The dynamics of rigid body motion are dependent on the inertial properties of the body - that is, the mass and moment of inertia. For complex systems, it may be necessary to derive these results empirically. Such is the case for manual wheelchairs, which can be modeled as a rigid body frame connected to four wheels. While 3D modeling software is capable of estimating inertial parameters, modeling inaccuracies and ill-defined material properties may introduce significant errors in this estimation technique and necessitate experimental measurements. To that end, this thesis discusses the design of a device called the iMachine that empirically determines the mass, location of the center of mass, and moment of inertia about the vertical (yaw) axis passing through the center of mass of the wheelchair. The iMachine is a spring-loaded rotating platform that freely oscillates about an axis passing through its center due to an initial angular velocity. The mass and location of the center of mass can be determined using a static analysis of a triangular configuration of load cells. An optical encoder records the dynamic angular displacement of the platform, and the natural frequency of free vibration is calculated using several techniques. Finally, the moment of inertia is determined from the natural frequency of the system. In this thesis, test results are presented for the calibration of the load cells and spring rate. In addition, objects with known mass properties were tested and comparisons are made between the analytical and empirical inertia results. In general, the mass measurement of the test object had greater than 99% accuracy. The average relative error for the x and y-coordinates of the center of mass was 0.891% and 1.99%, respectively. For the moment of inertia, a relationship was established between relative error and the ratio of the test object inertia to the inertia of the system. The results suggest that 95% accuracy can be achieved if the test object accounts for at least 25% of the total inertia of the system. Finally, the moment of inertia of a manual wheelchair is determined using the device (I = 1.213 kg-m²), and conclusions are made regarding the reliability and validity of results. The results of this project will feed into energy calculations for the Anatomical Model Propulsion System (AMPS), a wheelchair-propelling robot used to measure the mechanical efficiency of manual wheelchairs.

The relationship between crack path and test method was examined by comparing the performance of adhesive-adherend combinations (six) in quasi-static fracture, mixed-mode fatigue, and creep crack growth. Crack paths in creep and quasi-static fracture were similar due to similar crack-tip plastic zone sizes in the epoxy adhesive even though the crack growth rates in creep were much smaller. Under condensed moisture and mixed-mode, creep and threshold fatigue tests produced interfacial failure. Under room-temperature dry environment, near threshold mixed-mode fatigue was interfacial, but was not in creep or quasi-static fracture. Smaller plastic zone size and crack path proximity to the interface that followed increased the sensitivity of near threshold, mixed-mode fatigue to surface properties. Therefore, the interfacial or cohesive failure of an adhesive system, which may judge the quality of the bond, can be a function of the test being conducted and may not be an absolute indicator of joint quality.

The relationship between crack path and test method was examined by comparing the performance of adhesive-adherend combinations (six) in quasi-static fracture, mixed-mode fatigue, and creep crack growth. Crack paths in creep and quasi-static fracture were similar due to similar crack-tip plastic zone sizes in the epoxy adhesive even though the crack growth rates in creep were much smaller. Under condensed moisture and mixed-mode, creep and threshold fatigue tests produced interfacial failure. Under room-temperature dry environment, near threshold mixed-mode fatigue was interfacial, but was not in creep or quasi-static fracture. Smaller plastic zone size and crack path proximity to the interface that followed increased the sensitivity of near threshold, mixed-mode fatigue to surface properties. Therefore, the interfacial or cohesive failure of an adhesive system, which may judge the quality of the bond, can be a function of the test being conducted and may not be an absolute indicator of joint quality.

