Wood beam-column connections have traditionally been designed as simple shear connections, ignoring their potential moment capacity. A major reason for not utilizing such moment connections is linked to the brittle limit states that wood components exhibit. The purpose of this research was to develop and test a ductile and high-strength wood moment frame connection. A design procedure for such a connection is presented herein. The proposed glulam beam-column connection utilizes an embedded steel knife plate with a reduced section that acts as a ductile yield link, thus limiting the moment that can be transferred through the connection. This configuration is intended to fail through yielding of the ductile link, thus preventing non-ductile failure mechanisms of wood from occurring. In addition, the connection provides more wood cover over the embedded steel plate, which potentially may increase the connection's fire rating as compared to typical connections. Two specimens, based on a baseline connection developed using the design procedure presented, were monotonically loaded until failure. Unlike the first specimen, the second was reinforced in the perpendicular-to-grain direction using self-tapping screws. Failure mechanisms were analyzed, and performance characteristics related to the connection's strength, stiffness, and ductility were evaluated. Results indicated that the reinforced specimen exhibited higher strength, stiffness, and ductility compared to the unreinforced specimen. The reinforced specimen showed improvements of 9.49% and 42.2% in yielding and ultimate moment, respectively, compared to the unreinforced specimen. Moreover, an improvement of 31.3% in ductility was obtained using perpendicular-to-grain reinforcement. / Master of Science / Due to the variability of wood properties and its brittle behavior, the joints of wood buildings have traditionally been designed to resist gravity loads only. These types of loads result in predictable behavior of structural wood members at the joints, which helps in simplifying the design process. However, when wood structures are subjected to lateral loads, such as earthquake and wind loads, their joints are likely to fail abruptly as the building sways, resulting in sudden, unpredictable collapse. The purpose of this research was to develop and test a high-strength wood structural joint that can fail gradually and predictably. A design procedure for such a joint is presented herein. The proposed glue-laminated wood joint utilizes an embedded steel plate with a reduced section that acts as a ductile link. This configuration is intended to fail through gradual deformation of the ductile link, thus preventing brittle wood failure at the joint. In addition, this joint provides more wood cover over the embedded steel plate, which potentially may increase the fire resistance of the joint compared to typical configurations. Two specimens, based on a baseline joint developed using the design procedure presented, were subjected to slowly-increasing loads until failure. Unlike the first specimen, the second specimen was reinforced in the direction perpendicular to wood grain using long screws to prevent separation of wood layers. Failure mechanisms were analyzed, and the performance characteristics of the two specimens were evaluated and compared. Results indicated that the reinforced specimen exhibited higher strength and improved ductility at failure.

The paper presents details of a numerical analysis and simplified construction of strengthened glue laminated beams. Glue laminated beams are strengthened through the use of steel reinforcements embedded between the lamellas of the beams. The study compares the numerical results from reinforced and prestressed beams, simply reinforced beams and non-modified beams. Parametric studies were undertaken to evaluate the effects on reinforcement thickness, beam span, prestressing force and prestressing loss. Modified and prestressed beams with wide spans and large dimensions had a significantly higher design load compared to non-modified beams with similar geometry and span. In the most beneficial cases, a load increase of 438% was observed for point load at midspan and 346% for uniformly distributed load.

In this thesis, the behaviour of glued laminated timber combined with hardwood and softwood lamallae is investigated. The influence of hardwood in the tension and compression zone, in terms of strength and stiffness is evaluated.The basis of evaluation consists of determining the behaviour of beams with various combinations of hardwood solely in the tension zone along with beams with hardwood in the tension and compression zone. The influence of different amount of hardwood for both cases is studied my means of experimental and analytical methods.Experimental data attained by performing bending tests are evaluated for different combinations made from spruce and oak. By comparing the experimental and analytical data an increase in the strength and stiffness in various combinations is observed and potrayed which varies based on different wood species.

