abstract: The emergence of new technologies as well as a fresh look at analyzing existing processes have given rise to a new type of response characteristic, known as a profile. Profiles are useful when a quality variable is functionally dependent on one or more explanatory, or independent, variables. So, instead of observing a single measurement on each unit or product a set of values is obtained over a range which, when plotted, takes the shape of a curve. Traditional multivariate monitoring schemes are inadequate for monitoring profiles due to high dimensionality and poor use of the information stored in functional form leading to very large variance-covariance matrices. Profile monitoring has become an important area of study in statistical process control and is being actively addressed by researchers across the globe. This research explores the understanding of the area in three parts. A comparative analysis is conducted of two linear profile-monitoring techniques based on probability of false alarm rate and average run length (ARL) under shifts in the model parameters. The two techniques studied are control chart based on classical calibration statistic and a control chart based on the parameters of a linear model. The research demonstrates that a profile characterized by a parametric model is more efficient monitoring scheme than one based on monitoring only the individual features of the profile. A likelihood ratio based changepoint control chart is proposed for detecting a sustained step shift in low order polynomial profiles. The test statistic is plotted on a Shewhart like chart with control limits derived from asymptotic distribution theory. The statistic is factored to reflect the variation due to the parameters in to aid in interpreting an out of control signal. The research also looks at the robust parameter design study of profiles, also referred to as signal response systems. Such experiments are often necessary for understanding and reducing the common cause variation in systems. A split-plot approach is proposed to analyze the profiles. It is demonstrated that an explicit modeling of variance components using generalized linear mixed models approach has more precise point estimates and tighter confidence intervals. / Dissertation/Thesis / Ph.D. Industrial Engineering 2010

A disseminação de sistemas de aquisição de dados sobre a qualidade e o desempenho de produtos e processos de fabricação deu origem a novos tipos de dados. Dado funcional é um conjunto de dados que formam um perfil ou uma curva. No perfil, a característica de qualidade é uma função dependente de uma ou mais variáveis exploratórias ou independentes. A análise de dados funcionais é um tema de pesquisa recente praticado em diversas áreas do conhecimento. Na indústria, os dados funcionais aparecem no controle de qualidade. A ausência de métodos apropriados a dados funcionais pode levar ao uso de métodos ineficientes e reduzir o desempenho e a qualidade de um produto ou processo. A análise de dados funcionais através de métodos multivariados pode ser inadequada devido à alta dimensionalidade e estruturas de variância e covariância dos dados. O desenvolvimento teórico de métodos para a análise de dados funcionais na área de Engenharia da Qualidade encontra-se defasado em relação ao potencial de aplicações práticas. Este trabalho identificou a existência dos dados funcionais tratados por métodos ineficientes. Os métodos atuais para controle de qualidade de dados são adaptados a situações específicas, conforme o tipo de dado funcional e a fase do monitoramento. Este trabalho apresenta propostas para métodos de análise de dados funcionais aplicáveis a questões relevantes da área de pesquisa em Engenharia da Qualidade, tais como: (i) o uso da análise de variância em experimentos com dados funcionais; (ii) gráficos de controle para monitoramento de perfis; e (iii) a análise e seleção de perfis de fornecedores em projetos inovadores. / The dissemination of data acquisition systems on the quality and performance of products and manufacturing process has given rise to new types of data. Functional data are a collection of data points organized as a profile or curve. In profile, the quality characteristic is a function dependent on one or more exploratory or independent variables. The functional data analysis is a recent research topic practiced in various areas of knowledge. In industry, the functional data appears in quality control. The lack of suitable methods can lead to use of inefficient methods and reducing the performance and quality of a product or process. The analysis of functional data by multivariate methods may be inadequate due to the high dimensionality and variance and covariance structures of the data. The development of theoretical methods for the analysis of functional data in Quality Engineering area is lagged behind the potential for practical applications. This work identified the existence of functional data processed by inefficient methods. Current methods for data quality control are adapted to specific situations, depending on the type of functional data and the phase of monitoring. This paper presents proposals for functional data analysis methods applicable to relevant research questions in the area of Quality Engineering such as: (i) the use of analysis of variance in experiments with functional data; (ii) control charts for monitoring profiles; and (iii) the analysis and selection of supplier profiles on innovative projects.

