Gráficos de T² para o monitoramento de processos multivariados autocorrelacionados

In which refers to statistical process control, the analysis of univariate cases is not enough for many types of company, being necessary to resort to multivariate cases. Besides, it is usually supposed that the observations are independent. However, the violation of this hypothesis indicates the existence of autocorrelation in the process. In this work, by a basic quantitative approach for an exploratory and experimental research, the study target are the multivariate autocorrelated control charts, using Hotteling T². The ARL values were collected by simulations of a computational program on FORTRAN language, with objective of studying the charts properties, in addition to compare with the

Previsão da capacidade de carga de estacas escavadas em solo arenoso da região de Araquari-SC

In Geotechnical engineering the foundation projects depend on the bearing capacity and the acceptable displacements. One of the possible ways to predict the bearing capacity of foundations is through semi-empirical statistical methods which correlate in-situ tests (SPT and CPT). The piles breaking loads are defined by the interpretation of the load x head displacement curve and the experimental data acquired through the load test. In this work it is studied the behavior of bored piles executed in the Araquari/SC region, comparing the bearing capacity values predicted by the methods DECOURT & QUARESMA MODIFICADO (1996), AOKI & VELLOSO MODIFICADO MONTEIRO (2000), MILITITISKY E ALVES (1985), DECOURT & QUARESMA (1978), MÉTODO DE AOKI & VELLOSO (1975) e PHILOPANNAT (1986), with the results of the load test, evaluating their differences and discussing parameters that have direct effects on the prediction

Aplicação de métodos com múltiplos critérios em uma empresa petroquímica

The corporate world is increasingly competitive, and companies need to go deep into the routines and work them in order to understand them fully. The market is demanding more than simple improvements that bring advances - of small or great expression; however, in a longer term it will no longer meet the ideology of the market. Companies aimed at the world class must focus on projects that will continually bring returns to the company. As previously mentioned, understanding the processes in minute details is of paramount importance, as this knowledge can be acquired by analyzing the decisions that are necessary during the process. Once the complexity increases, the quantity and difficulty of the criteria that influence them grow accordingly. At this time, methods and tools that assist decisionmaking processes can be used as, besides being able to provide the best decision methods of MCDA (Multiple Criteria Decision Aid), they provide clear and assertive understanding of the whole decision process. In developing this study, we sought to explore the AHP (Analytic Hierarchy Process) method (a MCDA method) in the choice of access service, featured by the support service used to reach and be the basis of repairs in places of difficult access. This work proposes a study of the quantitative modeling approach in a real routine activity for a Brazilian petrochemical company. Decision-making processes are explored when we seek to analyze not only the decision makers but also what directly influences them on the use of the AIJ method. Once this is achieved, the understanding of decision-making is substantiated

Métodos de previsão auto-regressivo aplicado a uma série de volume de produção de caminhões

The national truck fleet has expanded strongly in recent decades. However, due to fluctuations in the demand that the market is exposed, it needed up making more effective strategic decisions of automakers. These decisions are made after an evaluation of guaranteed sales forecasts. This work aims to generate an annual forecast of truck production by Box and Jenkins methodology. They used annual data for referring forecast modeling from the year 1957 to 2014, which were obtained by the National Association of Motor Vehicle Manufacturers (Anfavea). The model used was Autoregressive Integrated Moving Average (ARIMA) and can choose the best model for the series under study, and the ARIMA (2,1,3) as representative for conducting truck production forecast