Para caracterizar a linha adesiva e a penetração do adesivo em de vigas de madeira lamelada colada (MLC), quatro espécies de madeira de reflorestamento Pinus sp, Schizolobium amazonicum (Paricá), Lyptus® e Tectona grandis (Teca) e quatro adesivos diferentes: (i) Purbond HB S390: adesivo à base de poliuretano; (ii) Jowapur 686.60: adesivo à base de poliuretano; (iii) Cascophen: adesivo à base de resina fenol-resorcinol e (iv) Melamina (MUF): adesivo à base de folmaldeído, foram utilizados na preparação de peças de madeira lamelada colada (MLC), seguindo instruções dos fabricantes dos adesivos. De cada viga, foram cortadas duas lâminas para analisar as características da adesão por meio de técnicas de microscopia, usando-se corpos de prova com dimensões adequadas a cada técnica utilizada. As linhas adesivas e as regiões de interfase foram submetidas à observação por microscopia óptica com luz polarizada, microscopia de fluorescência, microscopia eletrônica de varredura, microscopia confocal e microtomografia de raios X. Os adesivos Purbond HB S390 e Jowapur 686.60, apresentaram bolhas de ar na linha de cola após a cura, que podem ser relacionadas com comportamentos observados no cisalhamento das amostras, e o segundo ainda apresentou cristais e evidencias de tensão residual. Os adesivos MUF e Cascophen não apresentaram evidencias da presença de nenhum desses elementos. Os resultados das técnicas de microscopia mostraram que a penetração do adesivo resultante em cada caso, ocorreu nos vasos, no lúmen dos traqueídes e ao longo dos raios e está ligada à densidade das madeiras. As técnicas microscópicas mostraram-se complementares com informações características de cada uma delas. O tingimento das amostras facilitou a visualização das imagens na microscopia ótica, sendo útil na análise de tensão residual, formação de cristais e bolhas no adesivo, porém a técnica não permitiu a determinação de penetração de adesivo na interfase. A microscopia de fluorescência permitiu a visualização de todas as fases da linha adesiva, a profundidade da penetração grosseira e a identificação de bolhas. Porem os resultados têm pouco foco, baixa resolução e necessitam do uso de sonda fluorescente. A fluorescência confocal também permitiu identificação de todas as fases da região adesiva, profundidade de penetração grosseira e celular, com alta resolução e contraste de cores, porém gerou sombras nas regiões de irregularidade da superfície da amostra. A microscopia eletrônica de varredura permitiu a determinação precisa da espessura da linha adesiva na fase pura, a visualização do preenchimento de alguns lúmens, mas a escala de tons de cinza pode gerar dúvidas na diferenciação do adesivo e madeira. A micro CT permitiu determinar a penetração em três dimensões, com alta resolução, porém necessita de contraste adicionado ao adesivo para ensaio em madeiras de alta densidade. A partir desse estudo identificou-se a necessidade do uso das técnicas microscópicas como complemento nos estudos de ligações adesivas em MLC. Os resultados do presente trabalho indicam que as técnicas microscópicas devam passar a fazer parte das normas para testes de qualidade de MLC. / Four species of reforestation species Pinus sp, Schizolobium amazonicum (Paricá), Lyptus® e Tectona grandis (Teca) and four different adhesives (i) Purbond HB S390: poliurethane based adhesive; (ii) Jowapur 686.60: poliurethane based adhesive; (iii) Cascophen: phenol-resorcinol resin based adhesive and (iv) Melamine (MUF): formaldehyde based adhesive were used in the assembling of glulam beans, according to the adhesive manufacturer\'s instructions. In order to characterize the glue line and the penetration of the adhesive in the preparation of the beans. Two slices were cut from each bean to evaluate the adhesion characteristics by microscopy techniques, using specimens appropriate for each technique. Thus, the glue lines and interphase regions were submitted to polarized light optical microscopy, fluorescence microscopy, scanning electron microscopy, confocal microscopy and X-ray microtomography. Purbond HB S390 and Jowapur 686.60 adhesives presented air bubbles in the glue line after cure, that can be related with shear tests, while the second also presented crystals and evidences of residual tension. MUF and Cascophen do not presented evidences of the presence of any of these elements. The results from the microscopic techniques revealed details of the penetration of the adhesives in each case and occurred in the vessels, lumen of the tracheid and along of the radius and is deeply related to the density of each wood. Microscopic techniques reveled to be complimentary with characteristic information from each one of them. Dyeing of the samples make easier the observation of optical microscopy images being useful in the analysis of residual tension, presence of bubbles and crystals in the adhesive. However, this technique did not allowed the determination of the characteristics of penetration in the samples. Fluorescence microscopy allowed observing of all phases of the glue line, coarse penetration deepness and identification of bubbles, but the results has poor focus and low resolution and requires fluorescent probe. Confocal fluorescence microscopy also allowed the identification of all phases in the glue line, coarse and cell penetration deepness observation, with high resolution and color contrast, however it generated shade in the irregular regions of the sample surface. Scanning electron microscopy permitted the precise determination of the glue line thickness in the pure phase, visualization of the fulfilling of some lumens, but the grey scale can generate doubts concerning the differentiation of the adhesive and wood in the images. X-ray microtomography allowed determining the 3D penetration profile, with high resolution, however it needed that contrast in introduced in the adhesive when testing high density woods. Thus, from this work it was verified the need of using microscopic techniques as complementary tool in studding adhesion in glulam beans. The results of this work point that the microscopic techniques should be part of the standards for gluelam quality testing.

