Orientador: Flávio Vasconcelos da Silva / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Química / Made available in DSpace on 2018-08-23T22:24:17Z (GMT). No. of bitstreams: 1 Fonseca_RodolphoRodrigues_M.pdf: 9175458 bytes, checksum: f2dde4cc9ff6948eb7de39f51a322a4a (MD5) Previous issue date: 2013 / Resumo: Devido ao aumento da demanda nacional e mundial por combustíveis renováveis e novas tecnologias para melhoria de seus processos, é inegável a importância do desenvolvimento de novos controladores que possam garantir o funcionamento adequado destes sistemas. Neste contexto, este trabalho focou no desenvolvimento de um tipo de controlador não convencional baseado em inteligência artificial (Lógica Fuzzy) associado a uma estratégia Split-range para a manutenção da temperatura de reação de transesterificação do óleo de soja. Os ensaios foram conduzidos em um reator batelada totalmente instrumentado, monitorado e controlado via SCADA (Supervisory Control And Data Acquisition). Verificou-se que a melhor estratégia proposta para os sistemas de controle Fuzzy - Split-range na regulação da temperatura do reator foi a que empregou 147 regras sem a mistura de utilidades na jaqueta do reator, obtendo rápida estabilização da temperatura do reator, aproximadamente 15 minutos, e menor esforço de controle quando comparado às demais estratégias testadas. Como ferramentas de análise comparativa do sistema de controle foram utilizados os critérios de desempenho IAE, ISE e ITAE, além dos esforços de controle requeridos pelas válvulas durante os ensaios. Os resultados mostraram que a combinação Fuzzy - Split-range é viável no controle de temperatura, podendo ser estendida a demais processos industriais. / Abstract: In fact of national and international demand increasing for renewable fuels as biodiesel and also new technologies for process enhancement, it is worthy of attention the development of new controllers that guarantee adequate biodiesel production process control. In this context, this work applied the design of a non-conventional controller based on artificial intelligence (Fuzzy Logic) associated with Split-range strategy to regulate the temperature of soybean oil transesterification. The tests were conducted in a instrumented batch reactor, monitored and controlled by a SCADA (Supervisory Control And Data Acquisition) system. For the studied process control, the best combination set among the Fuzzy - Split-range strategies for the reactor's temperature control applied 147 set rules and no mixture of utilities in reactor's jacket. With fast temperature estabilization in almost 15 min, less control effort was required by the system among the strategies testeds. Performance criterions as IAE, ISE and ITAE were used to support comparative analysis, either control efforts by valves were used. The results show that Fuzzy - Split-range strategy is viable in biodiesel batch reactor temperature control, promising to application in others chemical processes. / Mestrado / Sistemas de Processos Quimicos e Informatica / Mestre em Engenharia Química

