Notre objectif a été d'étudier l'apport d'une analyse temps-fréquence par transformée en ondelettes dans l'évaluation stabilométrique de la stabilité posturale de sujets présentant une pathologie neurologique. A partir d'un corpus de données incluant 199 patients présentant une pathologie neurologique et 41 sujets témoins, nous avons étudié la reproductibilité et la validité des paramètres stabilométriques. Nous avons déterminé les valeurs d'erreur de mesure de paramètres issus d'une analyse temps-fréquence, d'une analyse fréquentielle par méthode FFT et d'une analyse linéaire du signal, chez des patients hémiplégiques vasculaires. Pour la moyenne de 3 acquisitions de 30 secondes, nous proposons de privilégier l'utilisation de paramètres issus d'une analyse temps-fréquence qui sont les plus reproductibles. L'évaluation stabilométrique a montré un intérêt au moins équivalent aux critères cliniques dans la prédiction du risque de chute chez le patient hémiplégique vasculaire. Le modèle de prédiction reste cependant insuffisant pour être appliqué de manière isolé en pratique clinique, où une évaluation multi-modale du patient reste nécessaire. Une étude des déterminants neurologiques et biomécaniques de la distribution du contenu fréquentiel a été réalisée avec l'analyse en ondelettes. Nous avons montré qu'en l'absence de tremblement postural caractérisé, la distribution du contenu fréquentiel était essentiellement liée au degré d'instabilité posturale. Nos perspectives sont d'approfondir l'intérêt de l'analyse temps-fréquence par transformée en ondelettes, qui permet d'étudier à la fois la stabilité posturale et les mécanismes de contrôle sous-jacents. / Our objective was to study the interest of time-frequency analysis by wavelet transform in stabilometric assessment of postural sway in neurological patients. From a corpus of data including 199 patients with neurological disease and 41 control subjects, we studied the reliability and validity of stabilometric parameters. We determined the error measurement from a time-frequency analysis, a FFT analysis and a linear analysis of the stabilogram, in stroke patients. For the average of 3 acquisitions of 30 seconds, we propose to use the time-frequency parameters that are most reliable. Stabilometric assessment showed an interest at least equivalent to clinical criteria in predicting the risk of falls in the stroke patient. The prediction model remains insufficient to be applied alone in clinical practice, where a multi-modal patient assessment is still required. A study of the neurological and biomechanical determinants of the distribution of frequency content was carried out with the wavelet analysis. We have shown that in the absence of postural tremor, the distribution of frequency content was mostly related to the degree of postural instability. Our outlook is to deepen the time-frequency analysis by wavelet transform, which allows to study both postural stability and the underlying control mechanisms.

No description available.

Ce manuscrit traite de l'étude métrologique d'un spectromètre Terahertrz (THz), de la caractérisation et de la quantification de produits pulvérulents par spectroscopie THz. Le sujet a été orienté afin de fournir les études préliminaires nécessaires à une thématique visant la détection de pesticides sur les aliments, à savoir la caractérisation métrologique de l'appareil (études de répétabilité, sensibilité...) ainsi que la faisabilité de la quantification de produits et l'application de méthodes chimiométriques lors du prétraitement des spectres. La thèse est ordonnée autour de trois publications. La première publication consiste en une revue de littérature servant à faire le point sur les applications concrètes existant en spectroscopie THz. La seconde porte sur la première partie de notre travail, à savoir la caractérisation métrologique du spectromètre THz sur lequel nous avons effectué nos mesures. Enfin, la troisième porte sur l'aspect quantification de la spectroscopie THz et la possibilité d'appliquer les mêmes principes et techniques chimiométriques qu'en spectroscopie infrarouge. / This manuscript concerns the metrological study of a Terahertz (THz) spectrometer, the characterization and the quantification of powder products by THz spectroscopy. The subject has been aimed to give preliminary analysis to a wider thematic of pesticides detection on aliments, which implies the metrological characterization of the device (repeatability, sensitivity...) and the quantification feasibility of these products and the application of chemometrics methods for spectral pretreatment. The thesis is organized around three publications. The first publication is a literature review, which aims to list but a few of the concrete applications of THz spectroscopy. The second one concerns the metrological characterization of the THz spectrometer we worked on. Lastly, the third one deals with the quantification aspect of THz spectroscopy and the possibility of using the same principles and chemometrics techniques that are used in infrared spectroscopy.