O objetivo deste trabalho é o estudo da resistência e rigidez das emendas dentadas à fadiga na tração com as finalidades de propor uma metodologia de ensaio e estimar um coeficiente redução da resistência à fadiga na tração das emendas dentadas em relação a madeira maciça, contribuindo para a caracterização destas ligações. Foram utilizadas duas espécies de madeira, Pinus caribea hondurensis e Eucalyptus grandis, e dois tipos de adesivos sendo um industrial à base de fenol-resorcinol e um poliuretano à base de resina de mamona desenvolvido no Instituto de Química de São Carlos. Foram realizados ensaios cíclicos para três níveis de freqüência (1 Hz, 5 Hz e 9 Hz) e três níveis de tensão (90%, 75% e 60% da resistência) a uma relação R = 0,1. Com base nos resultados foram propostos coeficientes de redução de resistência à fadiga da emenda dentada em relação à madeira maciça e verificar que não há variação significativa da rigidez em função do número de ciclos. / The aims of this work is the study of the fatigue strength and stiffness of the finger joints in tension, in order to propose an experimental methodology and to determine a strength reduction coefficient, in relation to solid wood, contributing to the characterization of these joints. Two wood species, Pinus caribea hondurensis and Eucalyptus grandis, and two adhesives types were used, being an industrial one phenol-resorcinol based and a poliuretane castor oil resin based, developed at the lnstitute of Chemistry of São Carlos. Cyclical tests were accomplished in three frequency levels (1 Hz, 5 Hz and 9 Hz) and three tension levels (90%, 75% and 60% of the strength) to a relationship R = 0,1. Based on the results, it were proposed strength reduction coefficients in relation to solid wood, for finger joints under fatigue in tension, and to concluded that there is no significant variation in stiffness with the cycle numbers.

Uma das técnicas para viabilizar a utilização de madeira de reflorestamento é a fabricação de Madeira Laminada Colada (MLC), que consiste na colagem de lâminas de madeira. Essas lâminas são obtidas da união de peças, por meio de emendas dentadas coladas. Para um bom desempenho de estruturas de MLC, a escolha do adesivo é fundamental. Atualmente, o Cascophen é o adesivo mais utilizado. Nesse contexto, o objetivo dessa pesquisa foi avaliar a eficiência, quanto à resistência à tração de emendas dentadas, de adesivos (Cascophen, Wonderbond, Cascorez, Melaminauréia formaldeído e Poliuretano à base de óleo de mamona) na colagem de madeira da espécie Pinus hondurensis, utilizando um padrão de emenda com perfil vertical. Também foram feitas comparações entre os perfis vertical e horizontal usando apenas o adesivo Cascophen. Os trabalhos experimentais foram baseados em ensaios de tração em peças com dimensão estrutural. Foram testadas 150 peças emendadas e 25 peças sem emendas (para comparações), com dimensões nominais de 3,5 x 7 x 200 cm. As principais análises feitas foram as comparações entre todos os adesivos com perfil vertical e entre os perfis vertical e horizontal, e uma análise dos modos de ruptura e suas relações com a porcentagem de madeira de inverno e o módulo de elasticidade. A comparação entre todos os adesivos com perfil vertical revelou que apenas o Cascophen e o Cascorez não possuem diferenças estatísticas. O adesivo que mais se destacou foi o Wonderbond, que teve desempenho superior ao do Cascophen, inclusive. Também se concluiu que não existe diferença estatística significativa na resistência à tração dos perfis vertical e horizontal / One of the techniques to allow for the use of reforestation lumber is the production of glued laminated timber (glulam), which consists of the gluing of lumber laminas. Those laminas are obtained of the union of pieces, through glued finger joints. For a good performance of glulam structures, the choice of the adhesive is fundamental. Traditionally, Cascophen is the most adhesive used in the production of glulam. In that context, the aim of the present research is to evaluate the efficiency, with relation to the strength, of adhesives (Cascophen, Wonderbond, Cascorez, Melamine urea formaldehyde and castor oil Polyurethane) in the gluing of Pinushondurensis lumber, using a joint pattern with vertical profile. Comparisons between the vertical and horizontal profiles using only Cascophen adhesive were made. The experimental analyses were based on tensile tests in pieces with structural dimension. For this purpose, 150 joint pieces and 25 pieces without joints (for comparisons) with nominal dimensions of 3,5 x 7,0 x 200 cm were tested. The main analyses performed were the comparison among all the adhesives with vertical profile, the comparison between the vertical and horizontal profiles for the Cascophen adhesive and, finally, an analysis of the failure modes and their relations with both latewood percentage and elasticity modulus. The comparison among all the adhesives with vertical profile revealed that Cascophen and Cascorez do. Not have statistical differences. The best adhesive was Wonderbond, which had a better performance than Cascophen. Concerning the vertical and horizontal profiles, no significant statistical difference was found in the tensile strength