Denna studie handlar om textila materials brännbarhet och antändlighet, framförallt bäddtextilier i bomull. Idag finns det lagkrav på att utföra antändlighetstester enligt SS- EN ISO 12952-1:2010 på bland annat fyllda bäddtextilier, till exempel täcke och kudde. Ikea of Sweden (IoS) har beslutat om att göra ytterligare antändlighetstester på ofyllda bäddtextilier så som underlakan, påslakan och örngott i säkerhetssyfte. Då det inte finns en testmetod för ofyllda påslakan i standarden har metoden för underlakan använts för test av påslakan. I detta arbete undersöks testmetodens relevans för test av påslakan. Det primära problemet med detta utförandet är att tester enligt ISO 12952-1 på ofyllda produkter resulterat i en stor spridning i utfallen mellan replikaten. Studien syftar därför till att undersöka om det finns faktorer som kan påverka denna inkonsekvens. För att göra denna undersökning har IoS egna testrapporter analyserats och testresultat sammanställts från år 2017. Ett urval av de produkter som visade spridda resultat valdes ut användes som testmaterial i egna utförda antändlighetstester enligt ISO 12952-1 på RISE, Research Institute of Sweden, i Borås. Litteratursökning, datainsamling och tester har legat till grund för denna studie. Projektet i sin helhet består av två delar; dels undersöks vilka faktorer som kan påverka utfallet och dels utreds frågan om testmetoden för lakan är relevant för test av påslakan. En hypotes om att val av tvättmedel kunde påverka testresultatet undersöktes närmare genom att fyra replikat tvättades med ett fosfatinnehållande tvättmedel före test. Tre test av fyra utförda tester visade godkänt resultat (75%), vilket ledde till slutsatsen om att tvättmedel är en känslig parameter och därför bör ej tvättmedlet avvika från standardens rekommendationer. Resterande tester, som tvättats enligt standardens rekommendationer, visade 100 % godkänt resultat, oavsett om lakan, påslakan (dubbelt lager) och påslakan (enkelt lager) testades. Utifrån testresultatet på de resterande testerna drogs en slutsats om att testmetoden för lakan är relevant för test av påslakan, oberoende om dubbelt eller enkelt lager testas. Dock ifrågasätts testets relevans ur ett miljöperspektiv, samt om tillräckligt många tester har utförts för att säkerställa trovärdigheten. / This study regards the combustibility and ignitability of textile materials, especially cotton bedlinens. Today, there is legal requirements for performing flammability tests according to SS-EN ISO 12952-1: 2010 on filled bedlinen, such as blankets and pillows. Ikea of Sweden (IoS) decided to make additional ignitability tests on unfilled bedlinen such as sheets, duvet covers and pillow cases for safety purposes. Since there is no test method for unfilled bed linens in the standard, the method for sheets was used for test of the duvet cover. In this report, the test method's relevance for the test of the duvet covers is examined. The primary problem with this is that tests according to ISO 12952-1 of unfilled products results in a spread in the output between replicates. Therefore the study aims to investigate if there is any factors that may cause this inconsistency. IoS test reports from yeas 2017 was analysed and complied. A few of the products that showed dispersed results were selected as test material for ignitability tests according to SS-EN ISO 12952-1: 2010 made at RISE, Research Institute of Sweden, in Borås. Literature research, data and test results have been the solid ground for this study.   The project consists of two main parts, one aims at investigating whether there are parameters that can affect the outcome of test results and the other analyses whether the test method for sheets is relevant for test of duvet covers. A hypothesis that choice of detergent had an effect on the result was further investigated by washing four replicates with a phosphate containing detergent before testing. Three tests of four showed approved results (75%), which led to the conclusion that detergent can be a sensitive parameter, thus the detergent should not deviate from the standard recommendations. The remaining tests, which were washed according to standard recommendations, showed 100% approved results, regardless of which test product. whether the sheet, duvet cover (double layer) and duvet cover (single layer) were tested. Based on the results of this study, the conclusion is given that the test method for the sheet is relevant for the test of the cover, regardless of double or single layer. However, the relevance of the test is questioned from an environmental perspective, and if enough tests have been conducted to strengthen credibility.