La précision des calculs théoriques d'électrodynamique quantique dans l'atome d'hydrogène est actuellement limitée par la constante de Rydberg et la distribution de charge du proton. La comparaison entre ces calculs et les mesures expérimentales de deux fréquences de transition dans l'hydrogène permet d'extraire ces deux constantes. La mesure de la transition 1S-2S est la plus précise à ce jour avec une incertitude relative de 10-15. L'objectif de mon travail de recherche est d'améliorer la précision de mesure de la fréquence de la transition 1S-3S, pouvant être utilisée comme la deuxième mesure nécessaire.La transition 1S-3S est sondée par une excitation à deux photons à 205 nm, permettant de s'affranchir de l'effet Doppler du 1er ordre. Ce faisceau UV est produit par somme de fréquence dans un cristal non linéaire. L'onde lumineuse délivrée par un laser Titane-saphir à 894 nm est sommée avec un faisceau à 266 nm produit par doublage d'un laser Nd-YO4. Cette somme de fréquence délivre un faisceau continu à 205 nm d'une puissance de 15 mWLa distribution de vitesse du jet atomique, dont la connaissance est indispensable pour évaluer l'effet Doppler du 2ème ordre, est déterminée grâce à l'effet Stark motionnel où l'action d'un champ magnétique produit un décalage en fréquence quadratique en vitesse.Les fréquences des deux lasers sources sont mesurées à l'aide un peigne de fréquence optique.La fréquence de la transition 1S-3S est finalement déterminée avec une incertitude relative de 10-12. Sa valeur conduit à une valeur préliminaire du rayon du proton qui serait en contradiction avec celle préconisée par le CODATA. / The uncertainty of the Quantum Electrodynamics calculations for hydrogen atom is currently limited by the knowledge of the Rydberg constant and the proton charge radius. Those two quantities can be extracted from the comparison between the theoretical predictions and two different frequency measurements on hydrogen.The 1S-2S transition frequency is one measured with the highest resolution with a relative uncertainty of 10-15. The aim of this thesis is to improve the determination of the 1S-3S transition, which can be used as the second precise measurement. The 1S-3S two-photon transition is excited at 205 nm. This UV light beam is generated by frequency mixing in a non-linear crystal. An 894 nm light delivered by a Ti:Sa laser is mixed with a 266 nm light beam generated by a quadrupled Nd:YVO4 laser. A reliable 15 mW continuous radiation at 205 nm is then produced. The frequencies of both lasers are measured simultaneously using an optical frequency comb referenced to a cesium clock. To evaluate the second-order Doppler effect, the velocity distribution of the atomic beam is determined thanks to a motional Stark effect. This effect is realized with a static magnetic field which induces a velocity-dependent quadratic frequency shift. Finally, the frequency of the 1S-3S transition is determined with a relative uncertainty of 10-12 which is accurate enough to contribute to the “proton size puzzle”. However, depending on the velocity distribution used in the analysis, the obtained value agrees or not with the present recommended CODATA value.