O uso de madeira laminada colada (MLC) em países da América do Norte e Europa é bastante difundido. No Brasil, apesar da técnica existir desde o século XX, a sua utilização é incipiente em relação à possibilidade de utilização desse produto. O desenvolvimento de produtos a partir de lâminas de madeira com pequenas dimensões se deve à necessidade de se eliminar defeitos nas peças estruturais, tais como nó e medula, os quais reduzem significativamente as características de resistência e elasticidade da peça, em comparação com a madeira sã. O aumento da resistência e da rigidez da viga de MLC pode ser conseguido por meio da classificação das lâminas, seguida de uma disposição adequada das mesmas. Outra maneira de se obter um aumento nas propriedades de rigidez e de resistência de elementos estruturais de MLC é a utilização de polímeros reforçados com fibras (PRF) na região mais tracionada da viga. Esta providência, além de melhorar as propriedades mecânicas do elemento estrutural, aumenta a confiabilidade em relação ao modo de ruptura. Dentro deste escopo, este trabalho procura estabelecer critérios para fabricação de vigas retas de MLC, sem e com reforço de fibras de vidro, a partir de informações obtidas na revisão bibliográfica, abordando aspectos relacionados à resistência da madeira (espécie e qualidade da madeira, quantidade de nós, inclinação das fibras em relação ao eixo longitudinal das peças, presença de medula), dimensões das lâminas, espaçamento mínimo entre emendas dentadas de uma mesma lâmina e posicionamento das emendas dentadas na viga, colagem e prensagem. Foi proposto um modelo numérico para cálculo do momento fletor resistente e da rigidez à flexão (EI) de vigas de MLC sem reforço e reforçadas com PRF. O modelo admite a validade da hipótese de distribuição linear de deformações ao longo da altura da viga, comportamento elásto-frágil-linear da madeira na tração paralela às fibras e bi-linear na compressão paralela às fibras. Este modelo foi implementado a um programa computacional que por meio de um processo interativo determina o valor do momento fletor resistente e da rigidez à flexão. Resultados experimentais obtidos por meio de ensaios em vigas de MLC reforçadas com fibra de vidro indicam a validade do modelo teórico proposto. Também foram realizados ensaios de delaminação para avaliar a eficiência dos adesivos Phenol-resorcinol e Epóxi AR-300 quando submetidos a variação de temperatura, umidade e pressão. Resultados obtidos indicaram a eficiência destes adesivos. / Glulam beam is used in countries of North America and Europe. In Brazil, the technique exists since the 20th Century, but its use is incipient in relation to the possibility of the product. The development of products that use wood with small dimensions is important to eliminate defects in the structural pieces, such as knot and marrow, which reduce the strength and stiffness of the piece. The increase of the strength and stiffness of the glulam beam can be achieved through the classification of the lumber, or by an appropriate disposition of the lumber. Another way to obtain an increase in the strength and stiffness of the glulam beam it is the use of fiber reinforced polymers (FRP). This way, besides improving the mechanical properties of the structural element, increases the reliability in relation to the rupture way. This work establish criteria for production of glulam beams, without and with reinforce of glass fiber, starting from information obtained in the bibliographical revision, approaching aspects related to the straight of the wood (species and quality of the wood, inclination of the fibers in relation to the longitudinal axis of the pieces, marrow presence), dimensions of the sheets, minimum spacing between bitten amendments of a same sheet and positioning of the jagged amendments in the beam. A numeric model for calculation of the straight moment and flexural stiffness (EI) of glulam beams was proposed. The model admits the validity of the hypothesis of lineal distribution of deformations along the height of the beam, elastic-fragile-lineal behavior of the wood in the parallel traction and bi-lineal in the parallel compression of the fibers. This model was implemented in a computer program that uses an interactive process to determine the value of the straight moment and of the flexural stiffness. Experimental results obtained indicate the validity of the theoretical model proposed. A test of delamination was done to evaluate the efficiency of the adhesives Phenol-resorcinol and Epoxy AR-300 when submitted to temperature, humidity and pressure variation. Results indicated the efficiency of these adhesives.

