Este trabalho apresenta algumas contribuições para métodos de controle baseados em dados obtidos em apenas um experimento, a fim de torná-los mais atrativos quanto à aplicação em processos industriais. A partir de dados obtidos em experimentos no processo, os métodos baseados em dados estimam os parâmetros de um controlador de estrutura fixa por meio da minimização do erro entre a saída do sistema real em malha fechada e uma saída desejada, dada por um modelo de referência. O método Virtual Reference Feedback Tuning - VRFT é o método mais expressivo na literatura que estima os parâmetros do controlador usando apenas uma batelada de dados, mas o mesmo apresenta alguns inconvenientes em sua formulação que limitam a sua aplicação. Neste trabalho, o método VRFT é modificado de forma que se obtém um método VRFT flexível, o qual minimiza um critério flexível, pelo qual são estimados tanto os parâmetros do controlador quanto os parâmetros do numerador do modelo de referência; assim, caso a planta que se deseja controlar seja de fase não-mínima, o critério é capaz de estimar esses zeros e os mesmos devem ser incluídos no modelo de referência que será utilizado no projeto do controlador. Além disso, em sistemas com ruído, o método VRFT necessita de uma variável instrumental para que a estimativa dos parâmetros do controlador seja não-polarizada. Para eliminar a necessidade de usar variáveis instrumentais, um novo método de controle baseado em dados é proposto, o qual é descrito sob a ótica de identificação. Este método pode ser visto como a identificação de um sistema, no qual a função de transferência do processo é reparametrizada em função do controlador ideal e do modelo de referência. Além disso, estende-se a teoria de projeto de experimento com solução baseada em restrições LMI para o caso da identificação do controlador ótimo. Todas essas contribuições são ilustradas através de simulações. / This work presents some contributions to data-based control methods where data is obtained in only one experiment, in order to make them more attractive to industrial process applications. Using data from experiments on the process, data-based methods estimate the parameters of fixed structure controller, through the minimization of the error between the closed loop response of the system and the desired response, given by a reference model. The Virtual Reference Feedback Tuning - VRFT method is the most expressive method in the literature that estimates the controller parameters using only one batch of data, but this method presents some inconveniences in its formulation which limit its application. In this work, the VRFT method is modified in order to obtain a flexible VRFT method, which minimizes a flexible criterion, and then obtains the controller parameters together with the parameters related to the reference model numerator; thus, if the plant which we want to control is non-minimum phase, then the criterion is able to estimate these zeros and they need to be included in the reference model that will be used in the control design. Besides, when dealing with noisy systems, the VRFT method needs an instrumental variable so the controller parameters estimate is unbiased. In order to eliminate this necessity, a new data-based control is proposed in this work, which is formulated using identification theory. This method can be seen as the identification of a system, where the process transfer function is reparameterized as a function of the ideal controller and the reference model. Besides, we extend the experimental design theory where the problem is solved using LMI constraints to the case of the optimal controller identification method. All these contributions are illustrated through simulations.

The objective of this thesis is to develop algorithms for application-oriented input design. This procedure takes the model application into account when designing experiments for system identification. This thesis is divided into two parts. The first part considers the theory of application-oriented input design, with special attention to Model Predictive Control (MPC). We start by studying how to find a convex approximation of the set of models that result in acceptable control performance using analytical methods when controllers with no closed-form control law, for e.g., MPC are employed. The application-oriented input design is formulated in time domain to enable handling of signals constraints. The framework is extended to closed-loop systems where two cases are considered i.e., when the plant is controlled by a general but known controller and for the case of MPC. To this end, an external stationary signal is designed via graph theory. Different sources of uncertainty in application-oriented input design are investigated and a robust application-oriented input design framework is proposed. The second part of this thesis is devoted to the problem of estimating the number of occupants based on the information available to HVAC systems in buildings. The occupancy estimation is first formulated as a two-tier problem. In the first tier, the room dynamic is identified using temporary measurements of occupancy. In the second tier, the identified model is employed to formulate the problem as a fused-lasso problem. The proposed method is further developed to be used as a multi-room estimator using a physics-based model. However, since it is not always possible to collect measurements of occupancy, we proceed by proposing a blind identification algorithm which estimates the room dynamic and occupancy, simultaneously. Finally, the application-oriented input design framework is employed to collect data that is informative enough for occupancy estimation purposes. / <p>QC 20170620</p>