Orientador: Flavio Vasconcelos da Silva / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Química / Made available in DSpace on 2018-08-19T07:29:45Z (GMT). No. of bitstreams: 1 Leite_ManuelaSouza_D.pdf: 7752104 bytes, checksum: ab0250de44ea5b7e99ae3740fec4f538 (MD5) Previous issue date: 2011 / Resumo: Através de processos de polimerização pode-se obter uma enorme diversidade de produtos de grande interesse industrial, devido a sua vasta gama de aplicações. Cada reação deve ser feita com finalidades específicas, e então os polímeros devem ser produzidos sob certas condições operacionais estabelecidas, que lhe atribuem características relacionadas às suas aplicações. Buscando soluções para os diversos desafios existentes em processos de polimerização, tem sido crescente o número de trabalhos que buscam novas estratégias de controle mais eficazes para tais sistemas. O controle de um reator de polimerização apresenta grande dificuldade devido a sua natureza altamente não-linear e complexidade do mecanismo cinético da reação. Muitos dos algoritmos convencionais de controle não atendem, em sua totalidade, as exigências cada vez mais especificas destes processos e, visando atender estas necessidades, tem ocorrido nos últimos anos, um crescimento considerável na aplicação de estratégias de controle avançado em processos de polimerização. Este trabalho teve como proposta a montagem de um protótipo experimental, utilização de tecnologia de automação, desenvolvimento e implementação de estratégias de controle baseadas em inteligência artificial, especificamente, lógica fuzzy. A configuração do sistema experimental permitiu o acompanhamento em tempo real das principais variáveis do processo, possibilitando o uso de medidas on-line de variáveis, como viscosidade e densidade, as quais estão relacionadas indiretamente com o peso molecular e conversão, respectivamente. As estratégias de controle foram desenvolvidas com o objetivo de proporcionar a manutenção da temperatura da reação de polimerização em solução, e em batelada, do estireno, uma vez que esta variável possui influência direta na qualidade do produto final. O estireno foi selecionado como estudo de caso devido a sua importância industrial e ampla faixa de aplicação. Foram implementados no sistema experimental controladores mono e multivariáveis, e com saídas do tipo incremental e posicional, utilizando-se de modelos Mamdani e Sugeno. A reação foi conduzida por 3 horas, a um set-point de 90°C, utilizando uma concentração do monômero es tireno de 50% em volume, e como solvente o tolueno, tendo o BPO (peróxido de benzoíla) como iniciador. Foi definido um sistema de controle tipo cascata-fuzzy, o qual trata-se de um sistema que apresenta inediticidade, visto que não foram encontradas na literatura aplicações com esta configuração. Os controladores inteligentes foram aplicados na malha principal, tendo como variável de saída (variável intermediária) a temperatura da camisa do reator, e a variação de potência da resistência imersa no fluido térmico da camisa, como variável manipulada deste processo. As estratégias de controle, implementadas para um mesmo estudo de caso, foram analisadas conforme suas características, e todas as estruturas apresentaram eficiência do controle da reação de polimerização. A eficiência do sistema foi avaliada através do comportamento das variáveis controlada e manipulada, análise de índices de desempenho dos controladores (ISE, ITSE, IAE e ITAE), consumo de energia elétrica, visando redução de custos operacionais e, análise das propriedades finais do polímero obtido tais como: peso molecular médio, polidispersidade e produtividade (conversão) / Abstract: The favorable properties of polymeric products such as its usage, flexibility, light weight, low cost and its ease of processing, results in increased on their demand. Polymerization reactors have nonlinear natures and they show time varying behaviour. Their dynamic nature and the wide variations in operating conditions during batch cycles can make the reactor control difficult and important. Temperature variations greatly affect the kinetics of polymerization process and the produced polymer. As a result, to keep the product quality constant, the temperature of the reactor should be efficiently controlled. However, the control of polymerization reactors in general and particularly batch polymerization reactors is very difficult due to its complex characteristics. Advanced control techniques can be used as a viable solution for controlling and improving the efficiency and productivity of such nonlinear processes. Until recently, application of intelligent system such as fuzzy logic control in batch polymerization reactor control has been realized. This work included the installation of an experimental prototype, automation techniques, development and implementation of fuzzy control strategies in a batch polymerization reactor. The experimental system allowed realtime monitoring of key process variables. The design enables on-line measurement of variables indirectly related to the molecular weight and conversion, such as viscosity and density. A free radical polymerization of styrene was chosen as a process for the investigation, because polystyrene is an important product in today's industrial polymers and it has a very wide range of applications. In this process, temperature control is the most important control problem. Styrene (50%, v/v), toluene and benzoylperoxide (BPO) were used as the monomer, solvent and initiator, respectively. The reaction was conducted for 3 hours at a set point of 90 °C. Setting an unprecede nted fuzzy-cascade was employed. The intelligent controllers have been applied in the primary loop. The secondary variable was the temperature of the reactor jacket, and the power variation of resistance immersed in a heat transfer fluid that circulates through the jacket, as manipulated variable in this process. The experimental results show the effectiveness of fuzzy controller strategies. System efficiency was evaluated through the behavior of the controlled and manipulated variables, analysis of performance indices of the controllers (ISE, ITSE, IAE and ITAE), energy consumption, to reduce costs operational and analysis of the final properties of the polymer obtained such as average molecular weight, polydispersity and conversion / Doutorado / Sistema de Processos Quimicos e Informatica / Doutor em Engenharia Química