A confecção e as características mecânicas e morfológicas de peças de madeira laminada colada (MLC) preparadas a partir de lamelas de Schizolobium amazonicum Herb, Paricá, coladas com três adesivos, Cascophen, Melamina e Pubond, após serem submetidas a dois tipos de tratamento químico preservante, arsanato de cobre cromatado (CCA) e borato de cobre cromatado (CCB) foram avaliadas. Após seleção visual das lamelas e seleção, foram determinados os módulos de elasticidade das mesmas visando a preparação de vigas com a distribuição mais uniforme possível deste parâmetro. A densidade aparente média da madeira de paricá determinada segundo a Norma NBR 7190:1997,usando todas as lamelas, após serem submetidas à seleção pela classificação visual foi igual a 370 kg/m3. As lamelas foram então coladas, prensadas e submetidas aos tratamentos CCA e CCB. A partir dessas peças foram cortados corpos de prova para ensaios de flexão estática, cisalhamento, delaminação e fluência, conforme a norma CSA 0112.9:2004. Os valores do módulo de elasticidade demonstraram que, vigas produzidas com Melamina, há aumento nos valores de MOE para as amostras tratadas por CCA (7656 ± 477 MPa) e CCB (8152 ± 570 MPa) em relação à amostra sem tratamento (7171 ± 519 MPa). No caso do CCA, esse aumento é menos significativo. No caso das amostras coladas com Cascophen praticamente não houve diferença estatística entre os módulos para as peças sem tratamento (9150 ± 672 MPa) e aquelas tratadas com CCA (8764 ± 908 MPa) e CCB (9822 ± 791 MPa), sendo a variação numérica maior no segundo tratamento. A mesma observação vale para as amostras coladas com Purbond sem tratamento (8119 ± 725 MPa), quando comparadas com aquelas tratadas com CCA (8154 ± 951 MPa) e com CCB (8331 ± 863 MPa). Nos ensaios de cisalhamento observou-se elevada incidência de eventos de compressão o que sugere que alta porosidade da madeira, tenha provocado a compressão do corpo de prova, em detrimento da ruptura do corpo de prova. Observou-se que a Melamina e o Purbond apresentaram a maior frequência de ruptura tipo 100% na madeira, sugerindo que esses apresentaram maior adesividade, não havendo diferença significativa entre os corpos de prova secos e saturados. Em relação à força de ruptura, notou-se, que os adesivos apresentam melhor desempenho quando secos (quando se apresentaram na ordem Melamina (3,7 ± 0,9 MPa), Purbond (3,5 ± 0,96 MPa) e Cascophen (3,4 ± 0,8 MPa).) Nenhum corpo de prova apresentou delaminação, todos permaneceram colados e praticamente intactos, o que sugere que os adesevos e os tratamentos químicos utilizados são adequados à preparação de MLC à partir de lamelas de Paricá. Os ensaios de fluência que a estabilização nos deslocamentos ocorreu em torno de 90 h, para os dois corpos de prova. Observou-se uma diminuição nos valores das linhas laterais e centrais, após ensaiadas, em relação aos valores antes do ensaio, o que pode estar relacionado com a porosidade e densidade da madeira, que permitiu a compressão irreversível do corpo de prova, sugerindo que as características da madeira são mais marcantes que o efeito do adesivo. / Glued Laminated Wood timber samples were confectioned from Schizolobium amazonicum Herb, Paricá, using Cascophen, Melamine and Purbond adhesives, and submitted to preservative chemical treatment by cromated copper arsenate (CCA) and cromated copper bromated. These samples were characterized regarding their mechanical and morphological characteristics. After visual evaluation and selection of the lamellas they had their elasticity modulus measured in order to prepare timbers with the most uniform distribution as possible. The apparent density of Paricá wood had been determined as 370 kg/m3 according to the NBR 7190 Standard, using all the lamellas after submitted to the visual inspection. Then the lamellas were glued pressed