<p>At Acreo AB located in Norrköping, Sweden, research and development in the field of organic electronics have been conducted since 1998. Several electronic devices and systems have been realized. In late 2003 a commercial printing press was installed to test large scale production of these devices. Prior to the summer of 2005 the project made significant progress. As a step towards industrialisation, the variability and yield of the printing process needed to bee studied. A decision to implement Statistical Process Control (SPC) and Design of Experiments (DOE) to evaluate and improve the process was taken.</p><p>SPC has been implemented on the EC-patterning step in the process. A total of 26 Samples were taken during the period October-December 2005. An - and s-chart were constructed from these samples. The charts clearly show that the process is not in statistical control. Investigations of what causes the variation in the process have been performed. The following root causes to variation has been found: </p><p>PEDOT:PSS-substrate sheet resistance and poorly cleaned screen printing drums. </p><p>After removing points affected by root causes, the process is still not in control. Further investigations are needed to get the process in control. Examples of where to go next is presented in the report. In the DOE part a four factor full factorial experiment was performed. The goal with the experiment was to find how different factors affects switch time and life length of an electrochromic display. The four factors investigated were: Electrolyte, Additive, Web speed and Encapsulation. All statistical analysis was performed using Minitab 14. The analysis of measurements from one day and seven days after printing showed that:</p><p>- Changing Electrolyte from E230 to E235 has small effect on the switch time</p><p>- Adding additives Add1 and Add2 decreases the switch time after 1 and 7 days</p><p>- Increasing web speed decreases the switch time after 1 and 7 days </p><p>- Encapsulation before UV-step decreases the switch time after 7 days</p>

Fluidized bed dryers are used in the pharmaceutical industry for the batch drying of pharmaceutical granulate. Maintaining optimal hydrodynamic conditions throughout the drying process is essential to product quality. Due to the complex interactions inherent in the fluidized bed drying process, mechanistic models capable of identifying these optimal modes of operation are either unavailable or limited in their capabilities. Therefore, empirical models based on experimentally generated data are relied upon to study these systems.<p> Principal Component Analysis (PCA) and Partial Least Squares (PLS) are multivariate statistical techniques that project data onto linear subspaces that are the most descriptive of variance in a dataset. By modeling data in terms of these subspaces, a more parsimonious representation of the system is possible. In this study, PCA and PLS are applied to data collected from a fluidized bed dryer containing pharmaceutical granulate. <p>System hydrodynamics were quantified in the models using high frequency pressure fluctuation measurements. These pressure fluctuations have previously been identified as a characteristic variable of hydrodynamics in fluidized bed systems. As such, contributions from the macroscale, mesoscale, and microscales of motion are encoded into the signals. A multiresolutional decomposition using a discrete wavelet transformation was used to resolve these signals into components more representative of these individual scales before modeling the data. <p>The combination of multiresolutional analysis with PCA and PLS was shown to be an effective approach for modeling the conditions in the fluidized bed dryer. In this study, datasets from both steady state and transient operation of the dryer were analyzed. The steady state dataset contained measurements made on a bed of dry granulate and the transient dataset consisted of measurements taken during the batch drying of granulate from approximately 33 wt.% moisture to 5 wt.%. Correlations involving several scales of motion were identified in both studies.<p> In the steady state study, deterministic behavior related to superficial velocity, pressure sensor position, and granulate particle size distribution was observed in PCA model parameters. It was determined that these properties could be characterized solely with the use of the high frequency pressure fluctuation data. Macroscopic hydrodynamic characteristics such as bubbling frequency and fluidization regime were identified in the low frequency components of the pressure signals and the particle scale interactions of the microscale were shown to be correlated to the highest frequency signal components. PLS models were able to characterize the effects of superficial velocity, pressure sensor position, and granulate particle size distribution in terms of the pressure signal components. Additionally, it was determined that statistical process control charts capable of monitoring the fluid bed hydrodynamics could be constructed using PCA<p>In the transient drying experiments, deterministic behaviors related to inlet air temperature, pressure sensor position, and initial bed mass were observed in PCA and PLS model parameters. The lowest frequency component of the pressure signal was found to be correlated to the overall temperature effects during the drying cycle. As in the steady state study, bubbling behavior was also observed in the low frequency components of the pressure signal. PLS was used to construct an inferential model of granulate moisture content. The model was found to be capable of predicting the moisture throughout the drying cycle. Preliminary statistical process control models were constructed to monitor the fluid bed hydrodynamics throughout the drying process. These models show promise but will require further investigation to better determine sensitivity to process upsets.<p> In addition to PCA and PLS analyses, Multiway Principal Component Analysis (MPCA) was used to model the drying process. Several key states related to the mass transfer of moisture and changes in temperature throughout the drying cycle were identified in the MPCA model parameters. It was determined that the mass transfer of moisture throughout the drying process affects all scales of motion and overshadows other hydrodynamic behaviors found in the pressure signals.

