The presented thesis deals with the issue of statistical quality control of a specific production process. The thesis presents a range of statistical tools that can be used to identify the factors causing a high proportion of non-standard pieces. The diploma thesis practically introduces the application of these quality management tools to the production process characterized by an increased proportion of non-standard pieces, in which the factors causing their occurrence are unknown, and as following the thesis in detail introduces the approach how to detect these factors. The last part of the work summarizes the recommendations handed over to the company in order to verify the conclusions of the thesis.

<p><a>Disruptive events during recent decades have highlighted the vulnerabilities of complex systems of systems to disruption propagation: </a>Disruptions that start in one part of a system and can propagate to other parts. Such examples include: Fire spreading in building complexes and forests; plant/crop diseases in agricultural production systems; propagating malware in computer networks and cyber-physical systems; and disruptions in supply networks. The impacts of disruption propagation are devastating, with fire causing annual US$23 billion loss in the US alone, plant diseases/crop reducing agricultural productivity 20% to 40% annually, and computer malware causing up to US$2.3 billion loss per event (as a conservative estimate). These problems, the response to disruption propagation (<a>RDP</a>) problems, are challenging due to the involvement of different problem aspects and their complex dynamics. To better design and control the responses to disruption propagation, a general framework and problem-solving guideline for the RDP problems is necessary.<br></p><p><br></p> <p> </p> <p>To address the aforementioned challenge, this research develops the Collaborative Response to Disruption Propagation (<a>CRDP</a>) unifying framework to classify, categorize, and characterize the different aspects of the RDP problems. The CRDP framework allows analogical reasoning across the different problem contexts, such as the examples mentioned above. Three main components applicable to the investigate RDP problems are identified and characterized: (1) The client system as the victims; (2) The response mechanisms as the rescuers/protectors; and (3) The disruption propagation as the aggressors/attackers. This allows further characterization of the complex interactions between the components, which augments the design and control decisions for the response mechanisms to better respond to the disruptions. The new Covering Lines of Collaboration (<a>CLOC</a>) principle, consisting of three guidelines, is developed to analyze the system state and guide the response decisions. The first CLOC guideline recommends the network modeling of potential disruption propagation directions, creating a complex network for better situation awareness and analysis. The second CLOC guideline recommends the analysis of the propagation-restraining effects due to the existence of the response mechanisms, and utilizing this interaction in optimizing response decisions. The third CLOC guideline recommends the development of collaboration protocols between the response decisions to maximize the coverage of response against disruption propagation.</p><p><br></p> <p> </p> <p>The CRDP framework and the CLOC principle are validated with three RDP case studies: (1) Detection of unknown disruptions; (2) Strategic prevention of unexpected disruptions; (3) Teaming and coordination of repair agents against recurring disruptions. Formulations, analytics, and protocols specific to each case are developed. TIE/CRDP, a new version of the Teamwork Integration Evaluator (<a>TIE</a>) software, is developed to simulate the complex interactions and dynamics of the CRDP components, the response decision protocols, and their performance. The evaluator is capable of simulating and evaluating the complex interactions and dynamics of the CRDP components and the response decision protocols. <a>Experiment results indicate that advanced CLOC-based decisions significantly outperform the baseline and less advanced protocols for all three cases, with performance superiority of 9.7-32.8% in case 1; 31.1%-56.6% in case 2; 2.1%-12.1% for teaming protocols, and at least 50% for team coordination protocols in case 3.</a></p>

No description available.

As the new paradigm of data collection enabled by the advancements in wireless technology and digital electronics, small sensing devices have started to be used in everyday life. These devices are capable of sensing, computing, communicating, and forming a wireless sensor network (WSN) which is necessary to provide sensing services and to monitor various conditions. In addition to WSNs, the idea of Internet of Things (IoT) has started to draw more attention. IoT is defined as an interconnection between identifiable devices within the internet for sensing and monitoring processes. This dissertation addresses the development of wireless network and process control for two challenging IoT applications, namely smart agriculture and industrial lyophilization.<div><br><div>Smart agriculture refers to the concept of using modern technology to increase the quantity and quality of agricultural products. This dissertation presents a novel hybrid large-area IoT network by combining the low-power wide-area network (LPWAN) as well as ultra-low-power wireless personal area networks (WPAN) that delivers wide-area coverage while maintaining low-power operation. </div><div><br></div><div>Lyophilization is the process of removing water content from a material with the objective of increasing its stability and, hence, its shelf life. Continuous inline process tracking is imperative to a successful lyophilization process in industrial pharmaceuticals. To address this need, this dissertation presents two wireless sensing systems that are capable of monitoring lyophilization process with detailed design and demonstration<br></div></div>

