Orientador: Robson Pederiva / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-08-04T03:26:01Z (GMT). No. of bitstreams: 1 Bezerra_RobertodeAraujo_D.pdf: 3842898 bytes, checksum: 23a6bdd65c87be47fe4cfa856795bab0 (MD5) Previous issue date: 2004 / Resumo: Neste trabalho, é feito um estudo comparativo entre diversas técnicas de detecção de falhas em rolamentos por análise de vibração. Inicialmente, as técnicas foram aplicadas a modelos matemáticos de falhas nas pistas interna, externa e nas esferas dos rolamentos; sendo, em seguida, feito um estudo comparativo entre as técnicas. As técnicas foram aplicadas também a rolamentos com falhas induzidas nas pistas e esfera com diferentes tamanhos de falhas e submetidos a diferentes velocidades, para uma melhor compreensão das técnicas. Finalmente, as técnicas foram usadas para o monitoramento da evolução das falhas em um conjunto de doze rolamentos que foram submetidos a condições próximas as de trabalho, o que possibilitou um estudo mais detalhado do processo de evolução dessas falhas. Os resultados obtidos mostraram que, de todas as técnicas utilizadas, o envelope com filtro adaptativo foi a mais eficiente, sendo capaz de detectar pequenos amassamentos e o surgimento de falhas na gaiola. O estudo em condições próximas a realidade possibilitou uma melhor compreensão do processo de evolução das falhas em rolamentos / Abstract: In this thesis, it is made a comparative study among several vibration analysis techniques of fault detection. Initially, the techniques were applied to inner and outer race and rolling element mathematical models of faults, and so, it was possible to compare the techniques. It was used the same techniques on the races and rolling elements with different size of induced faults, submitting to different speeds. With those studies it was possible to understand how to use the technique in a better way. Finally, we monitored a group of twelve bearings to analyse the evolution of faults, close to real conditions. The results showed that the most efficient techniques is the envelope with adaptive filter, it detects small dentings and the begging of cage failure. It was possible to get a better understanding of the failure evolution process in bearings, studying it in close conditions to the reality / Doutorado / Mecanica dos Sólidos e Projeto Mecanico / Doutor em Engenharia Mecânica

This research examines the plausibility of developing demand forecasting techniques which are consistently and accurately able to predict demand. Time Series Techniques and Artificial Neural Networks are both investigated. Deodorant sales in South Africa are specifically studied in this thesis. Marketing techniques which are used to influence consumer buyer behaviour are considered, and these factors are integrated into the forecasting models wherever possible. The results of this research suggest that Artificial Neural Networks can be developed which consistently outperform industry forecasting targets as well as Time Series forecasts, suggesting that producers could reduce costs by adopting this more effective method.

Analytical problems caused by over-fitting, confounding and non-independence in the data is a major challenge for variable selection. As more variables are tested against a certain data set, there is a greater risk that some will explain the data merely by chance, but will fail to explain new data. The main aim of this study is to employ a systematic and practicable variable selection process for the spatial analysis and mapping of historical malaria risk in Botswana using data collected from the MARA (Mapping Malaria Risk in Africa) project and environmental and climatic datasets from various sources. Details of how a spatial database is compiled for a statistical analysis to proceed is provided. The automation of the entire process is also explored. The final bayesian spatial model derived from the non-spatial variable selection procedure using Markov Chain Monte Carlo simulation was fitted to the data. Winter temperature had the greatest effect of malaria prevalence in Botswana. Summer rainfall, maximum temperature of the warmest month, annual range of temperature, altitude and distance to closest water source were also significantly associated with malaria prevalence in the final spatial model after accounting for spatial correlation. Using this spatial model malaria prevalence at unobserved locations was predicted, producing a smooth risk map covering Botswana. The automation of both compiling the spatial database and the variable selection procedure proved challenging and could only be achieved in parts of the process. The non-spatial selection procedure proved practical and was able to identify stable explanatory variables and provide an objective means for selecting one variable over another, however ultimately it was not entirely successful due to the fact that a unique set of spatial variables could not be selected.

This study presents the methodology for estimating the incidence rate from the age specific prevalence data of three different eye diseases. We consider both situations where the mortality may differ from one person to another, with and without the disease. The method used was developed by Marvin J. Podgor for estimating incidence rate from prevalence data. It delves into the application of logistic regression to obtain the smoothed prevalence rates that helps in obtaining incidence rate. The study concluded that the use of logistic regression can produce a meaningful model, and the incidence rates of these diseases were not affected by the assumption of differential mortality.