Ce mémoire présente la réalisation d'un dispositif expérimental de deuxième génération pour le projet FORCA-G (FORce de CAsimir et Gravitation à courte distance). L'objectif de ce projet est la mesure des interactions à faible distance entre un atome et une surface massive. La mesure de force est réalisée à l'aide d'interféromètres atomiques utilisant des atomes confinés dans un réseau optique 1D vertical basé sur le déplacement des atomes de puits en puits. La dégénérescence des niveaux d'énergies des atomes dans les puits du réseau est levée par la force que l'on cherche à mesurer. Des transitions Raman permettent de séparer les atomes dans des puits adjacents, puis de les recombiner, créant ainsi un interféromètre atomique qui permet de mesurer la différence d'énergie entre puits, liée à la fréquence de Bloch nu B du réseau. Ce travail présente la mise en place d'un dispositif proprement dédié au projet, qui permettra à terme de mesurer les forces à faible distance. Il rend compte des améliorations obtenues en configuration de gravimètre sur la sensibilité court terme de la mesure qui atteint 5. 10^-6 à 1 s. Il regroupe l'étude des limitations de la sensibilité, de l'exactitude et l'étude de la perte de contraste des interféromètres. Il présente aussi la mise en place d'une étape supplémentaire : l'implémentation d'un piège dipolaire visant obtenir un échantillon d'atomes plus dense et plus froid. / The thesis presents the set up of the second version of the experiment FORCA-G (CAsimir FORce and Gravitation at short range). The purpose of this experiment is the measurement of short-range interactions between an atom and a massive surface. The measurement is realised thanks to atom interferometers using atoms trapped in a 1D vertical optical lattice. The energy levels of atoms in such a trap are shifted from lattice site to another by the force we aim at measuring. We move the atoms from site to site using counter-propagating Raman transitions. The atoms are moved from Δm lattice sites only if the Raman frequency matches υHFS +∆m.υB where υHFS is the frequency of the hyperfine ground state transition and υB is the Bloch frequency and represents the difference of potential energy between two wells in the case where the atoms are far from the surface. This thesis presents the implementation of a setup properly dedicated to FORCA-G in which the measurement of short-range forces will be possible. It reaches an improved short-term relative sensitivity on the measurement of the Bloch frequency of at 3.9 10-6 at 1s. It contains the studies of the limits in the sensitivity, the accuracy and the contrast losses. It also presents the implementation of a dipolar trap to further cool the atoms and increase their density (crossed dipolar trap with a 1064 broadband laser).

A acuracidade dos processos de medição e de montagem que utilizam os desempenos como referência depende principalmente da planicidade dessas superfícies. Se a referência está fora das especificações é inútil utilizar instrumentos sofisticados e de alta acuracidade. Neste trabalho é apresentado um sistema para medição de desvios de planicidade de desempenos. O sistema é constituído por dois transdutores de deslocamento do tipo LVDT fixados no eixo z de uma máquina de medir a três coordenadas (MM3C). Durante a medição a intenção é avaliar apenas os desvios da superfície, no entanto os mancais das MM3C não se deslocam perfeitamente e as leituras obtidas são as componentes dos desvios da superfície e do movimento dos mancais. Para eliminar os componentes de erros da máquina dos dados medidos, pode-se usar as técnicas de separação de erros, e então, o desvio da superfície pode ser determinado. O sistema de medição proposto possui uma interface eletrônica que possibilita a aquisição dos sinais da régua óptica da MM3C e dos transdutores de deslocamento do tipo LVDT. Possui, também, um programa computacional que utiliza as técnicas de separação de erros para determinar o desvio de planicidade do mensurando. O sistema desenvolvido foi utilizado para medir o desvio de planicidade de uma superfície. Para verificar a eficiência do mesmo foi realizada uma comparação entre os valores de erro de planicidade obtidos através de medição com o sistema proposto e aqueles obtidos com interferômetro a laser e nível eletrônico. / The accuracy of measurement and assembly process using surface plates depends mainly on these surfaces flatness. If the surface plate does not meet the flatness specification, it is ineffective to apply high technology instruments of measurement. This research proposes is to develop a flatness measurement system for surface plates using two electronic comparators attached to the coordinate measuring machine (CMM). During a measurement process, the purpose is to evaluate the workpiece profile. However, the signals picked up by sensors include the workpiece profile and component motion error. In order to separate these errors, error separation techniques can be applied. The proposed measurement system has an electronic interface to collect data from the CMM optical scale and from the electronic comparators. The collected data are sent to a computer prepared with an algorithm for applying the error separation equations and for compute the flatness error. A surface was measured using the proposed measurement system. To evaluate its efficiency, the results were compared to the measurements made using electronic level and laser interferometric system.