Tillsammans med företaget Houdini Sportswear har det här projektets ämne och avgränsningar tagits fram. Houdini är ett svenskt klädföretag som designar friluftskläder av hög kvalitet och med stort fokus på funktion och komfort. De letar ständigt efter nya lösningar för utveckling kring smartare materialval, förbättrad funktion och komfort och mindre materialåtgång. I det här projektet vill Houdini undersöka hur man kan utveckla förhållandet mellan värmeisolering och vikt i flatstickade textila strukturer. Genom att öka textilens värmeisolerande förmåga med fokus på minskad vikt kan både materialåtgång och ökad funktion och komfort nås. En av de viktigaste faktorerna som bidrar till hög värmeisolerande förmåga är mängden instängd stilla luft i den textila strukturen. En tät ytstruktur och mycket volym med många luftfickor är därför intressant för att behålla och immobilisera luften i varan. Syftet med studien är att undersöka hur justeringar av parametrar i en distanstextil, stickad på flatstickmaskin påverkar den värmeisolerande förmågan. Tillsammans med Houdini har en viktbegränsning satts på 400 g/m2 och genom hela studien används endast en typ av ullgarn med garngrovlek Nm 48/2 för att skapa en vara som är i monomaterial och som relativt lätt kan brytas ner.   För att testa den värmeisolerande förmågan hos textilier idag använder Houdini och flera andra företag standardiserade testmetoder. På forskningsinstitutet Swerea IVF finns möjligheten att testa sina textilier med vald testmetod. Dessa metoder är dock otillgängliga och kräver avancerad utrustning som kan försvåra processen och ta lång tid. Studiens syfte är därför även att ta fram och jämföra en egen, mer lättillgänglig, framtagen testmetod för värmeisolering med två standardiserade testmetoder, SS-­ISO 5085-­1:2004 och ISO 11092:2014, för att se om och hur resultaten varierar mellan dem.   En förstudie genomförs där olika bindningar och två egna testmetoder för värmeisolering tas fram. En av testmetoderna och en distanstextil utvecklas vidare i huvudstudien. Fokus ligger på variationer av distanstrådens masklängd och placering. Samtliga prover från huvudstudien testas på den egna testmetoden och vissa utvalda prover testas på de två nämnda standardiserade testmetoderna på Swerea IVF. Sedan görs en jämförelse mellan resultaten.   Resultatet visar att det inte finns något tydligt samband mellan ökad masklängd och ökad värmeisoleringsförmåga. Däremot kan vissa slutsatser dras om att ökade intervaller av distanstrådens placering kan öka den värmeisolerande förmågan. Detta ökar dock även vikten. Resultat från mätningar med tre olika testmetoder visar att rangordningen av provers värmeisolerande förmåga skiljer sig mellan de olika metoderna. Den egna testmetoden bör därför vidareutvecklas men anses ändå ha potential för att jämföra provers värmeisolerande förmåga. / This study is a collaboration with the sportswear company Houdini Sportswear. They are looking for new ways to improve function and comfort in their garments with less material waste and choosing more sustainable materials. The aim of this study is to investigate the relationship between heat and weight in flat bed knitted spacer structures and what parameters that affects it. By creating a material with high heat insulating properties and with focus on reducing the weight it is possible to reduce the material waste and to achieve an improved function. One of the most important factors contributing to high insulation value is the amount of entrapped still air within the textile structure. This project therefore investigates how the stitch length and the placement of the spacer thread can affect the amount of air within the structure and thereby the thermal insulation. Together with Houdini a weight limit has been set to 400 g/m2 and only one type of wool yarn is used throughout the project in order to make a mono-­material fabric that can more easily degrade.   When measuring the thermal insulation properties of a fabric standardized methods are often used. These might have advanced equipment and be hard to access.The second aim of this project is therefore to develop a more accessible test method for measurement of thermal insulation in textiles and also to compare the results from the own method with the results from two standardized methods.   A pre study is made where different structures and two own developed test methods are being investigated. One of the test methods and a spacer binding is further developed in the main study where focus lies on reproducibility of the tests, and variations of the spacer thread stitch length and placement. All of the samples from the main study is tested on the own developed test method and some chosen samples are tested on two standardized test methods: ISO 5085-­1:2004 and ISO 11092:2014 at the Swedish research institute Swerea IVF. A comparison is made between the results from the three methods.   The results show that tests with our own developed test method has not been able to show a clear relationship between heat insulation and the changes of parameters in different textiles. Although, the samples that have been tested indicate that there might be a relationship between high heat insulating properties when the spacer thread is knitting more frequently. When comparing the results from the three different test methods the own developed method show different raking from the two standardized methods. The own developed test method should therefore be further developed. But still, after discussion with Valter Dejke, researcher at Swerea IVF, and Sibel Okcabol, quality manager at Swerea IVF, it is believed that the own test method could be used for comparing thermal insulating properties between textiles.