Orientador: Luiz Koodi Hotta / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matemática Estatística e Computação Científica / Made available in DSpace on 2018-08-24T12:49:57Z (GMT). No. of bitstreams: 1 MercadoLondono_SergioLuis_M.pdf: 1311356 bytes, checksum: 23c15e842c02af3c1dc7de3a2a46a5df (MD5) Previous issue date: 2014 / Resumo: Uma das quantidades mais importante definida na epidemiologia é o número de reprodução basal, ou básico, associado com a pandemia e denotado por $R_0$. Ele proporciona uma medida da intensidade das intervenções necessárias para o controle da epidemia. Ao mesmo tempo, os modelos epidemiológicos compartimentais SIR, SEIR, tanto no enfoque enfoque determinístico quanto no estocástico, têm sido de grande ajuda para a compreensão dos mecanismos de transmissão de doenças infecciosas em todo o mundo. Esta dissertação apresenta alguns métodos para estimar esta quantidade através da utilização dos modelos epidemiológicos compartimentais. São considerados os quatro métodos apresentados por Chowell et al. (Mathematical Biosciences, 2007, v. 208, p. 571-589). O primeiro método é baseado na taxa de crescimento (inicial) exponencial da epidemia. Dada a taxa de crescimento exponencial e o modelo subjacente temos uma estimativa de $R_{0}$. No caso dos métodos 2 e 3 o processo de estimação do $R_0$ baseia-se nos modelos compartimentais, modelos SIR e SEIR no método 2, e em um modelo SEIR estendido no método 3. O método 4 utiliza uma abordagem bayesiana do modelo SIR estocástico. O objetivo da dissertação é estudar as propriedades dos estimadores baseados nos métodos 1, 2 e 4. Através de simulações são estimados os vícios, os erros quadráticos médios, as cobertura e as larguras dos intervalos de confiança. Os métodos são estudados quando os verdadeiros processos geradores de dados são os modelos SIR ou SEIR estocásticos. Inicialmente foram estudados os métodos, como apresentados por Chowell et al. (2007), e depois apresentadas modificações para melhorar o desempenho dos estimadores. A dissertação está organizada da seguinte forma: o Capítulo 2 consiste na apresentação dos modelos compartimentais, SIR e SEIR para análise das doenças infecciosas; tanto na abordagem determinística quanto estocástica. Este capítulo apresenta também o número de reprodução basal. O Capítulo 3 apresenta os quatro métodos de estimação apresentados em Chowell et al. (2007) para estimação do número de reprodução basal. O Capítulo 4 apresenta uma comparação de três dos quatro métodos através de simulação, quando o processo gerador de dados é um modelo SIR ou SEIR estocásticos. Neste capítulo também são apresentadas as modificações dos métodos. A conclusão final e as sugestões de trabalhos futuros são apresentadas no Capítulo 5 / Abstract: The basic reproduction number, usually denoted by $R_0$, is one of the most important quantities defined in epidemiology and is associated with the potential of an infectious disease to spread through a population. It provides a measure of the intensity needed to control the epidemic interventions. At the same time, the compartmental epidemiological models SIR and SEIR , both in the deterministic and in the stochastic approach, have been very helpful for understanding the mechanisms of infectious diseases transmission. This paper considers the four methods presented by Chowell et al. (Mathematical Biosciences, 2007, v. 208, p. 571-589) to estimate $R_0$. All methods are based on compartmental epidemiological models. The first method is based on the epidemic (initial) exponential growth rate. Given an estimate of the exponential growth rate and an underlying compartmental model we have an estimate of $R_{0}$. The second method is based on fitting SIR or SEIR compartmental models, and the third method in fitting an extended SEIR model. The fourth method uses a Bayesian approach to a stochastic SIR model. The aim of this work is to study the properties of estimators based on methods 1, 2 and 4. The bias, the mean squared errors, the coverage and the widths of the confidence intervals are estimated through simulation. The methods are studied when the true data generating processes are the stochastic SIR or SEIR models. Initially the methods, as presented by Chowell et al. (2007), were studied and then presented modifications to improve the performance of the estimators. The dissertation is organized as follows: Chapter 2 consists of the presentation of compartmental SIR and SEIR models, the deterministic and stochastic approaches for analysis of infectious diseases. This chapter also presents the basic reproduction number. Chapter 3 explains the four estimation methods presented in Chowell et al. (2007) to estimate the basic reproduction number. Chapter 4 discusses and compares three of the four methods by simulation when the data generating process is a SIR or SEIR model. In this chapter the modifications of the methods are also considered. The final conclusion and suggestions for future work are presented in Chapter 5 / Mestrado / Estatistica / Mestre em Estatística