and submitted to the CCA and CCB treatments. From these pieces specimens for the elasticity, shear, delamination and creep tests were cut. The elasticity modulus showed that for samples glued with Melamine there was an increasing in the MOE values for samples treated with CCA (7656 ± 477 MPa) and CCB (8152 ± 570 MPa) when compared with the untreated sample (7171 ± 519 MPa). In the CCA such increase is less remarked. For Cascophen practically the same values were found for untreated (9150 ± 672 MPa) and CCA (8764 ± 908 MPa) or CCB (9822 ± 791 MPa) treated specimens. Same conclusion is valid for the Purbond glue samples once untreated (8119 ± 725 MPa) and CCA (8154 ± 951 MPa) or CCB (8331 ± 863 MPa) samples presented statistically similar MOE values. Shear tests presented high incidence of compression events, suggesting that the porosity of the Wood facilitates this behavior instead of rupture of the specimen. Melamine and Purbond glued specimens presented higher frequency of 100% in-wood rupture, reveling higher adhesion in such glues, without differences in dry or saturated samples. Regarding the rupture force all the adhesives promoted more resistence when dry, in the order: Melamine (3,7 ± 0,9 MPa), Purbond (3,5 ± 0,96 MPa) e Cascophen (3,4 ± 0,8 MPa). Any specimens presented delamination, remaining all of them intact showing that the adhesives and chemical treatments are adequate for the preparation of Laminated Glued Wood from Paricá. Creep tests revealed that the stabilization in the displacements stabilized after 90h of pressure in both the specimens that resisted until the end of the test. There was a diminishing on the length of both central and lateral glue lines, at the end of the test which can be related to the porosity and density of the Paricá wood allowing the irreversible compression of the specimens, suggesting that such characteristics are more significant than the adhesive and chemical treatment used.

O objetivo deste trabalho é o estudo da resistência e rigidez das emendas dentadas à fadiga na tração com as finalidades de propor uma metodologia de ensaio e estimar um coeficiente redução da resistência à fadiga na tração das emendas dentadas em relação a madeira maciça, contribuindo para a caracterização destas ligações. Foram utilizadas duas espécies de madeira, Pinus caribea hondurensis e Eucalyptus grandis, e dois tipos de adesivos sendo um industrial à base de fenol-resorcinol e um poliuretano à base de resina de mamona desenvolvido no Instituto de Química de São Carlos. Foram realizados ensaios cíclicos para três níveis de freqüência (1 Hz, 5 Hz e 9 Hz) e três níveis de tensão (90%, 75% e 60% da resistência) a uma relação R = 0,1. Com base nos resultados foram propostos coeficientes de redução de resistência à fadiga da emenda dentada em relação à madeira maciça e verificar que não há variação significativa da rigidez em função do número de ciclos. / The aims of this work is the study of the fatigue strength and stiffness of the finger joints in tension, in order to propose an experimental methodology and to determine a strength reduction coefficient, in relation to solid wood, contributing to the characterization of these joints. Two wood species, Pinus caribea hondurensis and Eucalyptus grandis, and two adhesives types were used, being an industrial one phenol-resorcinol based and a poliuretane castor oil resin based, developed at the lnstitute of Chemistry of São Carlos. Cyclical tests were accomplished in three frequency levels (1 Hz, 5 Hz and 9 Hz) and three tension levels (90%, 75% and 60% of the strength) to a relationship R = 0,1. Based on the results, it were proposed strength reduction coefficients in relation to solid wood, for finger joints under fatigue in tension, and to concluded that there is no significant variation in stiffness with the cycle numbers.