A disseminação de sistemas de aquisição de dados sobre a qualidade e o desempenho de produtos e processos de fabricação deu origem a novos tipos de dados. Dado funcional é um conjunto de dados que formam um perfil ou uma curva. No perfil, a característica de qualidade é uma função dependente de uma ou mais variáveis exploratórias ou independentes. A análise de dados funcionais é um tema de pesquisa recente praticado em diversas áreas do conhecimento. Na indústria, os dados funcionais aparecem no controle de qualidade. A ausência de métodos apropriados a dados funcionais pode levar ao uso de métodos ineficientes e reduzir o desempenho e a qualidade de um produto ou processo. A análise de dados funcionais através de métodos multivariados pode ser inadequada devido à alta dimensionalidade e estruturas de variância e covariância dos dados. O desenvolvimento teórico de métodos para a análise de dados funcionais na área de Engenharia da Qualidade encontra-se defasado em relação ao potencial de aplicações práticas. Este trabalho identificou a existência dos dados funcionais tratados por métodos ineficientes. Os métodos atuais para controle de qualidade de dados são adaptados a situações específicas, conforme o tipo de dado funcional e a fase do monitoramento. Este trabalho apresenta propostas para métodos de análise de dados funcionais aplicáveis a questões relevantes da área de pesquisa em Engenharia da Qualidade, tais como: (i) o uso da análise de variância em experimentos com dados funcionais; (ii) gráficos de controle para monitoramento de perfis; e (iii) a análise e seleção de perfis de fornecedores em projetos inovadores. / The dissemination of data acquisition systems on the quality and performance of products and manufacturing process has given rise to new types of data. Functional data are a collection of data points organized as a profile or curve. In profile, the quality characteristic is a function dependent on one or more exploratory or independent variables. The functional data analysis is a recent research topic practiced in various areas of knowledge. In industry, the functional data appears in quality control. The lack of suitable methods can lead to use of inefficient methods and reducing the performance and quality of a product or process. The analysis of functional data by multivariate methods may be inadequate due to the high dimensionality and variance and covariance structures of the data. The development of theoretical methods for the analysis of functional data in Quality Engineering area is lagged behind the potential for practical applications. This work identified the existence of functional data processed by inefficient methods. Current methods for data quality control are adapted to specific situations, depending on the type of functional data and the phase of monitoring. This paper presents proposals for functional data analysis methods applicable to relevant research questions in the area of Quality Engineering such as: (i) the use of analysis of variance in experiments with functional data; (ii) control charts for monitoring profiles; and (iii) the analysis and selection of supplier profiles on innovative projects.

A disseminação de sistemas de aquisição de dados sobre a qualidade e o desempenho de produtos e processos de fabricação deu origem a novos tipos de dados. Dado funcional é um conjunto de dados que formam um perfil ou uma curva. No perfil, a característica de qualidade é uma função dependente de uma ou mais variáveis exploratórias ou independentes. A análise de dados funcionais é um tema de pesquisa recente praticado em diversas áreas do conhecimento. Na indústria, os dados funcionais aparecem no controle de qualidade. A ausência de métodos apropriados a dados funcionais pode levar ao uso de métodos ineficientes e reduzir o desempenho e a qualidade de um produto ou processo. A análise de dados funcionais através de métodos multivariados pode ser inadequada devido à alta dimensionalidade e estruturas de variância e covariância dos dados. O desenvolvimento teórico de métodos para a análise de dados funcionais na área de Engenharia da Qualidade encontra-se defasado em relação ao potencial de aplicações práticas. Este trabalho identificou a existência dos dados funcionais tratados por métodos ineficientes. Os métodos atuais para controle de qualidade de dados são adaptados a situações específicas, conforme o tipo de dado funcional e a fase do monitoramento. Este trabalho apresenta propostas para métodos de análise de dados funcionais aplicáveis a questões relevantes da área de pesquisa em Engenharia da Qualidade, tais como: (i) o uso da análise de variância em experimentos com dados funcionais; (ii) gráficos de controle para monitoramento de perfis; e (iii) a análise e seleção de perfis de fornecedores em projetos inovadores. / The dissemination of data acquisition systems on the quality and performance of products and manufacturing process has given rise to new types of data. Functional data are a collection of data points organized as a profile or curve. In profile, the quality characteristic is a function dependent on one or more exploratory or independent variables. The functional data analysis is a recent research topic practiced in various areas of knowledge. In industry, the functional data appears in quality control. The lack of suitable methods can lead to use of inefficient methods and reducing the performance and quality of a product or process. The analysis of functional data by multivariate methods may be inadequate due to the high dimensionality and variance and covariance structures of the data. The development of theoretical methods for the analysis of functional data in Quality Engineering area is lagged behind the potential for practical applications. This work identified the existence of functional data processed by inefficient methods. Current methods for data quality control are adapted to specific situations, depending on the type of functional data and the phase of monitoring. This paper presents proposals for functional data analysis methods applicable to relevant research questions in the area of Quality Engineering such as: (i) the use of analysis of variance in experiments with functional data; (ii) control charts for monitoring profiles; and (iii) the analysis and selection of supplier profiles on innovative projects.