At Acreo AB located in Norrköping, Sweden, research and development in the field of organic electronics have been conducted since 1998. Several electronic devices and systems have been realized. In late 2003 a commercial printing press was installed to test large scale production of these devices. Prior to the summer of 2005 the project made significant progress. As a step towards industrialisation, the variability and yield of the printing process needed to bee studied. A decision to implement Statistical Process Control (SPC) and Design of Experiments (DOE) to evaluate and improve the process was taken. SPC has been implemented on the EC-patterning step in the process. A total of 26 Samples were taken during the period October-December 2005. An - and s-chart were constructed from these samples. The charts clearly show that the process is not in statistical control. Investigations of what causes the variation in the process have been performed. The following root causes to variation has been found: PEDOT:PSS-substrate sheet resistance and poorly cleaned screen printing drums. After removing points affected by root causes, the process is still not in control. Further investigations are needed to get the process in control. Examples of where to go next is presented in the report. In the DOE part a four factor full factorial experiment was performed. The goal with the experiment was to find how different factors affects switch time and life length of an electrochromic display. The four factors investigated were: Electrolyte, Additive, Web speed and Encapsulation. All statistical analysis was performed using Minitab 14. The analysis of measurements from one day and seven days after printing showed that: - Changing Electrolyte from E230 to E235 has small effect on the switch time - Adding additives Add1 and Add2 decreases the switch time after 1 and 7 days - Increasing web speed decreases the switch time after 1 and 7 days - Encapsulation before UV-step decreases the switch time after 7 days

Fluidized bed dryers are used in the pharmaceutical industry for the batch drying of pharmaceutical granulate. Maintaining optimal hydrodynamic conditions throughout the drying process is essential to product quality. Due to the complex interactions inherent in the fluidized bed drying process, mechanistic models capable of identifying these optimal modes of operation are either unavailable or limited in their capabilities. Therefore, empirical models based on experimentally generated data are relied upon to study these systems.<p> Principal Component Analysis (PCA) and Partial Least Squares (PLS) are multivariate statistical techniques that project data onto linear subspaces that are the most descriptive of variance in a dataset. By modeling data in terms of these subspaces, a more parsimonious representation of the system is possible. In this study, PCA and PLS are applied to data collected from a fluidized bed dryer containing pharmaceutical granulate. <p>System hydrodynamics were quantified in the models using high frequency pressure fluctuation measurements. These pressure fluctuations have previously been identified as a characteristic variable of hydrodynamics in fluidized bed systems. As such, contributions from the macroscale, mesoscale, and microscales of motion are encoded into the signals. A multiresolutional decomposition using a discrete wavelet transformation was used to resolve these signals into components more representative of these individual scales before modeling the data. <p>The combination of multiresolutional analysis with PCA and PLS was shown to be an effective approach for modeling the conditions in the fluidized bed dryer. In this study, datasets from both steady state and transient operation of the dryer were analyzed. The steady state dataset contained measurements made on a bed of dry granulate and the transient dataset consisted of measurements taken during the batch drying of granulate from approximately 33 wt.% moisture to 5 wt.%. Correlations involving several scales of motion were identified in both studies.<p> In the steady state study, deterministic behavior related to superficial velocity, pressure sensor position, and granulate particle size distribution was observed in PCA model parameters. It was determined that these properties could be characterized solely with the use of the high frequency pressure fluctuation data. Macroscopic hydrodynamic characteristics such as bubbling frequency and fluidization regime were identified in the low frequency components of the pressure signals and the particle scale interactions of the microscale were shown to be correlated to the highest frequency signal components. PLS models were able to characterize the effects of superficial velocity, pressure sensor position, and granulate particle size distribution in terms of the pressure signal components. Additionally, it was determined that statistical process control charts capable of monitoring the fluid bed hydrodynamics could be constructed using PCA<p>In the transient drying experiments, deterministic behaviors related to inlet air temperature, pressure sensor position, and initial bed mass were observed in PCA and PLS model parameters. The lowest frequency component of the pressure signal was found to be correlated to the overall temperature effects during the drying cycle. As in the steady state study, bubbling behavior was also observed in the low frequency components of the pressure signal. PLS was used to construct an inferential model of granulate moisture content. The model was found to be capable of predicting the moisture throughout the drying cycle. Preliminary statistical process control models were constructed to monitor the fluid bed hydrodynamics throughout the drying process. These models show promise but will require further investigation to better determine sensitivity to process upsets.<p> In addition to PCA and PLS analyses, Multiway Principal Component Analysis (MPCA) was used to model the drying process. Several key states related to the mass transfer of moisture and changes in temperature throughout the drying cycle were identified in the MPCA model parameters. It was determined that the mass transfer of moisture throughout the drying process affects all scales of motion and overshadows other hydrodynamic behaviors found in the pressure signals.

