantly different behavior in warp. Further development will be needed on the split-wedge model to be confident that it behaved as expected. / Master of Science

The majority of automatic speech recognition (ASR) systems rely on hidden Markov models (HMM), in which the output distribution associated with each state is modelled by a mixture of diagonal covariance Gaussians. Dynamic information is typically included by appending time-derivatives to feature vectors. This approach, whilst successful, makes the false assumption of framewise independence of the augmented feature vectors and ignores the spatial correlations in the parametrised speech signal. This dissertation seeks to address these shortcomings by exploring acoustic modelling for ASR with an application of a form of state-space model, the linear dynamic model (LDM). Rather than modelling individual frames of data, LDMs characterize entire segments of speech. An auto-regressive state evolution through a continuous space gives a Markovian model of the underlying dynamics, and spatial correlations between feature dimensions are absorbed into the structure of the observation process. LDMs have been applied to speech recognition before, however a smoothed Gauss-Markov form was used which ignored the potential for subspace modelling. The continuous dynamical state means that information is passed along the length of each segment. Furthermore, if the state is allowed to be continuous across segment boundaries, long range dependencies are built into the system and the assumption of independence of successive segments is loosened. The state provides an explicit model of temporal correlation which sets this approach apart from frame-based and some segment-based models where the ordering of the data is unimportant. The benefits of such a model are examined both within and between segments. LDMs are well suited to modelling smoothly varying, continuous, yet noisy trajectories such as found in measured articulatory data. Using speaker-dependent data from the MOCHA corpus, the performance of systems which model acoustic, articulatory, and combined acoustic-articulatory features are compared. As well as measured articulatory parameters, experiments use the output of neural networks trained to perform an articulatory inversion mapping. The speaker-independent TIMIT corpus provides the basis for larger scale acoustic-only experiments. Classification tasks provide an ideal means to compare modelling choices without the confounding influence of recognition search errors, and are used to explore issues such as choice of state dimension, front-end acoustic parametrization and parameter initialization. Recognition for segment models is typically more computationally expensive than for frame-based models. Unlike frame-level models, it is not always possible to share likelihood calculations for observation sequences which occur within hypothesized segments that have different start and end times. Furthermore, the Viterbi criterion is not necessarily applicable at the frame level. This work introduces a novel approach to decoding for segment models in the form of a stack decoder with A* search. Such a scheme allows flexibility in the choice of acoustic and language models since the Viterbi criterion is not integral to the search, and hypothesis generation is independent of the particular language model. Furthermore, the time-asynchronous ordering of the search means that only likely paths are extended, and so a minimum number of models are evaluated. The decoder is used to give full recognition results for feature-sets derived from the MOCHA and TIMIT corpora. Conventional train/test divisions and choice of language model are used so that results can be directly compared to those in other studies. The decoder is also used to implement Viterbi training, in which model parameters are alternately updated and then used to re-align the training data.

The increased popularity of electric vehicles and e-mobility among the people, have encouraged many automotive companies and research organisations to develop good strategies for drivetrain designs involving batteries. As seen in the department of Alternative Powertrains research is carried out on hybrid fuel cell and electric vehicles. This thesis deals with the development of lithium ion battery model for electric vehicle simulations. A novel approach using black box modelling is developed for development of battery model using only the available battery measurements. Furthermore, a measurement test strategy is formulated providing the process direction and measurement parameters to be considered. Developed battery model provide voltage estimates for given Charge rate,temperature and State of Charge (SOC). The comparison of experimentally obtained and model estimated values. The model developed has a very good accuracy in estimation.