L'objectif de cette thèse est de développer un système étalon de thoron afin de raccorder les instruments de mesure de l'activité volumique du thoron (²²⁰Rn) dans l'air. Le dispositif de mesure conçu est un volume composé d'un détecteur semi-conducteur silicium couplé à un champ électrique permettant la capture des descendants solides chargés du thoron à la surface du détecteur. Des simulations avec un code déterministe ont permis d'optimiser la forme et les dimensions du dispositif pour obtenir un champ électrique efficace pour la capture de ces descendants malgré un débit de circulation rapide à l'intérieur du volume de mesure. Des calculs Monte-Carlo ont aussi permis de définir le rendement de détection du dispositif, ainsi que la forme du spectre alpha. Les rendements de détection calculés pour la mesure du ²²²Rn et ²²⁰Rn ont été validés expérimentalement avec une atmosphère étalon de radon (²²²Rn) produite à l'aide d'un banc de dilution spécifiquement conçu. Une étude de sensibilité a montré que la réponse du dispositif est indépendante de la pression, de l'humidité et du débit de prélèvement dans la gamme de mesure considérée. L'analyse des spectres alpha obtenus a permis de qualifier précisément l'activité volumique d'une atmosphère en thoron avec une incertitude-type associée de 1%. Le dispositif, portable, a été transporté au laboratoire de métrologie de l'activité Italien pour réaliser des comparaisons avec leur dispositif en développement. Les résultats sont compatibles en tenant compte des incertitudes associées. / The goal of this PhD is to develop an activity standard for use in calibrating monitors used to measure the thoron (²²⁰Rn) concentration in air. The device, which has been designed to be the standard, is a volume with a silicon semi-conductor detector and an electric field which allows the charged decay products of thoron to be trapped on the detector surface. A finite element method has been used for the electric field simulations. This electric field is high enough to catch the decay products of thoron at the detector surface even with the high flow rate inside the volume. Monte-Carlo calculations were used to define the detection efficiency of the system and to optimise the geometry shape and size. The calculated detection efficiencies have been compared with the results obtained for a reference radon (²²²Rn) atmosphere produced with a new gas dilution setup. These experiments allowed the sensitivity of the system to be evaluated, as a function of the air properties. It has been demonstrated that the measurement system is independent of the pressure, the relative humidity and the flow rate for a large range of values. Through the analysis of measured alpha spectra the experimental gas detection efficiency was found to be consistent with the Monte-Carlo simulations. This portable system can now be used to evaluate precisely the thoron activity concentration with a well-defined associated uncertainty. Comparison measurements have been performed at the Italian metrological institute. Both systems are consistent within their uncertainties.

