Influence of joints on the seismic response of traditional timber frames in Turkey

Influence of joints on the seismic response of traditional timber frames in Turkey Güzide Aslankaya ADVERTIMENT La consulta d’aquesta tesi queda condicionada a l’acceptació de les següents condicions d'ús: La difusió d’aquesta tesi per mitjà del r e p o s i t o r i i n s t i t u c i o n a l UPCommons (http://upcommons.upc.edu/tesis) i el repositori cooperatiu TDX ( h t t p : / / w w w . t d x . c a t / ) ha estat autoritzada pels titulars dels drets de propietat intel·lectual únicament per a usos privats emmarcats en activitats d’investigació i docència. No s’autoritza la seva reproducció amb finalitats de lucre ni la seva difusió i posada a disposició des d’un lloc aliè al servei UPCommons o TDX. No s’autoritza la presentació del seu contingut en una finestra o marc aliè a UPCommons (framing). Aquesta reserva de drets afecta tant al resum de presentació de la tesi com als seus continguts. En la utilització o cita de parts de la tesi és obligat indicar el nom de la persona autora. ADVERTENCIA La consulta de esta tesis queda condicionada a la aceptación de las siguientes condiciones de uso: La difusión de esta tesis por medio del repositorio institucional UPCommons (http://upcommons.upc.edu/tesis) y el repositorio cooperativo TDR (http://www.tdx.cat/?localeattribute=es) ha sido autorizada por los titulares de los derechos de propiedad intelectual únicamente para usos privados enmarcados en actividades de investigación y docencia. No se autoriza su reproducción con finalidades de lucro ni su difusión y puesta a disposición desde un sitio ajeno al servicio UPCommons No se autoriza la presentación de su contenido en una ventana o marco ajeno a UPCommons (framing). Esta reserva de derechos afecta tanto al resumen de presentación de la tesis como a sus contenidos. En la utilización o cita de partes de la tesis es obligado indicar el nombre de la persona autora. WARNING On having consulted this thesis you’re accepting the following use conditions: Spreading this thesis by the i n s t i t u t i o n a l r e p o s i t o r y UPCommons (http://upcommons.upc.edu/tesis) and the cooperative repository TDX (http://www.tdx.cat/?localeattribute=en) has been authorized by the titular of the intellectual property rights only for private uses placed in investigation and teaching activities. Reproduction with lucrative aims is not authorized neither its spreading nor availability from a site foreign to the UPCommons service. Introducing its content in a window or frame foreign to the UPCommons service is not authorized (framing). These rights affect to the presentation summary of the thesis as well as to its contents. In the using or citation of parts of the thesis it’s obliged to indicate the name of the author. INFLUENCE OF JOINTS ON THE SEISMIC RESPONSE OF TRADITIONAL TIMBER FRAMES IN TURKEY DECEMBER 2019 GÜZİDE ASLANKAYA UNIVERSITAT POLITECHNICA DE CATALUNYA BARCELONATECH PhD. THESIS Department of Architectural Technology Architectural, Building Construction and Urbanism Technology Directors of Thesis : PhD. Albert Albareda-Valls PhD. Jaume Avellaneda Diaz-Grande Tutor of Thesis : PhD. Ana Maria Lacasta Palacio PhD. TESIS DICIEMBRE 2019 Directores de La Tesis: PhD. Albert Albareda-Valls PhD. Jaume Avellaneda Diaz-Grande Tutora de La Tesis: PhD. Ana Maria Lacasta Palacio GÜZİDE ASLANKAYA INFLUENCIA DE LAS JUNTAS EN LA RESPUESTA SÍSMICA DE ESTRUCTURAS DE MADERA TRADICIONALES EN TURQUIA UNIVERSITAT POLITECHNICA DE CATALUNYA BARCELONATECH Departamento de Tecnología de la Arquitectura Technologia de la Arquitectura, de la Edificion y del Urbanismo vi FOREWORD I hereby declare that all information in this thesis has been obtained and presented in accordance with academic rules and ethical conduct. Firstly, I would like to sincerely thank my thesis advisors Albert Albareda-Valls and Jaume Avellaneda Diaz-Grande for their inexhaustible patience and guidance. I am grateful to Cenk Ustundag for his great support and advices. I would particularly like to thank my family who have always encouraged me to go further. I would also like to thank workers of UPC Material Laboratory for their valuable helps during experiments. December 2019 Güzide ASLANKAYA vii viii CONTENTS FOREWORD ............................................................................................................. vi CONTENTS ............................................................................................................. viii ABBREVIATIONS ................................................................................................... xi LIST OF TABLES .................................................................................................. xiii LIST OF FIGURES ................................................................................................ xvi ABSTRACT ............................................................................................................ xxii RESUMEN ............................................................................................................. xxiii 1. INTRODUCTION .................................................................................................. 1 1.1 General ............................................................................................................... 1 1.2 Objectives ........................................................................................................... 5 1.3 Significance of the topic ..................................................................................... 6 2. STATE OF THE ART ........................................................................................... 7 2.1 Traditional Turkish Timber Structure (Hımış) ................................................... 7 2.2 Traditional Turkish Timber Joints ...................................................................... 9 2.3 Seismic Behaviour of Timber Structure ........................................................... 13 2.4 Strengthening of Timber Structures with FRP ................................................. 16 3. GLOBAL ANALYSIS ......................................................................................... 20 3.1 Definition of Model Parameters ....................................................................... 21 3.2 Dynamic Non-Linear Analysis ......................................................................... 24 3.3 Analysis of Results ........................................................................................... 26 3.4 Response Spectrum Analysis ........................................................................... 31 4. EXPERIMENTAL ANALYSIS OF TIMBER JOINTS ................................... 