Nowadays, timber is widely used in construction industry thanks to its availability and good properties. The use of solid (sawn) timber is not always proper since it is only available up to certain dimensions. Therefore, the so-called Engineered WoodProducts (EWPs) have been introduced to cope with the different design needs of structures. The Glued laminated Timber (glulam) is a type of EWPs that consists of smallsections of timber laminates glued together to form beams and columns. Glulam can be manufactured in almost any size and shape; it can also be tapered or notched. However, notching a beam at its end leads to a stress concentration at the re-entrantcorner of the notch due to the sudden change in the notched beam’s cross section. The concentration of shear and tensile stresses perpendicular to the grain can lead to a catastrophic brittle failure caused by the crack propagation from the notch corner. Crack opening due to tensile stresses perpendicular to grain is the most common failure at the notch corner and it is always taken into design consideration. However,shear component is usually exists and must be also considered in design to guarantee the safety of the structure. Currently, only the normal forces perpendicular to the beam’s axis are considered in the design of the reinforcement in design handbooks. The aim of this thesis was to study the structural behavior of notched glulam beams reinforced by adhered plywood panels and FRP. The carrying capacity of the notched glulam beams at their ends is the main subject of this thesis. In addition, a review of the notched beams design, reinforcements, and analysis theories are included. Experimental series of three point bending tests with notched glulam beams withdifferent configurations of reinforcement was carried out in lab. Deformations and forces were measured both with conventional techniques and with contact-free measurement systems - ARAMIS. On the other hand, a simple model of two dimensional plane stress element has been created of the unreinforced notchedbeam in ABAQUS. The normal and shear stresses were calculated for a horizontalpath of 100 mm in length starting from the notch tip. Afterwards, the mean stresseswere determined for the same path and have been used in calculations. The Mean Stress Approach has been adopted in the hand calculations to calculate the crack length and the failure load according to the ABAQUS model. Accordingly, the failure load was about 40 kN for the unreinforced beams. Also, Eurocode 5 has been used to calculate the failure load which gave a value of 20.2 kN for the unreinforced beams. The average maximum applied load in tests was 30 kN for the unreinforced beams while it reached about two and a half times this value for the CF-reinforced and the plywood-reinforced beams. / Tack vare sina goda egenskaper används trä i byggnadskonstruktioner i allt storeomfattning. Konstruktionsvirke (sågade trävaror) kan dock inte alltid användas pågrund av de begränsade dimensioner som finns tillgängliga. På grund av bl a dettahar ett flertal så kallade engineer wood products (EWP) utvecklats. Limträ är en typav EWP som består av sammanlimmade lameller som bygger upp tvärsnitt i balkareller pelare. Limträ kan tillverkas i nästan godtycklig storlek och form och kan enkeltförses med t ex urtag. Vid urtag i balkändar nära upplag uppstår högaspänningskoncentrationer vid urtagets horn på grund av geometrin. Koncentrationenav normalspänningar och skjuvspänningar kan leda till plötsligt brott på grund avsprickpropagering från urtagets hörn, något som måste tas hänsyn till viddimensionering. Dagens dimensioneringsmetoder är baserade på att man tar hänsyntill enbart normalspänningarna vinkelrät fiberriktningen.Målet med detta arbete har varit att studera beteendet hos limträbalkar med urtag vidupplag som förstärkts med fiberarmering eller plywood. Huvudmålet har varit attbestämma balkarnas bärförmåga, vilket skett genom att genomföra försök med olikakonfigurationer vad gäller förstärkningsmaterial och dess utformning. Vidare harolika dimensioneringsmetoder från litteraturen studerats.Kraft och förskjutning under provningarna uppmättes dels med traditionellamätmetoder, men deformationerna mättes även med beröringsfri metod, ARAMIS.En enkel tvådimensionell finit elementmodell skapades och analyserades i ABAQUSför analys av oförstärkt balk. Normalspänningar och skjuvspänningar beräknades ochmedelspänningarna längs en på förhand definierad sträcka beräknades.Medelspänningskriteriet användes sedan för att uppskatta balkens bärförmåga.Enligt FE-beräkningarna uppskattades bärförmågan för de oförstärkta balkarna till ca40 kN. Provningarna gav ett medelvärde på balkarnas bärförmåga på ca 30 kN,medan de förstärkta balkarna hade en 2,5 gånger högre bärförmåga. Skillnadenmellan FE-beräkningarna och provningarna kan förklaras med den osäkerhet somfinns vad gäller det aktuella trämaterialets egenskaper.Beräkningar enligt Eurokod 5 gav en karakteristisk bärförmåga på 20,2 kN.