Profile monitoring is a relatively new approach in quality control best used when the process data follow a profile (or curve). The majority of previous studies in profile monitoring focused on the parametric modeling of either linear or nonlinear profiles under the assumption of the correct model specification. Our work considers those cases where the parametric model for the family of profiles is unknown or, at least uncertain. Consequently, we consider monitoring Poisson profiles via three methods, a nonparametric (NP) method using penalized splines, a nonparametric (NP) method using wavelets and a semi parametric (SP) procedure that combines both parametric and NP profile fits. Our simulation results show that SP method is robust to the common problem of model misspecification of the user's proposed parametric model. We also showed that Haar wavelets are a better choice than the penalized splines in situations where a sudden jump happens or the jump is edgy. In addition, we showed that the penalized splines are better than wavelets when the shape of the profiles are smooth. The proposed novel techniques have been applied to a real data set and compare with some state-of-the arts.

We have seen an increased interest and extensive research in the monitoring of a process over time whose characteristics are represented mathematically in functional forms such as profiles. Most of the current techniques require all of the data for each profile to determine the state of the process. Thus, quality engineers from industrial processes such as agricultural, aquacultural, and chemical cannot make process corrections to the current profile that are essential for correcting their processes at an early stage. In addition, the focus of most of the current techniques is on the statistical significance of the parameters or features of the model instead of the practical significance, which often relates to the actual quality characteristic. The goal of this research is to provide alternatives to address these two main concerns. First, we study the use of a Shewhart type control chart to monitor within profiles, where the central line is the predictive mean profile and the control limits are formed based on the prediction band. Second, we study a statistic based on a non-linear mixed model recognizing that the model leads to correlations among the estimated parameters. / Doctor of Philosophy / Checking the stability over time of the quality of a process which is best expressed by a relationship between a quality characteristic and other variables involved in the process has received increasing attention. The goal of this research is to provide alternative methods to determine the state of such a process. Both methods presented here are compared to the current methodologies. The first method will allow us to monitor a process while the data is still being collected. The second one is based on the quality characteristic of the process and takes full advantage of the model structure. Both methods seem to be more robust than the current most well-known method.

Profile monitoring is a relatively new approach in quality control best used where the process data follow a profile (or curve) at each time period. The essential idea for profile monitoring is to model the profile via some parametric, nonparametric, and semiparametric methods and then monitor the fitted profiles or the estimated random effects over time to determine if there have been changes in the profiles. The majority of previous studies in profile monitoring focused on the parametric modeling of either linear or nonlinear profiles, with both fixed and random effects, under the assumption of correct model specification. Our work considers those cases where the parametric model for the family of profiles is unknown or at least uncertain. Consequently, we consider monitoring profiles via two techniques, a nonparametric technique and a semiparametric procedure that combines both parametric and nonparametric profile fits, a procedure we refer to as model robust profile monitoring (MRPM). Also, we incorporate a mixed model approach to both the parametric and nonparametric model fits. For the mixed effects models, the MMRPM method is an extension of the MRPM method which incorporates a mixed model approach to both parametric and nonparametric model fits to account for the correlation within profiles and to deal with the collection of profiles as a random sample from a common population. For each case, we formulated two Hotelling's T 2 statistics, one based on the estimated random effects and one based on the fitted values, and obtained the corresponding control limits. In addition,we used two different formulas for the estimated variancecovariance matrix: one based on the pooled sample variance-covariance matrix estimator and a second one based on the estimated variance-covariance matrix based on successive differences. A Monte Carlo study was performed to compare the integrated mean square errors (IMSE) and the probability of signal of the parametric, nonparametric, and semiparametric approaches. Both correlated and uncorrelated errors structure scenarios were evaluated for varying amounts of model misspecification, number of profiles, number of observations per profile, shift location, and in- and out-of-control situations. The semiparametric (MMRPM) method for uncorrelated and correlated scenarios was competitive and, often, clearly superior with the parametric and nonparametric over all levels of misspecification. For a correctly specified model, the IMSE and the simulated probability of signal for the parametric and theMMRPM methods were identical (or nearly so). For the severe modelmisspecification case, the nonparametric andMMRPM methods were identical (or nearly so). For the mild model misspecification case, the MMRPM method was superior to the parametric and nonparametric methods. Therefore, this simulation supports the claim that the MMRPM method is robust to model misspecification. In addition, the MMRPM method performed better for data sets with correlated error structure. Also, the performances of the nonparametric and MMRPM methods improved as the number of observations per profile increases since more observations over the same range of X generally enables more knots to be used by the penalized spline method, resulting in greater flexibility and improved fits in the nonparametric curves and consequently, the semiparametric curves. The parametric, nonparametric and semiparametric approaches were utilized for fitting the relationship between torque produced by an engine and engine speed in the automotive industry. Then, we used a Hotelling's T 2 statistic based on the estimated random effects to conduct Phase I studies to determine the outlying profiles. The parametric, nonparametric and seminonparametric methods showed that the process was stable. Despite the fact that all three methods reach the same conclusion regarding the –in-control– status of each profile, the nonparametric and MMRPM results provide a better description of the actual behavior of each profile. Thus, the nonparametric and MMRPM methods give the user greater ability to properly interpret the true relationship between engine speed and torque for this type of engine and an increased likelihood of detecting unusual engines in future production. Finally, we conclude that the nonparametric and semiparametric approaches performed better than the parametric approach when the user's model is misspecified. The case study demonstrates that, the proposed nonparametric and semiparametric methods are shown to be more efficient, flexible and robust to model misspecification for Phase I profile monitoring in a practical application. Thus, our methods are robust to the common problem of model misspecification. We also found that both the nonparametric and the semiparametric methods result in charts with good abilities to detect changes in Phase I data, and in charts with easily calculated control limits. The proposed methods provide greater flexibility and efficiency than current parametric methods used in profile monitoring for Phase I that rely on correct model specification, an unrealistic situation in many practical problems in industrial applications. / Ph. D.