Cette thèse présente les travaux de recherche effectués dans le but de concevoir et optimiser un instrument de métrologie du front d’onde et d’imagerie de phase sur faisceau synchrotron dans la gamme des rayonnements X durs. L’étude s’est focalisée sur la conception d’un interféromètre à réseau unique dont l’extraction des informations associées à l’analyse du front d’onde s’effectue par démodulation Fourier. Ces choix ont été déterminés par la volonté de concevoir un instrument robuste pouvant fonctionner sur une large gamme de conditions expérimentales sans avoir à modifier et accorder les paramètres et éléments constitutifs fondamentaux de l’instrument à chaque expérience. Ceci se résume par la contrainte forte de pouvoir réaliser une calibration absolue du système de façon à garantir la prise de mesure ultérieure par acquisition d’un interférogramme unique tout en s’affranchissant des erreurs de mesure déterministes de l’instrument.La variable expérimentale la plus importante correspond à l’énergie de travail; par conséquent la conception de l’interféromètre s’est organisée autour de la recherche de performances constantes sur une large bande spectrale pour des énergies entre 10 à 30keV, et a conduit à l’étude et à la mise en œuvre d’une configuration interférométrique innovante. Celle-ci exploite un régime diffractif particulier du réseau permettant d’accéder à la propriété d’achromaticité (non rigoureuse) par repliement des performances de mesure sur cette bande spectrale, et ce, pour un instrument reposant pourtant sur un composant diffractif fondamentalement chromatique.D’autre part, afin de garantir l’analyse quantitative de l’information portée par les modulations interférométriques générées par le réseau, nous avons également optimisé les traitements numériques et abouti au développement d’un algorithme de pré-traitement des interférogrammes permettant de s’affranchir des effets de bord lors de l’analyse d’images à support fini. Les artefacts rencontrés sont connus sous le nom de phénomènes de Gibbs et apparaissent dans le cas général de la transformée de Fourier d’un signal discontinu. Ainsi, annuler ces effets de bord permet également de gérer les problèmes d’éclairement partiel de la pupille de l’analyseur dont la gestion est essentielle en métrologie de front d’onde.Enfin nous présenterons les résultats expérimentaux de validation du concept interférométrique et de mesures applicatives en métrologie optique et en imagerie de phase sur des échantillons d’intérêt issus de domaines variés, de la biologie à la paléontologie. / This PhD dissertation presents the optimization and design of a wavefront analyzer dedicated to optical metrology and phase imaging on synchrotron sources in the hard X-ray regime. We chose to develop a single grating interferometer combined with a phase retrieval algorithm based on Fourier analysis. The main purpose here is to conceive a bulk instrument able to work in a great variety of experimental conditions without having to tune the parameters of the instrument in between experiments. This is thus related to the key constraint that is to calibrate the wavefront analyzer so that any further measurements could be corrected from any deterministic errors and allow single shot measurements of any sample.The key varying parameter in synchrotron experiments is the radiation energy (or wavelength). Therefore, the design of the interferometer aimed at minimizing the discrepancies of its performances over a broad spectral range from 10 to 30keV . This research lead in one hand to the description and implementation of an innovative interferometric configuration based on the achromatization of the instrument performances over this spectral range, despite the chromatic nature of the grating.On the other hand, in order to guaranty the quantitative analysis of interferograms, we also optimized the numerical approach to extract and treat the information they contain. This lead to the development of a conditioning procedure for a subsequent phase retrieval by Fourier demodulation. It fulfills the classical boundary conditions imposed by Fourier transform techniques and allows a nearly artifact-free extraction of the information.At last, experimental results demonstrate first, the viability of the grating achromatization concept, and then, the possibility to realize the metrology of grazing incidence optics at different wavelengths. The instrument was then used for phase imaging purposes of biological and archaeological samples.