Orientadores: Artur Uldesmann, Antônio Gonçalves de Oliveira Filho / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Ciências Médicas / Made available in DSpace on 2018-08-19T05:11:13Z (GMT). No. of bitstreams: 1 Bassanezi_BetinaSilviaBeozzo_D.pdf: 847548 bytes, checksum: c9637586230bf2f7cf661864b073a7fd (MD5) Previous issue date: 2011 / Resumo: A incidência de vômitos pós-operatórios em pediatria ainda é alta, apesar dos avanços da anestesia. A ocorrência de vômitos pós-operatórios gera grande insatisfação nos pacientes e seus familiares, pois compromete a qualidade de vida e recuperação das crianças neste período. A profilaxia de vômitos pós-operatórios tem como primeiro passo identificar os pacientes de risco e o grau deste através de uma pontuação. Até o momento existe apenas uma escala de risco proposta para a população pediátrica a escala de Eberhart baseada numa análise estatística de regressão logística. A lógica fuzzy é uma teoria matemática que reconhece não somente duas possibilidades, verdadeiro ou falso como na lógica clássica, mas graus diferentes de falso e verdadeiro, o que permite uma melhor análise de varáveis contínuas como: idade e tempo de anestesia. Este estudo teve como objetivo desenvolver uma nova escala utilizando a teoria fuzzy para predizer a probabilidade de vômitos pósoperatórios em pacientes pediátricos oncológicos a partir da determinação dos fatores de risco. Foram coletados e analisados dados de 198 pacientes menores de 19 anos. Estes dados foram submetidos inicialmente ao teste qui-quadrado e regressão logística para determinar quais variáveis eram estatisticamente significativas para o risco de vômitos pós-operatórios. Destacaram-se quatro fatores de risco: idade, tempo de anestesia, uso de opióides para analgesia pósoperatória e tipo de cirurgia. A partir destes fatores foi criado um sistema para o cálculo de risco baseado na lógica fuzzy com uma interface computacional. O modelo fuzzy desenvolvido foi comparado com o proposto por Eberhart e mostrou-se mais efetivo para esta população. A interface computacional está disponível para acesso na internet (www.hc.unicamp.br/downloads/VPO/) e é de fácil utilização, sendo capaz de predizer a probabilidade de VPO em crianças com câncer com boa acurácia, possibilitando um melhor planejamento da profilaxia anti-emética / Abstract: Despite impressive advances in the field of anaesthesia, postoperative vomiting (POV) has a higher incidence in children. PONV may decrease children and parental satisfaction after surgeries, and impact on quality of living during recovery. The first rule for rational approach of PONV control is identify the patient at risk using predictive factors. There is only one specific score that predicts POV in children, the Eberhart's score. Fuzzy logic is a mathematical theory that has emerged as a type of logic that recognizes more than simple true and false values and takes into account levels of continuous variables such as age or duration of the surgery. In this study, we developed a fuzzy model to predict the probability of POV in pediatric oncologic patients who underwent surgery. Preoperative potential risk factors for POV in 198 children (0-19 yr old) with malignancies were collected and analyzed. Data analysis was performed with the chi-square test and logistic regression to evaluate probable risk factors for POV. This study found four risk factors: age, time of anesthesia, use postoperative opioids and type of surgery. A system based on fuzzy logic was developed with the risk factors found in the logistic regression, and a computational interface was created to calculate the probability of POV. After the analysis, the model was compared with Eberhart's score in the same population and showed a better performance. The use of the computational interface is available through the internet(www.hc.unicamp.br/downloads/VPO/), it is very easy to use and can predict the chance of POV in children with cancer with good accuracy, allowing better planning for postoperative prophylaxis of vomiting / Doutorado / Fisiopatologia Cirúrgica / Doutor em Ciências

Orientador: Luiz Felipe Mendes de Moura / Made available in DSpace on 2018-08-04T03:19:09Z (GMT). No. of bitstreams: 1 CarvajalFlorez_FabioAlejandro_D.pdf: 13507897 bytes, checksum: 763803f559424480cb0dc2207f08a244 (MD5) Previous issue date: 2004 / Resumo: No presente trabalho se realiza o estudo do controle da válvula de expansão e da velocidade de rotação do motor do compressor para um sistema de ar condicionado com armazenamento térmico em banco de gelo. Utilizando os princípios de projeto térmico, projeto eletrônico, programação de computadores e técnicas de controle procuram-se o aumento da eficiência do sistema. Neste controle propostos usou-se a lógica difusa. Inicialmente controlou-se o grau de superaquecimento através do controle da abertura da válvula de expansão e trabalhando o compressor em sua rotação máxima. Em seguida foi implementado um controle que trabalha regulando simultaneamente a velocidade do compressor e a abertura da válvula de expansão / Abstract: In this work is realized a control analysis ofthe expansion valve and the rotacional speed of the compressor engine for an air conditioning system with thermal storage in an ice storage tank. The thermal project, electronical project, computer programming and control technics are integrated to increase the performance of the system. To implement the proposed controls the logic fuzzy is used. Inicially, the refrigerant superheat was controlled through the expansion valve with the compressor working at highest rotational speed. Then, it was implemented the control regulating simultananeously the compressor speed and the expansion valve / Doutorado / Termica e Fluidos / Doutor em Engenharia Mecânica