We consider a principal components based decomposition of the expected value of the multivariate quadratic loss function, i.e., MQL. The principal components are formed by scaling the original data by the contents of the loss constant matrix, which defines the economic penalty associated with specific variables being off their desired target values. We demonstrate the extent to which a subset of these ``loss-scaled principal components", i.e., LSPC, accounts for the two components of expected MQL, namely the trace-covariance term and the off-target vector product. We employ the LSPC to solve a robust design problem of full and reduced dimensionality with deterministic models that approximate the true solution and demonstrate comparable results in less computational time. We also employ the LSPC to construct a test statistic called loss-scaled T^2 for multivariate statistical process control. We show for one case how the proposed test statistic has faster detection than Hotelling's T^2 of shifts in location for variables with high weighting in the MQL. In addition we introduce a principal component based decomposition of Hotelling's T^2 to diagnose the variables responsible for driving the location and/or dispersion of a subgroup of multivariate observations out of statistical control. We demonstrate the accuracy of this diagnostic technique on a data set from the literature and show its potential for diagnosing the loss-scaled T^2 statistic as well.

Statistical process control (SPC) is a toolbox to detect changes in the output of a process distribution. It can serve as a valuable resource to maintain high quality in a manufacturing process. This report is based on the work on evaluating and implementing methods for SPC in the process of chromatography instrument manufacturing at GE Healthcare, Umeå. To handle low volume and non-normally distributed process output data, non-parametric methods are considered. Eight control charts, three for for Phase I analysis, and five for Phase II analysis, are evaluated in this study. The usability of the charts are assessed based on ease of interpretation and the performance to detect distributional changes. The later is evaluated with simulations. The result of the project is the implementation of the RS/P-chart, suggested by Capizzi et al (2013), for Phase I analysis. Of the considered Phase I methods (and simulation scenarios), the RS/P-chart has the highest overall probability, of detecting a variety of distributional changes. Further, the RS/P-chart is easily interpreted, facilitating the analysis. For Phase II analysis, the use of two control charts, one based on the Mann-Whitney U statistic, suggested by Chakraborti et al (2008), and one on the Mood test statistic for dispersion, suggested by Ghute et al (2014), have been implemented. These are chosen mainly based on the ease of interpretation. To reduce the detection time for changes in the process distribution, the change-point chart based on the Cramer Von Mises statistic, suggested by Ross et al (2012), could be used instead. Using single observations, instead of larger samples, this chart is updated more frequently. However, this efficiently increases the false alarm rate and the chart is also considered much more difficult to interpret for the SPC practitioner. / Statistisk processkontroll (SPC) är en samling verktyg för att upptäcka förändringar, i fördelningen, hos utfallen i en process. Det kan fungera som en värdefull resurs för att upprätthålla en hög kvalitet i en tillverkningsprocess. Denna rapport är baserad på arbetet med att utvärdera och implementera metoder för SPC i en monteringsprocess av kromatografiinstrument på GE Healthcare, Umeå. Åtta styrdiagram, tre för för fas I analys, och fem för fas II analys, studeras i denna rapport. Användbarheten hos styrdiagrammen bedöms efter hur enkla de är att tolka och förmågan att upptäcka fördelningsförändringar. Den senare utvärderas med simuleringar. Resultatet av projektet är införandet av RS/P-metod, utvecklad av Capizzi et al (2013), för analysen i fas I. Av de utvärderade metoderna, (och simuleringsscenarier), har RS/P-diagrammet den högsta övergripande sannolikheten, för att upptäcka en mängd olika fördelningsförändringar. Vidare är metodens grafiska diagram lätt att tolka, vilket underlättar analysen. För fas II analys, har två styrdiagram, ett baserat på Mann-Whitney's U teststatistika, som föreslagits av Chakraborti et al (2008), och ett på Mood's teststatistika för spridning, som föreslagits av Ghute et al (2014), implementerats. Styrkan i dessa styrdiagram ligger främst i dess enkla tolkning. För snabbare identifiering av processförändringar kan styrdiagrammet baserat på Cramer von Mises teststatistika, som föreslagits av Ross et al (2012), användas. Baserat på enskilda observationer, istället för stickprov, har styrdiagrammet en högre uppdateringsfrekvens. Detta leder dock till ett ökat antal falska larm och styrdiagrammet anses dessutom vara avsevärt mycket svårare att tolka för SPC-utövaren.