As the technology behind renewable energy sources becomes more advanced and cost-effective, these sources have become an ever-increasing portion of the generation portfolios of power systems across the country. While the shift away from non-renewable resources is generally considered beneficial, the fact remains that intermittent renewable sources present special challenges associated with their unique operating characteristics. Because of the high variability of intermittent renewables, the frequency performance of the system to which they are connected can degrade. Generators assigned to regulate frequency, keeping it close to the desired 60 Hz, are forced to ramp up and down quickly in order to offset the rise and fall of the variable resources (in addition to the rise and fall of load), causing transient frequency deviations, power swings, major interface transfer variations and other significant issues. This research measures the impact of intermittent renewable resource penetration level on power system frequency performance, and offers methods for managing that performance. Currently, the generally accepted amount of regulation (rapidly-dispatchable reserve, used as a supplement to base generation on a short time scale to avoid performance issues) is 1% of peak load. Because of the high variability associated with intermittent renewables, including wind generation (the focus of this thesis), it is expected that this amount of regulation must increase in order to maintain adequate system frequency performance. Thus, the primary objective of this thesis is to quantify the amount of regulation necessary to maintain adequate frequency performance as a function of the penetration level of wind generation. Presently, balancing resource requirements are computed, in both industry and in the research literature, using static models, which rely entirely on statistical manipulation of net load, failing to capture the intricacies of dynamic system and generator interactions. Using a dynamic model with high temporal resolution data, instead of these statistical models, this thesis confirms the need for additional regulation as wind generation penetration increases. But beyond that, our research demonstrates an exponentially increasing relationship between necessary regulation and wind generation percentage, indicating that, without further technological breakthroughs, there is a practical limit to the amount of wind generation that a typical system can accommodate. Furthermore, we compare our dynamic model results with those of the statistical models, and show that the majority of current statistical models substantially under-predict the necessary amount of regulation to accommodate significant amounts of wind generation. Finally, we verify that the ramping capability of the regulating generators impacts the amount of necessary regulation, although it is generally ignored in current analysis and related literature.

Sob o regime de metas de inflação cabe à autoridade monetária balisar seus instrumentos de política de forma a manter a estabilidade do nível geral de preços. Neste aspecto, pelo caráter volátil dos índices de inflação cheia os bancos centrais de todo o mundo utilizam o conceito de núcleo da inflação para tentar capturar com maior acurácia a tendência subjacente da taxa de inflação. Muitas vezes os índices de preços ao consumidor estão altamente sujeitos a volatilidades decorrentes de fatores temporários e muitas vezes localizados. E já que o objetivo da autoridade monetária está em zelar pela estabilidade \"real\" (ou de fato) do nível geral de preços, mudanças temporárias ou localizadas não afetam as taxas de inflação no longo prazo e, consequentemente, não cabe à autoridade monetária responder a tais mudanças, pois isso poderia gerar uma volatilidade desnecessária à política monetária com consequência sobre as flutuações da atividade econômica no período. Dessa forma, Bancos Centrais do mundo inteiro fazem uso de núcleos de inflação. Este trabalho aplica uma nova metodologia de cálculo de núcleo para a inflação brasileira, utilizando o modelo de fatores dinâmicos generalizados. Esta abordagem permite diferenciar fatores localizados (idiossincráticos) dos choques comuns (generalizados) em um grande conjunto de dados. Usamos o IPCA em seu nível mais desagregado e geramos o choque comum entre este conjunto. E a este choque chamamos de núcleo da inflação. Sua eficiência em termos de antecedência à inflação cheia no curto prazo foi testada por meio de uma cointegração, VEC, tais resultados foram comparados com o desempenho do núcleo por Exclusão, mostrando uma maior eficiência do núcleo aqui encontrado. / Under the inflation target system lies to the monetary authority the evaluation of the best tools to keep general price stability. In this context, due to the volatile character of the inflation, central banks around the world use the concept of the inflation core in attempt to capture in a more accurately way the prices trends. Several times, consumer prices indexes are subjected to very volatile prices, due to temporary or localized factors. As the vigilance of the monetary authority relies on the real stability of the general prices level, temporary or localized changes doesn\'t affect the inflation indexes in the long run and, therefore, it\'s not an issue to the central bank to respond to this variations, this could indeed create an unnecessary volatility to the monetary politics with consequences to the economic activity in the period. This way, central banks around the world calculate and use inflation core. This paper applies a new methodology to calculate the inflation core to the Brazilian inflation, using the generalized dynamic factor model. With this approach it\'s possible to differentiate the localized factors from the common (generalized) shocks in a great data set. We use IPCA on its more disaggregated level and create a common shock in the data set, and we name this shock the inflation core. We test the advance of this core to the inflation in the short run using a VEC, and compare with the results of the Exclusion core, we show that your core by using dynamic factor model is more effcient then Exclusion core.