[pt] Nessa dissertação é desenvolvido um modelo de previsão de curto prazo para cargas horárias empregando informações climáticas. Tal modelo é montado para a companhia de eletricidade LIGHT. O modelo proposto combina diferentes metodologias, são elas: Redes Neurais, Métodos Estatísticos e Lógica Nebulosa. Primeiramente, emprega-se o Mapa Auto-Organizável de Kohonen para identificar as curvas típicas de carga que são incluídas em um modelo de previsão estatística. Com intuito de melhorar o desempenho do modelo em termos do erro de previsão é adicionado, através de Lógica Nebulosa, o efeito da temperatura na carga. Por fim, é montado um procedimento com alguns conceitos de Lógica Nebulosa para identificar o tipo de curva de carga do dia a ser previsto. / [en] In the dissertation a new model to short-term forecasting of hourly loads using weather information is developed. This model was developed for the electricity distributing utility LIGHT and it combines different methodologies, namely: Neural Networks, Statistical Methods and Fuzzy Logic. First, the Kohonen Self-Organizing Map makes the identification of the load curves profiles and these are included in the statistical model. In order to improve the performance of the model in terms of forecasting error, the effect of temperature on the load is inserted by means of Fuzzy Logic. Finally, a procedure with some concepts of Fuzzy Logic was established to identify the type of curve of the day to be forecasted.

[pt] Este trabalho aborda o desenvolvimento de um ambiente gráfico, para simulação, de maneira simples, de um sistema de controle nebuloso. Este ambiente possibilita a introdução e alteração dos diversos parâmetros do controlador nebuloso e do sistema controlado, tais como equação da planta, conjuntos nebulosos, variáveis nebulosas, conjuntos de regras, etc. embora já existam atualmente ferramentas deste tipo, estas estão geralmente atreladas a controladores comerciais. O simulador desenvolvido visa oferecer uma ferramenta para uso também acadêmico, onde se possa ter uma noção melhor do funcionamento da técnica de controle nebuloso, aliada a uma facilidade de implementação e manipulação. São apresentados também alguns resultados de simulações feitas para comprovar o funcionamento do simulador. / [en] In this work a graphical environment for the simulation of a fuzzy control system is developed. In this environment, fuzzy controller and process parameters, such as plant equation, definitions of fuzzy sets, fuzzy variables, set of rules, etc., can be easily entered and changed. Althougth similar tools already exist, these are generaly linked to commercial controllers. The simulator developed in this work aims to provide a tool for academic use too, enabling the user to understand fuzzy control techniques. There are also shown results from experiments carried out in order to check and validate the simulator.

D.Phil. / Computer-based information systems will probably always have to contend with security issues. Much research have already gone into the field of information security. These research results have yielded some very sophisticated and effective security mechanisms and procedures. However, due to the ever increasing sophistication of criminals, combined with the ever changing and evolving information technology environment, some limitations still exist within the field of information security. Recent years have seen the proliferation of products embracing so-called artificial intelligence technologies. These products are in fields as diverse as engineering, business and medicine. The successes achieved in these fields pose the question whether artificial intelligence has a role to play within the field of information security. This thesis discusses limitations within information security and proposes ways in which artificial intelligence can be effectively applied to address these limitations. Specifically, the fields of authentication and risk analysis are identified as research fields where artificial intelligence has much to offer. These fields are explored in the context of their limitations and ways in which artificial intelligence can be applied to address these limitations. This thesis identifies two mainstream approaches in the attainment of artificial intelligence. These mainstream approaches are referred to as the "traditional" approach and the "non-traditional" approach. The traditional approach is based on symbolic processing, as opposed to the non-traditional approach, which is based on an abstraction of human reasoning. A representative technology from each of these mainstream approaches is selected to research their applicability within information security. Actual working prototypes of artificial intelligence techniques were developed to substantiate the results obtained in this research.

The quantification of information security risks is currently highly subjective. Values for information such as impact and probability, which are estimated during risk analysis, are mostly estimated by people or experts internal or external to the organization. Because the estimation of these values is done by people, all with different backgrounds and personalities, the values are exposed to subjectivity. The chance of any two people estimating the same value for risk analysis information is rare. There will always be a degree of uncertainty and imprecision in the values estimated. It is therefore during the data-gathering phase of risk analysis that the problem of subjectivity lies. To address the problem of subjectivity, techniques that mathematically deal with and present uncertainty and imprecision are used to estimate values for probability and impact. During this research a model for the objective estimation of probability was developed. The model uses mostly input values that are entirely objective, but also a small number of subjective input values. It is in these subjective input values that fuzzy logic and Monte Carlo simulation come into play. Fuzzy logic takes a qualitative subjective value and gives it an objective value, and Monte Carlo simulation complements fuzzy logic by giving a cumulative distribution function to the uncertain, imprecise input variable. In this way subjectivity is dealt with and the result of the model is a probability value that is estimated objectively. The same model that was used for the objective estimation of probability was used to estimate impact objectively. The end result of the research is the combination of the models to use the objective impact and probability values in a formula that calculates risk. The risk factors are then calculated objectively. A prototype was developed as proof that the process of objective information security risk quantification can be implemented in practice. / Prof. L. Labuschagne