Η παρούσα διατριβή αφορά στην ολοκλήρωση των μεθόδων Στατιστικού Ελέγχου Ποιότητας με τεχνικές Μηχανικής Μάθησης, για την καλύτερη εξυπηρέτηση των αναγκών των σύγχρονων επιχειρήσεων. Προς αυτή την κατεύθυνση, έγινε αρχικά μια λεπτομερής ανασκόπηση της σχετικής βιβλιογραφίας για τον εντοπισμό και την αναγνώριση των σημαντικότερων ελλείψεων του Στατιστικού Ελέγχου Ποιότητας. Στη συνέχεια, χρησιμοποιήθηκαν τεχνικές Μηχανικής Μάθησης για την αντιμετώπιση των παραπάνω ελλείψεων. Πιο συγκεκριμένα, προτάθηκε μια μεθοδολογία για αναγνώριση μέσων μετατοπίσεων σε αυτοσυσχετιζόμενα δεδομένα πολυμεταβλητών διεργασιών, τα οποία συναντώνται πολύ συχνά σε πραγματικές διεργασίες. Η προτεινόμενη μεθοδολογία δοκιμάζεται και ελέγχεται ως προς την απόδοσή της και την ικανότητά της για εφαρμογή σε δεδομένα διαφορετικής φύσεως σε δυο μελέτες περίπτωσης. Τα αποτελέσματα από τις μελέτες αυτές είναι ενθαρρυντικά καθώς επιτεύχθηκαν αρκετά υψηλά ποσοστά επιτυχών αναγνωρίσεων μέσων μετατοπίσεων. Η διατριβή ολοκληρώνεται με παράθεση μιας σειράς συμπερασμάτων, ανάδειξη της συμβολής της προτεινόμενης μεθοδολογίας και υπόδειξη μελλοντικών ερευνητικών κατευθύνσεων για την επέκτασή της. / This thesis concerns the integration of Statistical Quality Control methods with Machine Learning techniques for covering contemporary business needs. The proposed approach took into account a thorough review of the literature, which identified the major shortcomings of Statistical Quality Control. A consideration of Machine Learning techniques with respect to the above shortcomings was then performed. More specifically, a methodology was proposed for identifying mean shifts in auto-correlated multivariate data processes, which occurs very often in real processes. The proposed approach was tested through two different case studies for its performance and ability to implement data of different type. The results of these case studies were encouraging as quite high rates were achieved for the successful recognition of mean shifts. The thesis concludes by listing a series of findings, highlighting the contribution of the proposed approach and suggesting a series of future research directions.