The volume of data acquired in production systems continues to expand. Emerging imaging technologies, such as machine vision systems (MVSs) and 3D surface scanners, diversify the types of data being collected, further pushing data collection beyond discrete dimensional data. These large and diverse datasets increase the challenge of extracting useful information. Unfortunately, industry still relies heavily on traditional quality methods that are limited to fault detection, which fails to consider important diagnostic information needed for process recovery. Modern measurement technologies should spur the transformation of statistical process control (SPC) to provide practitioners with additional diagnostic information. This dissertation focuses on how MVSs and 3D laser scanners can be further utilized to meet that goal. More specifically, this work: 1) reviews image-based control charts while highlighting their advantages and disadvantages; 2) integrates spatiotemporal methods with digital image processing to detect process faults and estimate their location, size, and time of occurrence; and 3) shows how point cloud data (3D laser scans) can be used to detect and locate unknown faults in complex geometries. Overall, the research goal is to create new quality control tools that utilize high density data available in manufacturing environments to generate knowledge that supports decision-making beyond just indicating the existence of a process issue. This allows industrial practitioners to have a rapid process recovery once a process issue has been detected, and consequently reduce the associated downtime. / Ph. D.

With the advancement of sensor technology and data processing capacities, various types of high volume data are available for process monitoring and prognostics in manufacturing systems. In a broaching process, a multi-toothed broaching tool removes material from the workpiece by sequential engagement and disengagement of multiple cutting edges. The quality of the final part, including the geometric integrity and surface finish, is highly dependent upon the broaching tool condition. Though there has been a considerable amount of research on tool condition monitoring and prognostics for various machining processes, the broaching process is unique in the following aspects: 1) a broaching process involves multiple cutting edges, which jointly contribute to the final part quality; 2) the resharpening and any other process adjustments to the tool can only be performed with the whole broaching tool or at least a whole segment of the tool replaced. The overarching goal of this research is to explore how engineering knowledge can be used to improve process monitoring and prognostics for a complex manufacturing process like broaching. This dissertation addresses the needs for developing new monitoring and prognostics approaches based on various types of data. Specifically, the research effort focuses on 1) the use of in-situ force profile data for real-time process monitoring and fault diagnosis, 2) degradation characterization for broaching processes on an individual component level based on image processing; and 3) system-level degradation modeling and remaining useful life prediction for broaching processes based on multiple images. / Ph. D. / Big data have been providing both opportunities and challenges for product quality assurance and improvement in modern manufacturing systems. In aerospace industry, broaching processes are one of the most important manufacturing processes as they are used to produce the turbine discs in the jet engine. Nonconforming turbine disc quality, either in terms of compromised surface finish or geometry accuracy, will lead to malfunction or even catastrophic failures in the aircraft engines. One of the major sources that lead to nonconforming product quality is excessive tool wear accumulation and other abrupt malfunctions of the broaching tools. In broaching processes, multiple cutting edges are sequentially pushed or pulled through the workpiece, and each cutting edge is responsible to shape the workpiece into a specific intermediate shaped contour. Therefore, a broaching process can be regarded as a multistage manufacturing process with variation propagating through the multiple cutting edges. The overarching goal of this dissertation is to explore how process knowledge can be used to improve process monitoring and prognostics for a complex manufacturing process like broaching. This dissertation focuses on the quality assurance and improvement for broaching processes which includes: 1) timely abrupt process fault detection; 2) tool performance degradation quantification; and 3) remaining tool life prediction, which contributes to both methodological development and practical applications in advanced sensing analytics in manufacturing systems.