Researchers now have unprecedented access to free Landsat data, enabling detailed monitoring of the Earth's land surface and vegetation.  There are gaps in the data, due in part to cloud cover. The gaps are aperiodic and localized, forcing any detailed multitemporal analysis based on Landsat data to compensate.   Harmonic regression approximates Landsat data for any point in time with minimal training images and reduced storage requirements.  In two study areas in North Carolina, USA, harmonic regression approaches were least as good at simulating missing data as STAR-FM for images from 2001.  Harmonic regression had an R^2"0.9 over three quarters of all pixels. It gave the highest R_Predicted^2 values on two thirds of the pixels.  Applying harmonic regression with the same number of harmonics to consecutive years yielded an improved fit, R^2"0.99 for most pixels.   We next demonstrate a change detection method based on exponentially weighted moving average (EWMA) charts of harmonic residuals. In the process, a data-driven cloud filter is created, enabling use of partially clouded data.  The approach is shown capable of detecting thins and subtle forest degradations in Alabama, USA, considerably finer than the Landsat spatial resolution in an on-the-fly fashion, with new images easily incorporated into the algorithm.  EWMA detection accurately showed the location, timing, and magnitude of 85% of known harvests in the study area, verified by aerial imagery.   We use harmonic regression to improve the precision of dynamic forest parameter estimates, generating a robust time series of vegetation index values.  These values are classified into strata maps in Alabama, USA, depicting regions of similar growth potential.  These maps are applied to Forest Service Forest Inventory and Analysis (FIA) plots, generating post-stratified estimates of static and dynamic forest parameters.  Improvements to efficiency for all parameters were such that a comparable random sample would require at least 20% more sampling units, with the improvement for the growth parameter requiring a 50% increase. These applications demonstrate the utility of harmonic regression for Landsat data.  They suggest further applications in environmental monitoring and improved estimation of landscape parameters, critical to improving large-scale models of ecosystems and climate effects. / Ph. D.

La conception de souches efficacespour la production de composés d’intérêt offredes potentiels immenses qui restent trop peu exploitéspar manque de lien entre les étapes d’optimisationde la conception de souche et celle dela conduite du bioprocédé.Pour combler ce manque, cette thèse proposeune description des bioprocédés intégrant pleinementle fonctionnement interne des microorganismesimpliqués dans la production decomposés d’intérêt. Cette description permetd’optimiser simultanément la souche et le procédépour maximiser la production d’un composéd’intérêt en respectant les contraintes attachéesà ces deux étapes.Dans un premier temps, une nouvelle classe demodélisation de bioprocédés est développée, àl’interface entre les modèles intracellulaires degestion de ressources et les modèles macroscopiquesusuellement utilisés dans la commandede bioprocédés en bioréacteurs. Dans un secondtemps, des contraintes liées à l’implémentationbiologique de la stratégie de contrôle sont intégréesau problème. Ceci permet d’obtenir uneconception plus réaliste du point de vue de l’ingénieriedes génomes. Enfin, la dernière partiede la thèse montre que la méthodologie présentéejusqu’alors sur un modèle agrégé peut êtreétendue à des représentations détaillées du comportementdes micro-organismes. / The design of efficient strains forthe production of compounds of interest offerstremendous potentials that remain insufficientlyexploited due to the lack of link between theoptimization stages of strain design and that ofbioprocess control.This thesis proposes a description of bioprocessesthat fully integrates the internal functioningof micro-organisms involved in the productionof compounds of interest. This descriptionallows the strain and process to be optimizedsimultaneously to maximize the productionof a compound of interest while respecting theconstraints attached to these two stages.First, a new bioprocess modelling class is developedat the interface between intracellular resourceallocation models and macroscopic modelscommonly used in bioprocess control. In asecond stage, constraints linked to the biologicalimplementation of the control strategy are integratedinto the problem. This provides a morerealistic genome engineering design. Finally, thelast part of the thesis shows that the methodologypresented so far on an aggregate model canbe extended to detailed representations of thebehaviour of micro-organisms.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Civil Engineering; and, (M.S.)--Massachusetts Institute of Technology, Dept. of Architecture, 1975. / Bibliography: leaves 80-81. / M.S. / M.Arch.

No description available.

No description available.