Submitted by Claudia Rocha (claudia.rocha@udesc.br) on 2017-12-11T15:50:46Z No. of bitstreams: 1 PGEF17MA072.pdf: 1540009 bytes, checksum: 0a6a0c91bcb17bc18be105a058020b75 (MD5) / Made available in DSpace on 2017-12-11T15:50:46Z (GMT). No. of bitstreams: 1 PGEF17MA072.pdf: 1540009 bytes, checksum: 0a6a0c91bcb17bc18be105a058020b75 (MD5) Previous issue date: 2017-02-20 / Some organizations in the Brazilian forestry industry do not follow the development pace of other industries, or the adoption rate of management and quality methodologies and tools. As result, the low quality and high instability of processes create a lot of waste in the sector. This is intensified when it happens during the harvesting, one of the activities that most contributes to the production costs of a forestry business. This study aimed to evaluate the predictability of timber harvesting process based on critical points identified in the thinning and harvest operations. For this, research was divided in three stages. The first one was mapping the harvesting process and the elaboration of fluxograms. The second stage was the identification, through interviews, and assessment of critical points using the Failure Mode and Effect Analysis (FMEA) and Pareto chart. The third stage was the evaluation of the process using Statistical Process Control (SPC) through attributes and variable in the main failures. In the interviews conducted with workers, seven critical points were identified: damage to the remaining trees, sorting, stump height, dirt in load, knot on the second log, tree left in the field and safety. The evaluation with Pareto chart showed that 80% of the failures identified during harvest are attributed to three causes: damage to the remaining trees, sorting and stump height. The FMEA analysis showed that the failure with highest risk index was knots on the second log, followed by safety. For evaluation with the SPC, damage to remaining trees, sorting and stump height were selected. In general, control charts showed that the forest harvest process was considered unstable and unpredictable, even though it is within the limits defined by the company / No setor florestal brasileiro, algumas organizações não acompanham o ritmo do desenvolvimento de outros setores e da adoção de ferramentas e metodologias de gestão da qualidade. Em função deste atraso, a baixa qualidade e alta instabilidade dos processos, geram grandes desperdícios nas organizações. O fato se agrava quando ocorre na operação de colheita da madeira, uma das atividades que mais contribui nos custos de produção da empresa florestal. Este trabalho objetivou avaliar o processo de colheita florestal de uma empresa quanto a sua previsibilidade com base nos pontos críticos identificados nas operações de desbaste e corte raso. Para isso, a pesquisa foi dividida em três etapas, a primeira foi o mapeamento do processo de colheita e elaboração de fluxogramas. A segunda etapa foi a identificação, por meio de entrevistas, e avaliação dos pontos críticos utilizando a metodologia Failure Mode and Effect Analysis – FMEA e gráfico de Pareto. A terceira fase foi a avaliação do processo empregando o Controle Estatístico do Processo (CEP) por atributos e variáveis nas principais falhas. Nas entrevistas realizadas com os colaboradores, sete pontos críticos foram apontados: danos às árvores remanescentes, sortimento, altura de toco, sujeira na carga, nó na 2ª tora, árvores deixadas no talhão e segurança. Na avaliação com o gráfico de Pareto, observou-se que cerca de 80% dos problemas identificados na colheita da empresa são atribuídos às três primeiras causas supracitadas. A análise com o FMEA indicou que a falha com maior índice de risco foi o nó na 2ª tora, seguido da segurança. Para a avaliação com o CEP, selecionou-se os pontos críticos: danos às árvores remanescentes, sortimento e altura de toco. De maneira geral, os gráficos de controle indicaram que o processo da colheita florestal, apesar de estar dentro dos limites especificados pela empresa, foi considerado instável e não previsível

abstract: Many products undergo several stages of testing ranging from tests on individual components to end-item tests. Additionally, these products may be further "tested" via customer or field use. The later failure of a delivered product may in some cases be due to circumstances that have no correlation with the product's inherent quality. However, at times, there may be cues in the upstream test data that, if detected, could serve to predict the likelihood of downstream failure or performance degradation induced by product use or environmental stresses. This study explores the use of downstream factory test data or product field reliability data to infer data mining or pattern recognition criteria onto manufacturing process or upstream test data by means of support vector machines (SVM) in order to provide reliability prediction models. In concert with a risk/benefit analysis, these models can be utilized to drive improvement of the product or, at least, via screening to improve the reliability of the product delivered to the customer. Such models can be used to aid in reliability risk assessment based on detectable correlations between the product test performance and the sources of supply, test stands, or other factors related to product manufacture. As an enhancement to the usefulness of the SVM or hyperplane