Esse trabalho tem como objetivo estabelecer um método simples de ensaio que permita determinar a resistência à tração e módulo elástico de partículas individuais de agregados graúdos naturais submetidas à carga de compressão. O método de ensaio foi desenvolvido baseado no ensaio de carga pontual (Point Load Test) que permite obter a resistência à tração de partículas. Para a determinação do módulo de elasticidade foi acoplado ao método um LVDT e foi satisfeita a condição para aplicação da teoria de contato de Hertz (contato curvo-plano entre as partículas e as fixações de aplicação da carga). Inicialmente a metodologia foi avaliada utilizando como material de referência partículas de vidro (com geometrias similares aos agregados). Após a validação no vidro, o método de ensaio foi aplicado em agregados graúdos de granito. Propôs-se um método de seleção de partículas com base na sua distribuição de frequência de absorção, com o intuito de reduzir a quantidade de partículas testadas mecanicamente necessárias para obter a distribuição de Weibull da resistência à tração (e módulo de elasticidade). Para tanto, foi feita a determinação da absorção de água individual de centenas de partículas selecionadas por amostragem a esmo da população de agregados. Com base nos resultados, foi possível determinar a distribuição de Weibull da resistência à tração e módulo de elasticidade das partículas. Observou-se que as resistências à tração variaram de 3 a 15 MPa. Os agregados possuíam aproximadamente 10% da população de partículas com resistência à tração inferior a 5 MPa, o que pode influenciar as classes de resistência à compressão de concretos >50MPa. Os módulos elásticos dos agregados também foram variáveis (18-67 GPa) e aproximadamente 10% da população de partículas com módulo elástico <30 GPa. Isso pode limitar o módulo de elasticidade do concreto (que geralmente é em torno de 27 GPa), dependendo do processo de escolha dos agregados. As funções exponenciais inversas fundamentais entre essas propriedades mecânicas e a porosidade foram confirmadas (para valores médios). / The aim of this work is to establish a simple test method to determine the tensile strength and elastic modulus of individual natural aggregate particles subjected to a compression load. The test method was developed based on the Point Load Test, that allows to obtain the tensile strength of particles. For the determination of the elastic modulus, an LVDT was coupled to the method and the condition for application of Hertz contact theory (curved-plane contact between the particles and the load application fixations) was satisfied. Initially, the method was evaluated using glass particles as reference (with geometries similar to aggregates). After validation on the glass, the test method was applied to coarse granite aggregates. A method of particle selection was proposed based on its absorption frequency distribution, in order to reduce the amount of mechanically tested particles required to obtain the Weibull distribution of tensile strength (and elastic modulus). For this purpose, the determination of the individual water absorption of hundreds of particles selected by random sampling of the population of aggregates was made. Based on the results, it was possible to determine the Weibull distribution of the tensile strength and elastic modulus of the particles. The tensile strengths ranged from 3 to 15 MPa. The aggregates had approximately 10% of the particles population with tensile strength less than 5 MPa, which may influence the classes of concrete with compressive strength > 50MPa. The elastic modulus of the aggregates was also variable (18-67 GPa), with approximately 10% of the particle population with elastic modulus <30 GPa. This may limit the elastic modulus of the concrete (usually around 27 GPa) depending on how the aggregates are selected. The fundamental inverse exponential functions between these mechanical properties and the porosity were confirmed (for mean values).