Este trabalho apresenta um modelo de dinâmica de desenvolvimento distribuído de software, cujo objetivo é representar a realidade e os aspectos de ambientes de DDS (Desenvolvimento distribuído de software), a fim de torná-los observáveis e descritíveis qualitativa e quantitativamente. Um modelo preliminar foi elaborado a partir da revisão bibliográfica e de um caso de experimentação desenvolvido por LErario et al (2004). Para a construção e validação deste modelo, a metodologia de estudo de múltiplos casos foi aplicada em diversas organizações que desenvolvem software de maneira distribuída. Ao modelo preliminar foram adicionados estados e transições significantes para a dinâmica do desenvolvimento distribuído de software, originando então o M3DS (Modelo de Dinâmica de Desenvolvimento Distribuído de Software). Duas versões do M3DS são apresentadas. Uma versão construída sobre uma máquina de estados, cujo objetivo é representar apenas a transições entre os estados. Outra versão equivalente, porém mais formal, é apresentada no formato de redes de Petri, na qual é possível visualizar a dependência entre transições e mudanças de estado. Com este modelo, é possível compreender o funcionamento de um projeto distribuído e auxiliar na eficácia da gestão da rede de produção, além de auxiliar as demais entidades e pessoas envolvidas a obterem um posicionamento na rede mais preciso. O M3DS pode, também, auxiliar a detecção proativa de problemas originados a partir do desenvolvimento a distância. Os resultados apresentados neste trabalho respondem a questão de como as organizações desenvolvedoras de software produzem software de maneira distribuída. A riginalidade da pesquisa centra-se na construção de um modelo de dinâmica do desenvolvimento distribuído elaborado com os dados levantados a partir de seis estudos de casos. / This work presents a dynamic model of distributed development of software, whose objective is to represent the reality and the aspects of DDS environments, in order to turn them qualitatively and quantitatively observable. A preliminary model was elaborated from the bibliographical revision and an experimentation case developed by L\'Erario et al (2004). The construction and validation of this model used the methodology multiple cases study in several organizations that develop software in a distributed way. After this, states and transitions were added in the dynamics model of the distributed development of software creating the M3DS. (Dynamics Model of Distributed Development of Software). Two versions of M3DS are presented. A version built on a state machine whose objective is demonstrating the transitions among the states. Another version equivalent, however more formal, it is presented in the format of Petri nets. The second version makes possible to visualize the dependence between transitions and state changes. With this model it is possible to understand the operation of a distributed project, aiding in the effectiveness of the manager of the network production and people can obtain a precise positioning in network. Besides, M3DS can also aid the proactive detection of problems originated from the development at the distance. The results presented in this work answer the question: how the development software organizations produce software in a distributed way. The originality of the research is the construction of a model of dynamics of the distributed development elaborated from data of six cases studies.

In the face of the internationally tightened requirements and regulations for CO2 emissions from the transportation sector, waste heat recovery using a thermoelectric generator (TEG) has become the most significant research interest. A vehicular TEG, converting otherwise wasted thermal energy from engines to electricity directly for use in the vehicle systems, is a promising approach for vehicle original equipment manufacturers (OEMs) to reduce fuel consumption and lower CO2 emissions. This thesis aims to explore the main challenges to be faced in the commercialization of TEGs. Based on a review of the literature, four research gaps have been identified, which are respectively: * Translating the material improvements into TEG Performance, * Transient behaviors of vehicular TEGs under driving cycles, * Fuel saving percentage and cost-benefit estimation of TEG, * Bidirectional characteristic of TEM and bifunctional vehicular TEG. To directly address these research gaps, a quasi-static TEM model, a dynamic TEG model, a semi-empirical vehicular TEG model, and a dual-model TEM model have been respectively developed and validated through experiments on both TEM test rigs and TEG engine test benches. These developed models are used as tools to investigate the performance of TEG, parameters sensitivity, and integration effects. Model-based TEG control, TEG cost benefit ratio and feasibility of a bifunctional TEG are also explored based on the developed models. The simulation results show that TEG power generation is highly sensitive to the heat transfer coefficient of hot side heat exchanger and thermal contact resistance. The TEG installation position is identified as the most important integration effect. It has been found by the simulation result that the fuel saving with TEG installed upstream of the three-way catalyst (TWC) is 50% higher than the fuel saving with TEG installed downstream of the TWC. The fuel saving percentage for a skutterudite vehicular TEG, which can generate around 400-600W in constant speed 120km/h, is 0.5-3.6% depending on the integration position in the exhaust line. A 3-minute faster warm-up effect of engine oil can be obtained when the bifunctional TEG works in engine warm-up mode with electrical current applied.