This research aims at finding a foundation for assessment of capabilities and applying the concept in a human resource selection. The research identifies a common ground for assessing individuals’ applied capability in a given job based on literature review of various disciplines in engineering, human sciences and economics. A set of criteria is found to be common and appropriate to be used as the basis of this assessment. Applied Capability is then described in this research as the impact of the person in fulfilling job requirements and also their level of usage from their resources with regards to the identified criteria. In other words how their available resources (abilities, skills, value sets, personal attributes and previous performance records) can be used in completing a job. Translation of the person’s resources and task requirements using the proposed criteria is done through a novel algorithm and two prevalent statistical inference techniques (OLS regression and Fuzzy) are used to estimate quantitative levels of impact and utilisation. A survey on post graduate students is conducted to estimate their applied capabilities in a given job. Moreover, expert academics are surveyed on their views on key applied capability assessment criteria, and how different levels of match between job requirement and person’s resources in those criteria might affect the impact levels. The results from both surveys were mathematically modelled and the predictive ability of the conceptual and mathematical developments were compared and further contrasted with the observed data. The models were tested for robustness using experimental data and the results for both estimation methods in both surveys are close to one another with the regression models being closer to observations. It is believed that this research has provided sound conceptual and mathematical platforms which can satisfactorily predict individuals’ applied capability in a given job. This research has contributed to the current knowledge and practice by a) providing a comparison of capability definitions and uses in different disciplines, b) defining criteria for applied capability assessment, c) developing an algorithm to capture applied capabilities, d) quantification of an existing parallel model and finally e) estimating impact and utilisation indices using mathematical methods.

Abstract Handoff is the procedure providing the connection to the backbone network while a mobile terminal is moving across the boundaries of coverage of two wireless points of connection. The complexity of the handoff decision process has led to the examination of a number of traditional and pattern recognition handoff decision algorithms for wireless networks. Traditional algorithms use a received signal strength measurement and an optional threshold, hysteresis, or a dwell timer to determine the handoff decision. Degradation of the signal level, however, is a random process, and simple decision mechanisms result in a ping–pong effect whereby several consecutive handoffs degrade the service provided by the network. Consequently, more complex pattern recognition algorithms are needed to decide on the optimal time for handoff. In these algorithms, the handoff decision receives off line training to create a reference database of possible handoff locations in an environment with an associated handoff "fingerprint" at those locations. This dissertation introduces newly designed neural network and adaptive network based fuzzy inference system (ANFIS) pattern recognition algorithms. To select appropriate algorithms for a specific wireless network, we need to create an analytical framework for performance evaluation. The design of a framework for comparative performance evaluation of different handoff algorithms is a complex problem as different networks have different performance evaluation criteria. This dissertation divides wireless networks into three categories according to their topology and wireless service application: traditional cellular phone networks, heterogeneous wireless data networks, and rate adaptive wireless data networks. For each category of wireless networks we define a performance evaluation scenario and using Monte Carlo simulations, Monte Carlo calculations, and direct mathematical analysis we analyze the effects of different handoff decision algorithms. The Manhattan micro-cellular scenario is used for traditional cellular phone networks. Using Monte Carlo simulations on this scenario, the performance of traditional and our neural network and ANFIS handoff decision algorithms are compared. A moving-in moving-out performance evaluation scenario for heterogeneous wireless data networks is defined to characterize intertechnology roaming between two networks with substantially different data rates. We use Monte Carlo calculations to define the optimum handoff location for a mobile terminal in this scenario. Using Monte Carlo simulations and the optimal handoff location, we perform comparative performance evaluation of newly introduced asymmetric traditional and pattern recognition algorithms designed for intertechnology handoff. Finally, we introduce two performance evaluation scenarios for rate adaptive wireless networks to characterize user mobility in rate adaptive networks with random and grid deployments. For the first scenario we provide mathematical analysis for the effects of handoff using relative power to calculate the average throughput observed by the mobile terminal for different distances between the two wireless points of connection. For the second scenario designed for grid deployment we present a comparative performance analysis using Monte Carlo calculations for four handoff decision algorithms.