Due to an increase in the number of large-scale façade fires around the world the interest in the fire hazards of façades has also increased. The hazards of such fires have been acknowledged since earlier and many different test methods have therefore been developed to enable evaluation of the fire properties of wall assemblies. The purpose of this study is to map differences and similarities between existing full-scale test methods that are currently used to evaluate façade systems based on their performance when influenced by fire. The study also includes a review of previous research into parameters and conditions that influence a façade fire. Some past incidents will be used to enable comparison between the tests and a real fire scenario. The review of previous research has been focused in the areas fire spread, fire properties influencing the heat flux from a fire, and comparative studies evaluating differences between existing test methods. For the study of parameters in existing test methods 21 test methods have been identified and included in the study. A detailed compilation of information on each of the included test methods can be found in the tables included in appendix A. The mapping of the differences and similarities of the included methods has resulted in comparisons of wall and specimen specifications, ignition source parameters, measuring points and approval criteria. The increase in interest for façade fire hazards has also led to some new methods being developed and old methods being revised. New methods and unfinished revisions have not been included in the comparison study but are mentioned.  The conclusion of this study is that although the variation between the tests on the detailed methodology level are very high, if you consider the conditions used for approval and evaluation, they can all be linked to identified hazards of façade fires. However, there are some parameters that need to be re-examined. Previous research indicates that the fire load may be the parameter with most influence on the fire scenario. This is one of the parameters that vary the most and at same time it is likely to contribute to different results between the test methods. Another parameter is the influence of wall openings representing windows to overlaying compartments. This is a parameter that could have a significant effect on the outcome of a test.

Large-scale laminated safety glass is increasingly used in the public and private sector for example museums, jewelers or villas. Special safety is provided by these glasses in layered combination with polycarbonate sheets. Due to the low weight and its high tensile elongation comparing glass of the same thickness, highly transparent polycarbonate is suitable for safety-relevant components. Because of the dimensional limitation in the extrusion process, joining polycarbonate compounds are necessary for large-area joint of this laminated safety glass. According to the present state of the art, a transparent weld joint of polycarbonate is not possible. Within the scope of the project, polycarbonate joints were produced by infrared welding and adhesive bonding. The transparency of the joints was realized by varying the process parameters, for example the heating time. Additionally the present study focusses on a new innovative light intensity measurement, because conventional optical test methods can only subjectively assess the transparency and quality of the compound and the welding seam. The principle is based on the transmission of the welded and polished polycarbonate sample by means of a laser and the measurement of the current at the connected photoelectric cell, which is proportional to the intensity of the laserlight. This enables a qualitative evaluation of the joining dimension and -quality such as structural changes over the entire sample width. The results of this objective method correlate with the width of the stress distribution around the joining level, which are made visible with the photoelasticity using a polariscope and thus contribute to a better process understanding.

The work aims to study different methods suitable for de-icing and ice prevention of vehicle headlamps and tail lamps, especially LED-lights. Furthermore, the work aims to investigate the scale of the problem with insufficient or lack of de-icing on automotive lamps depending on the region and the environment the vehicle operates in. The problem with insufficient de-icing in automotive lamps was investigated by observations, tests of various lamps and a driver survey. Deicing methods were identified through a litera-ture review. The methods were studied in detail, and some were also evaluated by tests. The tests were narrowed down to temperature measurements and de-icing measurements. The latter were performed using a test method especially developed for the task. The collected data was used to evaluate whether actions are needed to be taken and to form recommen-dations for future developments. The number one priority should be to improve the tail lamps de-icing ability. Headlamps can also be improved but there is no imminent need. Insufficient de-icing of headlamps and tail lamps can potentially be a problem in all areas subjected