[pt] É importante investigar formas de aumentar a eficiência de gráficos de controle estatístico de processos por atributos, porque o controle por atributos tipicamente exige tamanhos de amostra muito grandes para se obter um tempo rápido de sinalização de desvios no processo, e estes tamanhos de amostra podem ser altamente indesejáveis ou mesmo inviáveis em diversas situações práticas. Os esquemas adaptativos, que consistem em variar um ou mais dos parâmetros do gráfico de controle (tamanho de amostra, intervalo de tempo entre amostras e abertura dos limites de controle) de acordo com a informação precedente fornecida pelo gráfico (isto é, de acordo com a posição do último ponto registrado), proporcionam, para a detecção de desvios pequenos a moderados no processo, uma melhor eficiência que os esquemas tradicionais, de parâmetros fixos. Enquanto que esquemas adaptativos têm sido propostos para gráficos de controle por variáveis, para gráficos por atributos - apesar da necessidade e oportunidade acima mencionada de investigar formas de aumentar a eficiência de gráficos de controle estatístico de processos por atributos - há uma lacuna na literatura, a ser preenchida. Esta é a motivação do presente trabalho, que propõe um esquema adaptativo para gráficos decontrole por atributos, aplicável tanto a gráficos de np como de c, variando todos os parâmetros do gráfico. Desenvolve-se o modelo matemático para cálculo das medidas de desempenho; este modelo foi implementado em uma planilha eletrônica, o que permitiu analisar quantitativamente o desempenho do esquema em uma larga gama de casos, comparando-o ainda com o desempenho de gráficos tradicionais (com parâmetros fixos), e com o de outros esquemas adaptativos, com menor número de parâmetros variando. O esquema proposto mostrou-se sensivelmente mais eficiente na maioria das situações de interesse, respondendo portanto à necessidade, que motivou o trabalho, de buscar formas de aumentar a eficiência de gráficos por atributos. Resultados adicionais do trabalho são: a identificação das situações em que cada esquema é mais eficiente (ou do esquema mais eficiente em cada situação) e conjuntos de valores recomendados para os parâmetros dos gráficos em cada situação. Estes resultados facilitam a operacionalização da ferramenta para uso na prática. / [en] It is important to investigate ways to improve the efficiency of the statistical control charts for attributes, because the controls for attributes typically require very large sample sizes for quick detection of changes in the process, and these sample sizes can be extremely undesirable or eventually unfeasible in many practical situations. The adaptive schemes, which vary one or more parameters of the control charts (sample size, sampling interval and control limit width) according to the most recent information about the process (the position of the last sample point in the chart), were proven to be more efficient than the traditional (fixed parameters) scheme in detecting small to moderate changes in the process. Although there is a large volume of work on adaptive schemes applied to variables control charts, there is a lack of adaptive schemes for control charts for attributes in the literature, despite the already mentioned need and opportunity of investigating ways to improve the efficiency of the statistical control charts for attributes. This is the motivation of this work. A fully adaptive scheme is proposed for both np and charts. The mathematical model is developed for the measures of performance; this model was implemented in a spreadsheet, and used for the quantitative evaluation of the scheme efficiency, and comparison with the fixed-parameter (traditional) scheme and also with other adaptive schemes, with only one or two parameters variable. The proposed scheme proved to be significantly more efficient in the majority of the situations of interest, answering the necessity of finding ways to improve the efficiency of control charts for attributes. Additional results of this work are: the identification of the situations in which each of the schemes here analyzed is the most efficient (or, the other way around: indication of the most efficient scheme for each situation), and a set of recommended values for the design parameters for each scheme, in each situation. These results are meant to make easier the adoption of the scheme in practice, increasing thereby its usefulness. / [es] El desarrollo de técnicas que aumenten la eficiencia de gráficos de control estadístico de procesos por atributos tiene gran importancia porque el control por atributos típicamente exige tamaños de muestra muy grandes para obtener un tiempo rápido de señalización de desvíos en el proceso. En diversas situaciones prácticas estos tamaños de muestra pueden ser altamente indeseables o incluso inviables. Los esquemas adaptativos, que consisten en variar uno o más parámetros del gráfico de control (tamaño de muestra, intervalo de tiempo entre muestras y abertura de los límites de control) de acuerdo con la información precedente ofrecida por el gráfico (esto es, de acuerdo con la posición del último punto registrado), proporcionan, para la detección de desvíos pequeños a moderados, una mayor eficiencia que los esquemas tradicionales, de parámetros fijos. Mientras que los esquemas adaptativos han sido propuestos para gráficos de control por variables, para gráficos por atributos - a pesar de la necesidad y oportunidad mencionada de investigar formas de aumentar la eficiencia de gráficos de control estadístico de procesos por atributos - no existen muchas referencias en la literatura. Esta es la motivación del presente trabajo, que propone un esquema adaptativo para gráficos de control por atributos, aplicable tanto a gráficos de np como de c, variando todos los parámetros del gráfico. Se desarrolla el modelo matemático para el cálculo de las medidas de desempeño; este modelo fue implementado en una planilla electrónica, que permitió analizar cuantitativamente el desempeño del esquema en una amplia gama de casos, comparando con el desempeño de gráficos tradicionales (con parámetros fijos), y con el de otros esquemas adaptativos, con menor número de parámetros variando. El esquema propuesto se mostró sensiblemente más eficiente en la mayoría de las situaciones de interés que motivaron el trabajo. Resultados adicionales del trabajo son: la identificación de las situaciones en que cada esquema es más eficiente (o del esquema más eficiente en cada situación) y conjuntos de valores recomendados para los parámetros de los gráficos en cada situación. Estos resultados facilitan la operacionalización de la ferramenta para uso en la práctica.