[pt] O controle das cores, seja visual ou instrumental, necessita de padrões de comparação. No caso visual esse padrão tem que ser uma amostra real, mas no caso instrumental existe a possibilidade de definir padrões virtuais em forma de valores de refletância espectral. Os padrões reais não são permanentes. A cor muda com o tempo dependendo do substrato, dos corantes ou pigmentos e dependendo da forma de armazenamento essa mudança se torna perceptível após algumas semanas, meses ou anos. Outra limitação dos padrões reais é a dificuldade de reproduzi-los em número suficiente com reprodutibilidade aceitável. Nas indústrias é muito comum usar coleções de cores padrões (por exemplo, Pantone, NCS, RAL, ColorCurve), mas mesmo assim sempre há uma diferença entre amostras até da mesma edição, e muito mais entre edições diferentes. Os padrões virtuais não têm essas desvantagens, mas para poder utilizá-los, tem-se que determinar o efeito da estrutura da amostra (não mostrada pelos valores de refletância) na cor percebida, e os limites de tolerância permissíveis na iluminação primária (luz do dia) e em outras iluminações (incandescente, fluorescente). Apresentam-se no texto conceitos relevantes e básicos sobre colorimetria limitados ao escopo abordado no trabalho. O objetivo do trabalho é verificar quantitativamente a diferença instrumental de cor de tecidos com diferentes estruturas, a eficiência de diferentes métodos de cálculo computadorizado de correção de receitas de tingimento e a influência do fenômeno da metameria na tolerância colorimétrica. As amostras têxteis com diferentes estruturas serão tingidas em diversas cores e medidas para a determinação da relação da estrutura com a tolerância instrumental, para vários padrões virtuais e vários iluminantes. Demandou-se a preparação de amostras padrão e medições espectrofotométricas, com elevado controle metrológico, realizadas nas dependências do SENAI-CETIQT, mais especificamente nas UOEQ (Unidade Operacional de Ensaios Químicos) e UOC (Unidade Operacional de Colorimetria). Para a análise dos resultados utilizou-se de métodos estatísticos e matemáticos, auxiliada por planilhas eletrônicas. Complementando a formação acadêmica, o desenvolvimento da pesquisa realizou-se nas instalações do SENAI/CETIQT, que possui comprovada competência técnica e científica na área colorimétrica e uma adequada infra-estrutura laboratorial em metrologia da cor que deu suporte ao trabalho. A presente pesquisa de mestrado voltada ao equacionamento de problemas de interesse industrial desenvolveu-se no contexto de um convênio celebrado entre o Programa de Pós-Graduação em Metrologia da PUC-Rio e o SENAI/CETIQT, que somam esforços para consolidar no País uma nova área de pesquisa em metrologia da cor. Por essa razão, o trabalho foi incluído dentre os 10 projetos-piloto que participaram do Convênio FINEP/MCT no. 22.01.0692.00, Referência 1974/01, que aportou recursos do Fundo Setorial Verde Amarelo para direcionar o esforço de pesquisa em metrologia para a solução de problemas de interesse industrial. Foi com esse propósito que padrões virtuais e tolerâncias colorimétricas no controle instrumental das cores foi selecionado como tema central da pesquisa, cujo desenvolvimento beneficiou-se do ambiente acadêmico e de pesquisa da universidade e da excelente infra-estrutura laboratorial em colorimetria do SENAI/CETIQT. A cooperação inter-institucional estabelecida evidencia o potencial da parceria na solução de complexos problemas de interesse industrial. / [en] The control of colors, either visual or instrumental, needs comparison standards. In the visual case this standard has to be a real, physical sample, but in the instrumental case, it is possible to define virtual standards in the form of spectral reflectance values. The real standards are not permanent. Their color changes over time depending on the substrate, the dyes or pigments, and - depending on the storage form - this change becomes perceptible after some weeks, months or years. Another limitation of the real standards is the difficulty to reproduce them in sufficient number with acceptable reproducibility. In industry it is very common to use collections of colors standards (for example, Pantone, NCS, RAL, Color-Curve), but even then there is always a difference between samples even of the same edition, and much more between different editions. Virtual standards do not have these disadvantages, but to be able to use them, the effect of the structure of the sample on the perceived color (not shown by the values of spectral reflectance), as well as the acceptable limits of tolerance under the primary illumination (daylight) and under other illuminations (incandescent, fluorescent) have to be determined. Relevant and basic concepts of colorimetry are presented in the text, limited to the scope of the work. The objective of the work is to quantitatively verify the instrumental color difference between fabrics with different structures, the efficiency of different methods of computerized dye recipe correction calculation and the influence of the phenomenon of metamerism on colorimetric tolerances. Textile samples with different