to cold winter climate. Tail lamps should be fitted with electrical heating in order to improve the de-icing ability. They should be positioned so that snow and ice does not stack on top of them. The de-icing time of tail lamps should be less than 10 min. Truck drivers needs to be better in scraping their headlamps and tail lamps. The time needed for de-icing Scania’s H7 headlamps is 20 min for halogen version and 35 min for xenon version at -18 °C. This should be compared to the BMW LED-headlamp which needs 65 min to complete de-icing. LED-headlamps are probably limited to a de-icing time of approximately 60 min unless additional heat is added to the headlamp lens. The fastest and most efficient way to de-ice the headlamps is to use hot washer fluid. Electrically heated lenses are also effective but the de-icing process is slower. The proposed test method is a simple and effective way to compare and evaluate headlamps and tail lamps without knowing internal airflows and light sources. The way the ice layer is created on the device under test is unique to this method. The created ice layer is extremely uniform and the results are easy to evaluate. / Sammanfattning Arbetets syfte är att undersöka olika avisningsmetoder och isförebyggande åtgärder för hu-vudstrålkastare och baklyktor, speciellt LED-lampor. Ytterligare syfte är att undersöka hur stort problemet med otillräcklig avisning är beroende på vilken region och miljö som fordonen körs i. Problemet med otillräcklig avisning av fordonsbelysning har undersökts genom observatio-ner, en förarenkät samt genom tester av olika strålkastare och baklyktor. Avisningsmetoder har identifierats genom en litteraturstudie. Metoderna har studerats i detalj och några har även utvärderats genom tester. Testerna har utförts enligt en för uppgiften framtagen testme-tod. De insamlade uppgifterna har sedan används för att utvärdera om åtgärder behöver vid-tas och för att ge rekommendationer för framtida utvecklingsprojekt. Första prioritet bör vara att förbättra baklyktornas avisningsförmåga. Huvudstrålkastarna kan även de förbättras men det föreligger inte i dagsläget något akut behov av det. Otillräcklig avisning av huvudstrålkastare och bakljus kan potentiellt vara ett problem i alla miljöer med ett kallt vinterklimat. Baklyktorna bör utrustas med en eluppvärmd lins för att förbättra avis-ningen. Lamporna bör placeras på ett sådant sätt att snö och is inte kan ansamlas ovanpå dem. Avisningstiden för en baklykta bör inte överstiga 10 min. Lastbilschaufförerna behöver bli bättre på skrapa av is och snö från huvudstrålkastare och baklyktor. Avisningstiden för Scanias H7 huvudstrålkastare är 20 min för halogen versionen och 35 min för xenon versionen vid -18 °C. Detta kan jämföras med att BMW:s LED-huvudstrålkastare behöver 65 min för att avisas. LED-huvudstrålkastare är troligtvis begränsade till en avis-ningstid kring 60 min om inte någon extra värme tillförs. Den snabbaste metoden för att avisa en huvudstrålkastare är att använda varm spolarvätska men elektrotermisk avisning kan också vara mycket effektivt. Den föreslagna testmetoden är ett enkelt sätt att jämföra och utvärdera olika huvudstrålkastare och baklyktor. Det som är unikt med testmetoden är hur isskiktet på lamporna bildas. Isskiktet som skapas är extremt jämt och lätt att utvärdera.

Cette thèse fait partie des programmes de recherche sur le vieillissement des ouvrages de génie civil nucléaire lancés par l’IRSN (Institut de Radioprotection et de Sûreté nucléaire) dans le cadre du projet de prolongement de la durée d’exploitation des centrales nucléaires françaises. Le but étant de contribuer à la connaissance des mécanismes de vieillissement des matériaux constitutifs des ouvrages non-remplaçables d’une centrale nucléaire, tels que les enceintes de confinement. En effet, leur comportement représente un point important pour l’évaluation de la sûreté des installations nucléaires, car elles représentent la troisième et la dernière barrière de confinement contre le relâchement éventuel des produits radioactifs dans l'environnement.Parmi les phénomènes intervenants lors du vieillissement des ouvrages, le développement éventuel des pathologies, notamment les réactions de gonflements internes (RGI) qui sont susceptibles de dégrader le béton et conduire à l’apparition des fissurations dans les structures touchées. Ces réactions de gonflement internes (RGI) englobent la réaction sulfatique interne (RSI), la réaction alcali-granulat (RAG) et la concomitance de ces deux. Il s’agit de réactions endogènes, se produisant suite à l’interaction des composants initiaux du matériau. Ces interactions peuvent avoir lieu suite à l’utilisation de granulats réactifs en cas de RAG et/ou suite à un échauffement important au jeune âge en cas de RSI. Dans les ouvrages nucléaires, l’éventualité de ces phénomènes ne peut pas être écartée. En effet, ils comportent des éléments structuraux massifs en béton armé pour lesquels un échauffement important a pu se produire lors de la prise du béton et des granulats réactifs ont pu être utilisés pendant la confection. Ce risque doit donc être étudié d’autant plus que le prolongement de la durée d’exploitation des installations est envisagé.Cependant, la