Submitted by Silvana Teresinha Dornelles Studzinski (sstudzinski) on 2016-06-14T15:45:26Z No. of bitstreams: 1 Rafael Herzer_.pdf: 1397260 bytes, checksum: a9941bd0932b535c5699cf0b35a815dc (MD5) / Made available in DSpace on 2016-06-14T15:45:26Z (GMT). No. of bitstreams: 1 Rafael Herzer_.pdf: 1397260 bytes, checksum: a9941bd0932b535c5699cf0b35a815dc (MD5) Previous issue date: 2016-02-29 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O presente trabalho tem como objetivo propor um método de monitoramento para gestão de portfólio, o qual, através de um sistema multi-critério e de um modelo econométrico, identifica variações no cenário econômico e em indicadores da empresa sendo então, a partir do monitoramento de resíduos, possível definir o momento exato para alteração de portfólio de produtos. A Gestão de Portfólio de produtos vem atraindo interesse dos gestores das corporações e deste modo, é difícil encontrar alguma organização que não possua uma carteira de produtos e projetos para gerenciar. A Gestão de Portfólio trata das decisões de alocação de recursos e de como ficará a carteira dos produtos atuais, sendo uma ferramenta de extrema importância para o resultado, principalmente financeiro, das organizações. Encontram-se vários métodos na literatura para realizar a Gestão do Portfólio, dentre os quais modelos financeiros, modelos probabilísticos financeiros, modelos de escores e checklists, abordagens de hierarquia analítica, abordagens comportamentais e abordagens de mapas ou diagrama de bolhas são os mais relevantes. Mesmo existindo diversos métodos na literatura para realizar a gestão do portfólio, não há consenso sobre qual método deve ser utilizado em cada etapa específica. Esses métodos também necessitam de intervenção dos gestores, levando em consideração que geralmente as informações disponíveis para tomada de decisão não são completas ou exatas. Para este estudo, foi realizado um estudo de simulação Monte Carlo para avaliar a sensibilidade dos diversos elementos que compõem o método. Os resultados mostraram taxas de alarmes falsos e tempo médio para detectar a mudança semelhantes a estudos anteriores. Esse processo de gestão e tomada de decisão é considerado complexo para os gestores das empresas, uma vez que o portfólio necessita ser periodicamente revisado, buscando sempre maximização de valor e equilíbrio ideal de produtos no mercado. Por fim, a aplicação do modelo é ilustrada por um caso real, utilizando dados fornecidos por uma empresa multinacional do segmento agrícola. / Product Portfolio Management is attracting the interest of the managers of the corporations. With the competitiveness of the market, it is difficult to find an organization that does not have a portfolio of products to manage. The Portfolio Management deals with resource allocation decisions and how will the portfolio of current products be compouse, being an extremely important tool for the result, especially financial, for the organizations. This process of management and decision making is considered complex to company managers, since the portfolio needs to be periodically revised, always seeking to maximize value and correct balance of products on the market. There are several methods in the literature to perform portfolio management, among which financial models, financial probabilistic models, scores and checklists models, analytical hierarchy of approaches, behavioral approaches and approaches map or diagram bubbles are the most relevant. While there are several methods in the literature to make the portfolio management, there is no consensus about which method should be used in each specific step. These methods also require the intervention of managers, taking into account that generally available information for decision-making are not complete or accurate. This paper aims to propose a method, which, through a multi-criteria system containing an econometric model, identifies changes in the economic environment and business indicators and then, from the profile monitoring, can set the exacly time for change portfolio of products. We performed the Monte Carlo simulation study to assess the sensitivity of the various parts that make up the method. The results showed false alarm rate and mean time to detect changes similar to previous studies. Finally, the application of the model is illustrated by a real case using data provided by a multinational company, agricultural segment.