structures will be dyed in diverse colors and the effect of the structure on the instrumental tolerance for some virtual standards and some illuminants will be determined by instrumental measurements. Sample preparation and spectrophotometric measurements were performed under careful metrological control, carried out in the installations of SENAI-CETIQT, more specifically in the UOEQ (Chemical Testing Laboratory) and UOC (Applied Colorimetry Laboratory). Statistical and mathematical methods, assisted by electronic spread sheets, were used for the analysis of the results. Complementing the academic development, the experimental research took place at SENAI/CETIQT, an institution with proven scientific and technical capabilities in the field of colorimetry and outstanding laboratorial infrastructure in color metrology, that gave support to the work. The present M.Sc. Dissertation aiming at the solution of problems of industrial interest was developed in connection with an agreement celebrated between the Post-graduate Programme in Metrology of PUC-Rio and SENAI/CETIQT which cooperate to implement in the country a new area of research in metrology applied to colorimetry. For this reason, the research work was included among the 10 pilot-projects which benefited from the financial support received from the FINEP/MCT Agreement no. 22.01.0692.00, Reference 1974/01 (Green-Yellow governmental Fund) driving the research effort in metrology to the solution of industrial problems. This is why virtual standards and colorimetric tolerances in the instrumental control of the colors was selected as the central topic of this research, which successful development strongly benefited from the academic and scientific environment provided by the university and also by the outstanding laboratory infrastructure in colorimetry offered by SENAI/CETIQT. The agreement established proved to speed up the solution of complex problems requiring advanced knowledge of metrology.

[pt] Recentemente, a umidade vem sendo um dos parâmetros cujas necessidades de medição e controle mais se desenvolvem especialmente em processos os farmacêuticos, de biotecnologia, automobilísticos, além de instalações de ar condicionado. A necessidade de expansão da rede de laboratórios de calibração de umidade, rastreados à Rede Brasileira de Calibração, torna-se cada vez maior para atender a demanda da indústria, além de contribuir para a melhoria dos níveis de incerteza de medição. Este trabalho tem como objetivo uma análise crítica das diferentes metodologias de medição de umidade para uso em laboratórios de calibração. Foram avaliadas as mais utilizadas expressões para a pressão parcial de vapor, visando o cálculo de umidade, e foi realizada uma comparação com a formulação de Sonntag, a mais aceita na comunidade metrológica internacional. Foi também analisada uma comparação entre alguns dos softwares disponíveis na internet para o cálculo dos principais parâmetros de umidade, tomando como referência a mesma formulação. Dentre os fatores de possível influência nas medições de umidade, foi avaliado o comportamento da mistura ar seco-vapor d´água como gás ideal e real, bem com as alterações na composição do ar. Foi feita uma análise das principais fontes de incerteza de medição seguindo a metodologia de medição de umidade a partir da temperatura de ponto de orvalho. Finalmente, o objetivo deste trabalho é apresentar aos laboratórios de calibração de umidade uma análise crítica das metodologias disponíveis para medição de umidade, juntamente com as incertezas alcançadas. / [en] Recently, the need of humidity measurement and control has been shown to be of growing interest, especially in processes like pharmaceuticals, biotechnology, automotives, as well as in air conditioning installations. Increasing the number of laboratories, which are traceable to the Brazilian Calibration Network, is a requirement for attending the industry demands, besides contributing towards improving the measurement uncertainty level. This work has as an objective to critically analyze different methods for humidity measurement, to be used by calibration laboratories. The mostly used expressions for calculating vapor partial pressure were evaluated, aiming humidity determination in comparison with Sonntag´s one, the most accepted formulation by the international metrological community. Also, a comparison was made between different softwares, which were available through internet, for calculating the most important humidity parameters, in comparison with results obtained by the same formulation. Among several factors, the dry air and water vapor mixture properties were calculated, considering it as an ideal or a real one, as well as the effect air composition changes on humidity measurement. The contribution of several sources of errors on the combined uncertainty of humidity measurement was examined, following the methodology that uses the dew point as a starting measurement parameter. Finally, the objective of this work is to present the humidity calibration laboratories a critical analysis of the available methodologies for measuring humidity, together with the achieved uncertainties.