cinétique d’apparition des désordres suite au développement de ces réactions dans les ouvrages est en général lente allant jusqu’à quelques dizaines d’années, d’où le recourt à des essais accélérés. Cependant les solutions pratiquées dans les laboratoires de recherche, consiste à utiliser des protocoles d’essais accélérés mis au point sur des éprouvettes de faibles dimensions.Le travail de recherche réalisé dans le cadre de cette thèse comporte trois grandes axes : la mise au point d’un protocole expérimental visant à accélérer les réactions de gonflement interne du béton à l’échelle de structures massives, l’analyse multi-échelle des phénomènes et la mise au point d’une instrumentation innovante adaptée à cette problèmatique. Trois maquettes (de dimensions 2,4 x 1,4 x 1 m3) représentatives d’un ouvrage massif ont été réalisées dans des conditions maitrisées et optimisées afin de favoriser respectivement le développement de la réaction sulfatique interne (RSI), de la réaction alcali-granulat (RAG) et de la concomitance des deux pathologies, RAG et RSI. Ceci a permis d’étudier ces pathologies et de les caractériser à une telle échelle.L’effet d’échelle sur la cinétique de gonflement est mis en évidence grâce à une étude expérimentale multi-échelle permettant de comparer l’évolution des réactions dans les maquettes avec leur évolution dans des éprouvettes en béton reconstitué et des carottages.Une méthode d’instrumentation innovante par fibres optiques a été spécifiquement développée pour ces maquettes, permettant de suivre l’évolution des pathologies dans la masse et d’en déduire les champs de déformation dans plusieurs plans d’intérêt. Les mesures sont utilisées pour décrire finement les mécanismes évolutifs du gonflement et la dégradation des structures. Ces mesures sont aussi utilisées comme référence pour des simulations numériques ultérieures. / This thesis is part of a research program on the aging of nuclear infrastructures launched by the IRSN (Institute for Radiological Protection and Nuclear Safety) within the project aiming to extend the lifespan of the French nuclear power plants. The goal is to contribute to the knowledge of aging mechanisms touching the constituents of non-replaceable structures of a nuclear power plant, such as the containment building. Matter of fact, the behavior of such structures is an important point for the safety assessment of nuclear installations because they represent the third and last barrier against the dispersion of radioactive particles in the environment. Among the phenomena involved in aging mechanisms, internal swellings reactions (ISR) are pathologies that can degrade concrete by causing swelling, cracking and major disorders in the affected structures. Internal swelling reactions (ISW) include the delayed ettringite formation (DEF), the alkali-aggregate reaction (AAR) and the concomitance of the two. These pathologies are endogenous reactions, occurring as a result of the interaction between the initial components of the material. These interactions take place following the use of reactive aggregates in case of AAR and/or following a significant warm-up at the concrete’s early age in case of DEF. In nuclear facilities, the possibility of these phenomena cannot be ruled out. Matter of fact, they include massive reinforced concrete elements for which a significant heating could have occurred at the early age. Moreover, reactive aggregates may have been used during the construction phase. Therefore those risks ought to be studied especially if the extension of the lifespan of these installations is envisaged.However, the kinetics of these reactions are generally slow and it can take for up to a few decades for disorders to appear in the structure. Among the solutions used in research laboratories, one is to use accelerated test methods developed on small specimens.The research work carried out in this context includes three major axis: the development of an experimental protocol to accelerate internal swelling reactions of concrete on a massive structure scale, multi-scale analysis of the phenomena involved and the development of innovative instrumentation adapted to this problem. The purpose is to allow better observation and understanding of swelling reactions at such scale. Hence, three representative massive concrete mock-ups whose dimensions are 2,4 x 1,4 x 1 m3 were realized under controlled and optimized conditions for the development of delayed ettringite formation (DEF), alkali silica reaction (ASR) and the concomitant ASR-DEF case respectively.The scale effect on swelling kinetics is demonstrated by an experimental study comparing the evolution of swelling reaction in the massive samples with the evolution in reconstituted concrete laboratory specimens as well as coring specimens.An innovative instrumentation method using optical fibers is also developed thus allowing the monitoring of the pathologies in the massive structures and the tracking of the respective deformation fields in several two-dimensional plans of interest. The measurements are to finely describe and help in understanding the evolutionary mechanisms of structural degradation and can be used subsequently as a reference for numerical simulations.