Orientador: Carlos Eduardo Angeli Furlani / Coorientador: Rouverson Pereira da Silva / Banca: Carlos Alessandro Chioderoli / Banca: Paulo Roberto Arbex Silva / Resumo: O Brasil é o maior produtor mundial de cana-de-açúcar, devido a expansão de áreas cultivadas, a adubação mecanizada torna-se muito importante para o aumento do desempenho operacional nesta cultura, porém esta adubação pode demonstrar alguns problemas na distribuição de fertilizantes, como a não realização da caracterização dos fertilizantes por meio do ângulo de repouso, granulometria e densidade dos nutrientes, entre outros. Dessa forma, torna-se necessário o desenvolvimento de novas tecnologias que possibilitem uma melhora na aplicação de fertilizantes na cultura da cana-de-açúcar. À vista disso, um novo conceito de adubadora está em desenvolvimento considerada como protótipo, pois esta realiza aplicação individualizada de nitrogênio, fósforo e potássio proporcionando maior eficiência na distribuição destes em relação às outras adubadoras. Com este trabalho o objetivo foi avaliar aplicação mecanizada de N-P-K individualizada na cultura da cana-de-açúcar. O experimento foi desenvolvido no município de Matão em área de cana-de-açúcar pertencente à Fazenda Cascavel, possuindo 1,66 ha aproximadamente de área experimental. O delineamento experimental foi Inteiramente Casualizado (DIC), com três tratamentos e trinta repetições por tratamento. Este delineamento foi estabelecido conforme os critérios do controle de qualidade, sendo o monitoramento das variáveis realizado durante a operação de adubação. Ao final do período de avaliação foram coletados 90 pontos amostrais no total, sendo 30 pontos por tratamento. Os tratamentos foram: 1- Adubação mecanizada, sem aplicação de herbicida; 2- Operação conjugada (aplicação simultânea de herbicida e adubação); e 3- Duas operações (aplicação separada de herbicida e adubo). Concluiu-se que a melhor qualidade operacional por meio das cartas de controle foi o tratamento 3, sendo duas operações (aplicação separada de herbicida e adubo)... / Abstract: The Brazil is the largest producer of sugarcane, due to expansion of cultivated areas, mechanized fertilizer becomes very important for the increase in operating performance in this culture, but this fertilization may show some problems in the distribution of fertilizers, so that when not performing the characterization of fertilizers through the angle of repose, particle size and density of nutrients, among other. Thus, it becomes necessary to develop new technologies that enable an improvement in the application of fertilizers in the cultivation of sugarcane. In view of this, a new concept of fertilizer in development considered as a prototype because this performs individualized application of nitrogen, phosphorus and potassium, providing greater efficiency in the distribution of these over other fertilizer machine. With this work the objective was to evaluate the mechanical application of NPK individualized culture of sugarcane. The experiment was conducted in the Matão municipality in the area of sugarcane belonging to the Fazenda Cascavel, possessing approximately 1.66 ha experimental area. The experimental design was completely randomized (DIC), with three treatments and thirty replications for treatment. This design was established according to the criteria of quality control, and monitoring of variables held during fertilization operation. At the end of the evaluation period were collected 90 sample points in total, with 30 points per treatment. The treatments were: 1- mechanized fertilized, without herbicide; 2- Operation combined (simultaneous application of herbicide and fertilizer); and 3- Two operations (separate application of herbicide and fertilizer). It was concluded that the best operational quality through control charts was the third treatment, two operations (separate application of herbicide and fertilizer) for presenting less variability. The right side of fertilizer machine was the best for having applied ... / Mestre