classifier within this context, L-moments and the Western Electric Company (WECO) Rules are used to augment or replace the native process or test data used as inputs to the classifier. As part of this research, a generalizable binary classification methodology was developed that can be used to design and implement predictors of end-item field failure or downstream product performance based on upstream test data that may be composed of single-parameter, time-series, or multivariate real-valued data. Additionally, the methodology provides input parameter weighting factors that have proved useful in failure analysis and root cause investigations as indicators of which of several upstream product parameters have the greater influence on the downstream failure outcomes. / Dissertation/Thesis / Ph.D. Electrical Engineering 2011

Submitted by Carla Segatto Strini Paixão (ca_paixao@live.com) on 2018-06-14T16:31:26Z No. of bitstreams: 1 Tese_Carla.pdf: 1311888 bytes, checksum: 5200851f1f76ad510e670ba7f602977e (MD5) / Rejected by Neli Silvia Pereira null (nelisps@fcav.unesp.br), reason: Solicitamos que realize correção na submissão seguindo as orientações abaixo: 1 - Está faltando o certificado de aprovação no arquivo em pdf. 2 - Encaminhar novamente o arquivo todo, com o certificado de aprovação, em pdf. Agradecemos a compreensão. on 2018-06-15T17:24:28Z (GMT) / Submitted by Carla Segatto Strini Paixão (ca_paixao@live.com) on 2018-06-25T00:54:36Z No. of bitstreams: 1 Tese_Carla-2.pdf: 1920361 bytes, checksum: 1a15616e02ddd976f02a2606d549e8f8 (MD5) / Approved for entry into archive by Neli Silvia Pereira null (nelisps@fcav.unesp.br) on 2018-06-25T18:46:11Z (GMT) No. of bitstreams: 1 paixao_css_dr_jabo.pdf: 1920361 bytes, checksum: 1a15616e02ddd976f02a2606d549e8f8 (MD5) / Made available in DSpace on 2018-06-25T18:46:11Z (GMT). No. of bitstreams: 1 paixao_css_dr_jabo.pdf: 1920361 bytes, checksum: 1a15616e02ddd976f02a2606d549e8f8 (MD5) Previous issue date: 2017-12-19 / A colheita mecanizada de soja é uma operação fundamental para a finalização de seu ciclo produtivo, porém a ausência de metodologias eficientes para quantificação das perdas, não tem contribuído para melhoria continua deste processo agrícola. Associado a isto, ainda existem inúmeros fatores que podem influenciar a qualidade da operação da colheita de soja, sendo difíceis de serem analisados e interpretados, pois as relações entre as variáveis são complexas. Para tentar suavizar essa complexidade, abordagens multivariadas, como Análise de Componentes Principais (ACP) e Análise Fatorial (AF), podem ser uma alternativa para extrair informações da base de dados gerados durante a colheita. E por fim, para monitorar a operação com base no nível de qualidade que a mesma está sendo realizada, o controle estatístico de processo, com uso de gráficos multivariados torna-se essencial, para as variáveis que possuem correlação. Diante disto, objetivou-se neste trabalho determinar as variáveis que mais afetam a qualidade operacional da colheita mecanizada de soja, por meio de análises uni e multivariada. O trabalho foi realizado em março de 2016, em área agrícola no município de Ribeirão Preto-SP, sendo utilizada uma colhedora da marca John Deere, modelo 1470, com sistema de trilha do tipo tangencial e separação por saca-palhas. O processo foi considerado incapaz de manter as perdas da colheita mecanizada de soja em níveis aceitáveis durante toda a operação para as duas armações. A análise fatorial permitiu a seleção de quatro indicadores da colheita mecanizada de soja, explicando 76,4% da variância total. As cartas de controle multivariadas foram mais eficazes para determinar a não aleatoriedade no monitoramento de processo com variáveis correlacionadas. / The mechanical harvesting of soybeans is a fundamental operation to the end of its productive cycle, but the absence of efficient methodologies to quantify losses has not contributed to the continuous improvement of this agricultural process. Associated with this, there are still many factors that can influence the quality of the soybean harvesting operation, being difficult to analyze and interpret because the relationships among the variables are complex. To attempt to soften this complexity, multivariate approaches such as Principal Component Analysis and Factorial Analysis may be an alternative to extract information from the database generated during harvesting. Finally, to monitor the operation based on the level of quality that is being performed, the statistical process control, using multivariate graphs becomes essential, for the variables that have correlation. Therefore, the aimed of this study was to determine the variables that most affect the operational quality of the soybean mechanized harvest, through uni and multivariate analyzes. The work was carried out in March 2016, in an agricultural area in the city of Ribeirão Preto-SP, using a John Deere brand model 1470 harvester with a tangential type track system and straw picking. The process was found to be unable to keep soybean harvest losses at acceptable levels throughout the operation for the two frames. The factorial analysis allowed the selection of four indicators of the soybean mechanized harvest, explaining 76.4% of the total variance. Multivariate control charts were more effective in determining non-randomness in process monitoring with correlated variables.