Identifying the parameters with the largest influence on the predicted outputs of a model revealswhich parameters need to be known more precisely to reduce the overall uncertainty on themodel output. A large improvement of such models would result when uncertainties in the keymodel parameters are reduced. To achieve this, new experiments could be very helpful,especially if the measurements are taken at the spatio-temporal locations that allow estimate the parameters in an optimal way. After evaluating the methodologies available for optimal sensor location, a few observations were drawn. The method based on the Gram determinant evolution can report results not according to what should be expected. This method is strongly dependent of the sensitivity coefficients behaviour. The approach based on the maximum angle between subspaces, in some cases, produced more that one optimal solution. It was observed that this method depends on the magnitude of outputs values and report the measurement positions where the outputs reached their extrema values. The D-optimal design method produces number and locations of the optimal measurements and it depends strongly of the sensitivity coefficients, but mostly of their behaviours. In general it was observed that the measurements should be taken at the locations where the extrema values (sensitivity coefficients, POD modes and/or outputs values) are reached. Further improvements can be obtained when a reduced model of the system is employed. This is computationally less expensive and the best estimation of the parameter is obtained, even with experimental data contaminated with noise. A new approach to calculate the time coefficients belonging to an empirical approximator based on the POD-modes derived from experimental data is introduced. Additionally, an artificial neural network can be used to calculate the derivatives but only for systems without complex nonlinear behaviour. The latter two approximations are very valuable and useful especially if the model of the system is unknown.

The aim of Master's thesis is to apply statistical methods in production process. The assignment of the thesis is to describe and evaluate pressing process which is the part of posistor production. The first theoretical part contains introduction to the mathematical statistics, verifying assumptions about the data, describing the regression analysis, the analysis of variance and the part of the design of experiment. The second practical part is focused on design, performing and interpreting of experiments. The thesis is developed according to requirements of company EPCOS s.r.o. Šumperk. Statistical program MINITAB Rlease 14, which is supported by company, was used for all computations.

This thesis is focused on experimental work in the domain of vacuum evaporation efficiency improvement. The first part introduces theory of liquids evaporation and its influencing factors, then multi-stage flash distillation technology. The next part of theory is an overview of other technologies used for desalination of sea water and introduction to the design of experiments. In the practical part are selected factors affecting efficiency of device and experiment plan is designed for them. When the experiment is done, optimal operating conditions of technology are designed on basis of analysis.

No description available.

Multi-fold Time Domain Electromagnetics (TDEM) is a novel experimental approach that couples elements of traditional land-based TDEM survey designs to obtain a robust data set. This design inherently accommodates a broad range of possible Earth models through a rich combination of analysis opportunities making it ideally suited for reconnaissance. Kentland Farms, VA was chosen as the test site, for its ease of access and interesting geologic features such as river terraces and karstic landscape. Three independent methods of analyzing the 3-component data set each provided unique insights into the subsurface electrical structure through a complementary interpretation. Synthesis of log-normalized ∂tB_z pseudo-sections provided a first-order analysis of the lateral and vertical heterogeneities of the profile. A Zero-Crossing Moveout (ZCMO) analysis used a brute-force grid-search inversion to estimate the two-layer Earth model that best-fit the observed moveout times for a range of interface depths. By using the ZCMO result as an initial model, regularized 1D Occam inversions determined a 3-layer electrical structure consisting of a 3.5 m – 5 m thick resistive upper layer, over an 12.5 m – 15 m thick conductive layer, overlying a resistive half-space. From correlation of the inverse solutions with ZCMO derived conductivity models and prior resistivity information, the depth to the limestone bedrock was approximated to be 16 – 20 m. The delineation of the bedrock depth provided additional support for the fill-cut terrace formation model (Ward et al., 2005), as well as possible evidence of groundwater drainage on the 40 m terrace at Kentland Farm. / Master of Science

Geological Carbon Storage (GCS) is considered as one of the key techniques to reduce the rate of atmospheric emissions of CO2 and thereby to contribute to controlling the global warming. A successful application of a GCS project requires the capability of the formation to trap CO2 for a long term. In this context, processes related to CO2 trapping and also possible leakage of CO2 to the near surface environment need to be understood. The overall aim of this thesis is to understand the flow and transport of CO2 through porous media in the context of geological storage of CO2. The entire range of scales, including the pore scale, the laboratory scale, the field experiment scale and the industrial scale of CO2 injection operation are addressed, and some of the key processes investigated by means of experiments and modeling.  First, a numerical model and laboratory experimental setup were developed to investigate the CO2 gas flow, mimicking the system in the near-surface conditions in case a leak from the storage formation should occur. The system specifically addressed the coupled flow and mass transport of gaseous CO2 both in the porous domain as well as the free flow domain above it. The comparison of experiments and modelling results showed a very good agreement indicating that the model developed can be applied to evaluate monitoring and surface detection of potential CO2 leakage. Second, the field scale CO2 injection test carried out in a shallow aquifer in Maguelone, France was analyzed and modeled. The results showed that Monte Carlo simulations accounting for the heterogeneity effects of the permeability field did capture the key observations of the monitoring data, while a homogeneous model could not represent them. Third, a numerical model based on phase-field method was developed and model simulations carried out addressing the effect of wettability on CO2-brine displacement at the pore-scale. The results show that strongly water-wet reservoirs provide a better potential for the dissolution trapping, due to the increase of interface between CO2 and brine with very low contact angles. The results further showed that strong water-wet conditions also imply a strong capillary effect, which is important for residual trapping of CO2. Finally, numerical model development and model simulations were carried out to address the large scale geological storage of CO2 in the presence of impurity gases in the CO2 rich phase. The results showed that impurity gases N2 and CH4 affected the spatial distribution of the gas (the supercritical CO2 rich phase), and a larger volume of reservoir is needed in comparison to the pure CO2 injection scenario. In addition, the solubility trapping significantly increased in the presence of N2 and CH4.