Research in soft robot hardware has led to the development of platforms that allow for safer performance when working in uncertain or dynamic environments. The potential of these platforms is limited by the lack of proper dynamic models to describe or controllers to operate them. A common difficulty associated with these soft robots is a representation for torque, the common electromechanical relation seen in motors does not apply. In this thesis, several different torque models are presented and used to construct linear state-space models. The control limitations on soft robots are induced by natural compliance inherent to the hardware. This inherent compliance results in soft robots that are commonly underdamped and present significant oscillations when accelerated quickly. These oscillations can be mitigated through model-based controllers which can anticipate these oscillations. In this thesis, multiple model predictive controllers are implemented with the torque models produced and results are presented for an inflatable single-DoF pneumatically actuated soft robot. Larger, multi-DoF, soft robots present additional issues with control, where flexibility in one joint impacts control in others. In this thesis a preliminary method and results for controlling multiple joints on an inflatable multi-DoF pneumatically actuated soft robot are presented. While model predictive controllers are capable, their control commands are defined by solving an optimization constrained by model dynamics. This optimization relies on minimizing the cost of a user-defined objective function. This objective function contains a series of weights, which allow the user to tune the importance of each component in the objective function. As there are no calculations that can be performed to tune model predictive controllers to achieve superior control performance, they often need to be tuned tediously by a skilled operator. In this thesis, a method for automated discrete performance identification and model predictive controller weight tuning is presented. This thesis constructs multiple state-space models for single- and multi-DoF underdamped, antagonistic, pneumatically actuated soft robots and shows that these models can be used with model predictive control, tuned for performance, to achieve accurate joint position control.

The research aims at modeling and analyzing the interactions among functional process variables (FPVs) for process control in semiconductor manufacturing. Interaction is a universal phenomenon and different interaction patterns among system components might characterize the system conditions. To monitor and control the system, process variables are normally collected for observation which could vary with time and present in a functional form. These FPVs interact with each other and contain rich information regarding the process conditions. As an example in one of the semiconductor manufacturing processes, changes of interactions among FPVs like temperature and coefficient of friction (COF) might characterize different process conditions. This dissertation systematically developed a methodology to study interaction among FPVs through statistical and physical modeling. Three main topics are discussed in this dissertation: (1) Interaction patterns of FPVs under varying process conditions are studied both through experiments and statistical approaches. A method based on functional canonical correlation analysis (FCCA) is employed to extract the interaction patterns between FPVs and experiments of wafer polishing processes are conducted to verify the patterns of FPVs under varying process conditions. (2) Interaction among FPVs is further studied based on physics for process condition diagnosis. A mathematical model based on nonlinear dynamics is developed to study the strength of interaction and their directionalities, and advanced statistical control charts followed by this nonlinear dynamics model are established for process monitoring. (3) Complex interaction structures among multiple FPVs are analyzed based on nonlinear dynamics for a better understanding of process mechanism. An approach with extended nonlinear dynamics model is proposed to characterize process conditions, and combined engineering knowledge, complex interaction structure patterns are concluded accordingly for interpretation of process mechanism. The main contribution of this dissertation is to propose a novel methodology based on nonlinear dynamics, which could investigate interactions between components of systems and provide physical understanding of process mechanism for process monitoring and diagnosis. Through studies on interaction among FPVs in semiconductor manufacturing, this research provides guidance for improvement of manufacturing processes. Not limited to manufacturing, the developed methodology can be applied to other areas such as healthcare delivery.