DSpace Communidade:
http://www.repositorio.ufc.br/handle/riufc/478
2020-02-27T15:48:34ZDetecção de falhas de curto-circuito em motores de indução trifásicos usando classificadores baseados em protótipos
http://www.repositorio.ufc.br/handle/riufc/49906
Título: Detecção de falhas de curto-circuito em motores de indução trifásicos usando classificadores baseados em protótipos
Autor(es): Sousa, Diego Perdigão
Abstract: Three-phase induction motors are one of the most important equipment of modern industry.
However, in many situations, these equipment are subject to inappropriate conditions such as in
environments with high temperatures and humidity, abrupt variations of load above speciﬁed,
excessive vibrations, among others. These conditions make motors more susceptible to various
failures, whether external or internal, which are obviously undesirable in industrial processes. In
this context, the predictive maintenance plays a relevant role, where the detection and correct
diagnosis of failures in a timely manner leads to increasing the useful life of the motor and,
consequently, to the reduction of costs with production stoppage due to corrective maintenance.
Considering these factors, this dissertation proposes a methodology for detecting short-circuit
failures in three-phase induction motors, which involves prototypes-based algorithms. To this
end, both unsupervised techniques - such as the K-means and supervised algorithm, such as the
LVQ (Learning Vector Quantization) family classiﬁers are used.
The methodology starts with the seeking of the optimal number of prototypes from the unsu-
pervised analysis of clusters and techniques clustering validation. Then, the prototypes that
were found are used in the supervised training of various classiﬁers of the LVQ family. The
inﬂuence that each type of clustering validation criterion exerts on the various LVQ classiﬁers
implemented is deeply evaluated. In particular, the GRLVQ (Generalized Relevance Learning
Vector Quantization) classiﬁer obtained the best results where it presented a maximum classiﬁca-
tion rate of 98.3%, with the Dunn and Silhouette criteria standing out as the most efﬁcient in
determining the optimal quantity of prototypes.2019-02-11T00:00:00ZHistograma de quadratura em tomografia de estado quântico de máxima verossimilhança
http://www.repositorio.ufc.br/handle/riufc/49531
Título: Histograma de quadratura em tomografia de estado quântico de máxima verossimilhança
Autor(es): Silva, José Leonardo Esteves da
Abstract: Quantum state tomography aims to determine the quantum state of a system from measured data
and is an essential tool for quantum information science. When dealing with continuous variable
quantum states of light, tomography is often done by measuring the ﬁeld amplitudes at different
optical phases using homodyne detection. The quadrature-phase homodyne measurement outputs
a continuous variable, so to reduce the computational cost of tomography, researchers often
discretize the measurements. We show that this can be done without signiﬁcantly degrading the
ﬁdelity between the estimated state and the true state. In this thesis, we study different strategies
for determining the histogram bin widths. We show that computation time can be signiﬁcantly
reduced with little loss in the ﬁdelity of the estimated state when the measurement operators
corresponding to each histogram bin are integrated over the bin width.2018-07-20T00:00:00ZAnálise do viés em tomografia do estado quântico de máxima verossimilhança
http://www.repositorio.ufc.br/handle/riufc/48906
Título: Análise do viés em tomografia do estado quântico de máxima verossimilhança
Autor(es): Silva, George Barbosa da
Abstract: Maximum likelihood quantum state tomography yields estimators that are
consistent, provided that the likelihood model is correct, but the maximum likelihood
estimators may have bias for any nite data set. The bias of an estimator is the di erence
between the expected value of the estimate and the true value of the parameter being
estimated. This paper investigates bias in the widely used maximum likelihood quantum
state tomography. Our goal is to understand how the amount of bias depends on factors
such as the purity of the true state, the number of measurements performed, and the
number of di erent bases in which the system is measured. For that, we perform numer-
ical experiments that simulate optical homodyne tomography under various conditions,
perform tomography, and estimate the bias in the purity of the estimated state. We nd
that estimates of higher purity states exhibit considerable bias, such that the estimates
have lower purity than the true states.2016-07-13T00:00:00ZEnsino de Engenharia via e-learning: relação entre uso de simuladores, ambiente virtual e design instrucional do conteúdo
http://www.repositorio.ufc.br/handle/riufc/48095
Título: Ensino de Engenharia via e-learning: relação entre uso de simuladores, ambiente virtual e design instrucional do conteúdo
Autor(es): Araújo, Régia Talina Silva
Abstract: The main objective of this research is to statistically assess the learning effectiveness in e-learning engineering courses. Given this context, the following research question was outlined: Do educational resources available in on-line engineering courses, such as Virtual Learning Environment (VLE), Design of Content and simulations, impact on learning effectiveness?
To fulfill this objective, we have observed these learning resources in two classes of distance
learning courses: a pilot team electrical-engineering course offered by a Federal institute of
technological education, part of the Brazilian Open-Tech-School Network E-TEC Brazil, in the
discipline of Electric Power Systems; and an extension course on Image Processing and Pattern
Recognition using Python / NumPy – PIRP. As a procedure, the use of an interactive simulator
was observed in the respective discipline of the Electrotechnical course, as well as the AVA
analysis and instructional content design. The development of the simulator was based on virtual
reality techniques and cognitive-learning-basis theories, in order to provide training courses in
power distribution theme. In PIRP, AVA and instructional content design were also analyzed, and
programming simulations available in Adessowiki environments were observed. An evaluation
instrument was also applied to the students of both courses. In order to test the effectiveness
learning of these resources, we have used the following statistical methods: descriptive and
multivariate statistics, namely the principal component analysis (PCA), as well as reliability
analysis and canonical correlation analysis (CCA). The results of the statistical analysis confirmed
the importance of the three investigated educational resources for learning effectiveness. In
this research, we introduced descriptive statistical analysis in terms of a representative index
for each question of the educational resource under test. Thus, it possible to observe relevant
characteristics to be considered in online engineering courses. Another important contribution of
the descriptive statistical analysis is that it enabled identifying opportunities to increase student
satisfaction by illustrating institutional strengths and weaknesses. The multivariate analysis by
canonical correlation over PIRP course data confirmed the positive contribution of the tested
resources through AVA Functionality, Design of Content and Simulations in the effectiveness
learning.2017-06-14T00:00:00Z