The emergence of new Internet applications and technologies have resulted in an increased complexity as well as a need for lower latency, higher bandwidth, and increased reliability. This ultimately results in an increased complexity of network operation and management. Manual management is not sufficient to meet these new requirements. There is a need for data driven techniques to advance from manual management to autonomous management of network systems. One such technique, Machine Learning (ML), can use data to create models from hidden patterns in the data and make autonomous modifications. This approach has shown significant improvements in other domains (e.g., image recognition and natural language processing). The use of ML, along with advances in programmable control of Software- Defined Networks (SDNs), will alleviate manual network intervention and ultimately aid in autonomous network operations. However, realizing a data driven system that can not only understand what is happening in the network but also operate autonomously requires advances in the networking domain, as well as in ML algorithms. In this thesis, we focus on developing ML-based network architectures and data driven net- working algorithms whose objective is to improve the performance and management of future networks and network applications. We focus on problems spanning across the network protocol stack from the application layer to the physical layer. We design algorithms and architectures that are motivated by measurements and observations in real world or experimental testbeds. In Part I we focus on the challenge of monitoring and estimating user video quality of experience (QoE) of encrypted video traffic for network operators. We develop a system for REal-time QUality of experience metric detection for Encrypted Traffic, Requet. Requet uses a detection algorithm to identify video and audio chunks from the IP headers of encrypted traffic. Features extracted from the chunk statistics are used as input to a random forest ML model to predict QoE metrics. We evaluate Requet on a YouTube dataset we collected, consisting of diverse video assets delivered over various WiFi and LTE network conditions. We then extend Requet, and present a study on YouTube TV live streaming traffic behavior over WiFi and cellular networks covering a 9-month period. We observed pipelined chunk requests, a reduced buffer capacity, and a more stable chunk duration across various video resolutions compared to prior studies of on-demand streaming services. We develop a YouTube TV analysis tool using chunks statistics detected from the extracted data as input to a ML model to infer user QoE metrics. In Part II we consider allocating end-to-end resources in cellular networks. Future cellular networks will utilize SDN and Network Function Virtualization (NFV) to offer increased flexibility for network infrastructure operators to utilize network resources. Combining these technologies with real-time network load prediction will enable efficient use of network resources. Specifically, we leverage a type of recurrent neural network, Long Short-Term Memory (LSTM) neural networks, for (i) service specific traffic load prediction for network slicing, and (ii) Baseband Unit (BBU) pool traffic load prediction in a 5G cloud Radio Access Network (RAN). We show that leveraging a system with better accuracy to predict service requirements results in a reduction of operation costs. We focus on addressing the optical physical layer in Part III. Greater network flexibility through SDN and the growth of high bandwidth services are motivating faster service provisioning and capacity management in the optical layer. These functionalities require increased capacity along with rapid reconfiguration of network resources. Recent advances in optical hardware can enable a dramatic reduction in wavelength provisioning times in optical circuit switched networks. To support such operations, it is imperative to reconfigure the network without causing a drop in service quality to existing users. Therefore, we present a ML system that uses feedforward neural networks to predict the dynamic response of an optically circuit-switched 90-channel multi-hop Reconfigurable Optical Add-Drop Multiplexer (ROADM) network. We show that the trained deep neural network can recommend wavelength assignments for wavelength switching with minimal power excursions. We extend the performance of the ML system by implementing and testing a Hybrid Machine Learning (HML) model, which combines an analytical model with a neural network machine learning model to achieve higher prediction accuracy. In Part IV, we use a data-driven approach to address the challenge of wireless content delivery in crowded areas. We present the Adaptive Multicast Services (AMuSe) system, whose objective is to enable scalable and adaptive WiFi multicast. Specifically, we develop an algorithm for dynamic selection of a subset of the multicast receivers as feedback nodes. Further, we describe the Multicast Dynamic Rate Adaptation (MuDRA) algorithm that utilizes AMuSe’s feedback to optimally tune the physical layer multicast rate. Our experimental evaluation of MuDRA on the ORBIT testbed shows that MuDRA outperforms other schemes and supports high throughput multicast flows to hundreds of nodes while meeting quality requirements. We leverage the lessons learned from AMuSe for WiFi and use order statistics to address the performance issues with LTE evolved Multimedia Broadcast/Multicast Service (eMBMS). We present the Dynamic Monitoring (DyMo) system which provides low-overhead and real-time feedback about eMBMS performance to be used for network optimization. We focus on the Quality of Service (QoS) Evaluation module and develop a Two-step estimation algorithm which can efficiently identify the SNR Threshold as a one time estimation. DyMo significantly outperforms alternative schemes based on the Order-Statistics estimation method which relies on random or periodic sampling.

This diploma thesis deals with the development of individual financial indicators of a selected company using the methods of financial analysis, interval regression analysis and time series analysis. Based on the results of these analyzes there is evaluated the financial situation of the company and created a forecast of the evolution of the selected indicators for the next two years. The selected company is described by an analysis of the internal and external surroundings. Possible risks are identified by a risk analysis and proposals are made to reduce the value of individual risks to an acceptable level.

Subject of the project is a concept of utilization of the basic quality devices during monitoring of critical parameters of silicon monocrystals’ production.

The Master´s thesis is concerned with an analysis of financial statements of a selected stock company using statistical methods and further comparisons with two major branch competitors. On the basis of the previous steps is settled comprehensive assessment of the current situation and future strategic steps are proposed. These steps are supposed to bring an economical growth to the company. The work is built up on the necessary theoretical economic and strategic backgrounds.