Der verhaltensökonomischen Literatur entsprechend behandeln die drei Kapitel dieser Dissertation unterschiedliche Aspekte des menschlichen Verhaltens, welches als "nicht-rational" zu bezeichnen ist. Jedes dieser Kapitel leistet einen Beitrag zum aktuellen Stand der Forschung auf dem Gebiet der Verhaltensökonomik mit Hilfe von entweder experimentellen, empirischen oder methodischen Ansätzen. Das erste Kapitel schlägt ein einfaches verhaltensökonomisches Modell vor und unterzieht dieses einer Reihe von experimentellen Tests. Das Modell erweitert die Literatur zur Fehlwahrnehmung von multiplikativen Wachstumsprozessen und hilft somit typische Fehlinvestitionen in der langen Frist zu erklären. Im Rahmen des zweiten Kapitels werden Daten einer Online-Kreditbörse genutzt, um empirisch zu untersuchen, ob sich private Investoren entsprechend den Vorhersagen der standardmäßigen ökonomischen Fachliteratur verhalten und einzig die erwartete Rendite berücksichtigen oder ob sie von anderen nicht-finanztechnischen Attributen eines Schuldners beeinflusst werden. Der Schwerpunkt der Analyse liegt dabei auf Geschlechterdiskriminierung im Rahmen dessen unterschiedliche Diskriminierungskonzepte getestet werden. Das dritte Kapitel wählt einen methodischen Ansatz und schlägt ein innovatives Experiment-Design vor, welches den empirisch gut dokumentierten Schwierigkeiten bzgl. der Angabe von subjektiven Wahrscheinlichkeiten von Teilnehmern an Umfragen und Laborexperimenten Rechnung trägt. Ein Binary-Choice-Ansatz eingebettet in ein adaptives Experiment-Design minimiert den Aufwand für die Befragten und ermöglich somit eine praktikable und effiziente Elizitierung der subjektiven Meinungen. / In the line with the literature on behavioral economics, the three chapters of this dissertation shed light on different aspects of human behavior that are at odds with rationality. Each chapter contributes to the existing behavioral economic research using either experimental, empirical, or methodological tools. First, by proposing and experimentally testing a simple behavioral model that extends the literature on the misperception of multiplicative growth processes, Chapter 1 aims to explain common money mistakes that people often make with long-term investments such as retirement savings plans. Second, in Chapter 2, real-life investment data of an online-lending platform are used to empirically investigate if private investors behave as the standard economic literature would predict and solely consider an investment’s expected return or if they also care about other non-financial attributes of a debtor. The focus of the analysis is on gender discrimination, thereby defining and econometrically testing different concepts of how investors discriminate between male and female borrowers. Third, Chapter 3 takes a methodological path and proposes a novel experimental design that accounts for the empirically well-documented difficulties that survey respondents typically have when asked to state subjective probabilities. A binary choice approach embedded in an adaptive experimental design helps to minimize effort of the respondents, thus allowing for a more practical belief elicitation in both the lab and the field.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-01-11T11:13:50Z No. of bitstreams: 1 pedrolinojonecoutinho.pdf: 1490441 bytes, checksum: 2d30a406bff1dc2da6123c10c2e44617 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-01-25T17:04:18Z (GMT) No. of bitstreams: 1 pedrolinojonecoutinho.pdf: 1490441 bytes, checksum: 2d30a406bff1dc2da6123c10c2e44617 (MD5) / Made available in DSpace on 2016-01-25T17:04:18Z (GMT). No. of bitstreams: 1 pedrolinojonecoutinho.pdf: 1490441 bytes, checksum: 2d30a406bff1dc2da6123c10c2e44617 (MD5) Previous issue date: 2015-05-22 / Atualmente, um dos mais graves problemas relacionado com a poluição ambiental é a contaminação da água, principalmente, devido ao descarte inadequado de resíduos industriais e agrícolas. Deste modo, o presente trabalho tem como objetivo avaliar a remoção de Al3+ em meio aquoso por adsorção em casca de arroz in natura e cascas de arroz modificadas quimicamente. Para tal foram realizados experimentos em escala laboratorial e em batelada verificando-se a remoção da concentração de alumínio quando em contato com o biossorvente até atingir-se o tempo de equilíbrio. Foram utilizados três tipos de biosorventes: casca de arroz in natura (CN) e cascas modificadas quimicamente com solução ácida (CTA) e solução básica (CTB), a fim de avaliar qual delas proporciona a melhor adsorção do alumínio solúvel. Foram realizados três planejamentos estatísticos de experimento fatoriais 33 Box-Behnken para a otimização das condições de modo a encontrar a melhor resposta para a adsorção do alumínio em sistema de batelada. As condições ótimas encontradas através do planejamento para CTB, que teve melhor resposta de adsorção do metal, com uma recuperação de 97,60 % foram: (X1) massa 0,1500 g; (X2) volume 7,5 mL e (X3) tempo de contato solução de alumínio e o biossorvente de 5 horas. Durante a execução do planejamento de experimento fatorial 33 Box-Behken foi mantido em comum o pH 5,0 e a concentração do Al3+ em 30,0 mg L-1. A técnica analítica utilizada para quantificação do metal na solução após a adsorção foi à espectrometria de absorção atômica com chama (F AAS). Encontradas as condições ótimas prosseguiu-se com os demais estudos, a exemplo do estudo de efeito de pH considerando a faixa de 4,0 a 6,0 e estudos de isotermas de adsorção adotando os modelos de Langmuir e Freundlich. O melhor pH para adsorção do Al3+ foi 5,0 sendo que acima deste valor ocorrem perdas por precipitação em forma de hidróxido. Essas perdas devem ser levadas em consideração durante os cálculos após a sua adsorção. O modelo de Langmuir, foi o que teve valores de coeficientes de determinação (R2) altos e relativamente próximo de 1, apresentando desse modo o devido ajuste das isotermas linearizadas, comparativamente ao modelo de Freundlich que apresentou coeficientes de determinação inferiores. Desse modo o biossorvente usado neste trabalho mostrou-se eficiente na remoção do alumínio, tanto na forma modificada CTB assim como na forma CN, em detrimento da forma modificada CTA que apresentou sempre baixas capacidades adsortivas de Al3+, nas condições estabelecidas e consideradas ótimas neste trabalho. / Nowadays, one of the most serious problems related to environmental pollution is the contamination of water, mainly, through improper discard of industrial and agricultural waste. Thus, the main objective of the present study is to evaluate the removal of Al3+ in aqueous environment by adsorption on raw rice husk and chemically modified rice husks. For this purpose experiments were performed on a laboratory scale batch and was seen that actually occur a removing of aluminum concentration when it is in contact with the biosorbent until to be achieved an balance adsorption time. Three types of biosorbents were used: raw rice husk (R-RH), and chemically modified rice husk with an acid (CM-RHA) and basic solution (CM-RHB) in order to evaluate which one has the best adsorption of soluble aluminum. Three statistical 33 Box-Behnken experiment designs were performed to optimize the conditions in order to find the best answer for aluminum adsorption in batch system. The optimum conditions found for CM-RHB, which had better metal adsorption response, with a recovery about 97.60% were: (X1) mass 0.1500 g; (X2) volume 7.5 mL and (X3) contact time between the aluminum solution and biosorbent which is about 5 hours. During the execution of factorial experiment design 33 Box-Behnken was kept in common, the value of pH (pH ~ 5) and the concentration of Al3+ 30,0 mg L-1. The flame atomic absorption spectrometric (F AAS), was used to quantify the metal in solution after its adsorption. Found the optimal conditions, were carried on other studies, such as the study of effect pH where was considered the range of 4,0 to 6,0 as well the study of adsorption isotherms adopting the models of Langmuir and Freundlich. The best pH for adsorption of Al3+ was 5,0 and above this value there are losses by precipitation in form of hydroxide. These losses should be taken into account during the calculations after its adsorption. The Langmuir model was the one which showed high determination coefficient values (R2) and relatively close to 1, thereby presenting the best fit of the linearized isotherms compared to Freundlich model which showed lower determination coefficient values. Thus the biosorbent used in this work was efficient in aluminum removal, either in CM-RHB form and R-RH form rather than CM-RHA form that always showed low adsorption capacity of Al3+ under the conditions established and considered optimum in this work.