36 4.1 General Mechanical Behaviour of Timber ....................................................... 37 4.2 Timber Characterization Tests ......................................................................... 40 4.2.1. Compression parallel to the grain ............................................................ 40 4.2.2. Compression perpendicular to the grain .................................................. 44 4.2.3. Bending .................................................................................................... 48 4.2.4. Discussion of test resuts ........................................................................... 56 4.3 Lap Joint Tests ................................................................................................. 57 4.3.1. Monotonic tests on unreinforced specimens ............................................ 57 4.3.2. Monotonic tests on reinforced specimens ................................................ 62 4.3.3. Cyclic tests on reinforced specimens ....................................................... 67 4.3.4. Discussion of test results .......................................................................... 74 4.4 Mortise Tenon Joint Tests ................................................................................ 75 4.4.1. Monotonic tests on unreinforced specimens ............................................ 76 4.4.3. Cyclic tests on unreinforced specimens ................................................... 85 4.4.4. Cyclic tests on reinforced specimens ....................................................... 92 4.4.5. Discussion of test results .......................................................................... 98 5. NUMERICAL ANALYSIS ................................................................................. 99 5.1 Material Model ................................................................................................. 99 ix 5.2 Failure Criteria ............................................................................................... 103 5.3 Numerical Modelling of Lap Joint ................................................................. 103 5.3.1. Geometric constraints, mesh and loading .............................................. 104 5.3.2. Analysis of the unreinforced model under monotonic loading .............. 106 5.3.3. Analysis of the reinforced model under monotonic loading .................. 109 5.3.4. Comparison of analysis results .............................................................. 113 5.4 Numerical Modelling of Mortise Tenon Joint ................................................ 114 5.4.1. Geometric constraints, mesh and loading .............................................. 114 5.4.2. Analysis of the unreinforced model under monotonic loading .............. 116 5.4.3. Analysis of the reinforced model under monotonic loading .................. 119 5.4.4. Comparison of analysis results .............................................................. 125 6. GLOBAL SEMI-RIGID ANALYSIS ............................................................... 126 6.1 Definition of Model Parameters ..................................................................... 127 6.2 Lateral Load Analysis .................................................................................... 130 6.3 Analysis Results of Unreinforced and CFRP Reinforced Structures ............. 131 7. CONCLUSIONS ................................................................................................ 152 REFERENCES ....................................................................................................... 157 STANDARDS ......................................................................................................... 160 xvi LIST OF FIGURES Figure 2. 1: Turkish timber house (Turkish Timber Association, 2018) ..................... 8 Figure 2. 2: Timber frame members ............................................................................ 8 Figure 2. 3:Traditional Turkish wooden slabs. Double plates with one-way (a), single slab (b) and two-way slabs (c) using double plates.............................................. 9 Figure 2. 4: Timber joints in frame (Turkish Timber Association, 2018) ................. 12 Figure 2. 5: Timber joints, mortise-tenon (a) and half-lap (b) joints ......................... 12 Figure 2. 6: Timber joints, tongued-grooved (c) and notch (d) joints ....................... 12 Figure 2. 7: The beaviour of timber wall during test (a) and crack pattern in masonry infill (b) (Poletti, 2014) ...................................................................................... 14 Figure 3. 1: The frame configuration for analysis in Aktas´s study (Aktas, 2016) 22 Figure 3. 2: The frame configuration for analysis ..................................................... 22 Figure 3. 3: The non-linear plastic curve of masonry ................................................ 24 Figure 3. 4: Acceleration ............................................................................................ 25 Figure 3. 5: General model of frame .......................................................................... 25 Figure 3. 6: Definition of dead loads ......................................................................... 26 Figure 3. 7: Definition of live loads ........................................................................... 26 Figure 3. 8: Normal forces ......................................................................................... 27 Figure 3. 9: Shear forces ............................................................................................ 27 Figure 3. 10: Bending moment .................................................................................. 28 Figure 3. 11: Global deformations ............................................................................. 28 Figure 3. 12: Axial stresses in sigma-x at masonry surface ....................................... 30 Figure 3. 13: Axial stresses in sigma-y at masonry surface ....................................... 30 Figure 3. 14: Acceleration spectrum .......................................................................... 31 Figure 3. 15: Mode shape 1 and mode shape 2 .......................................................... 32 Figure 3. 16: Mode Shape 3 and mode shape 4 ......................................................... 32 Figure 3. 17: Mode Shape 5 and mode shape 6 ......................................................... 32 Figure 3. 18: The deformation in mode shape 1 ........................................................ 33 Figure 3. 19: Envelope of normal forces .................................................................... 33 Figure 3. 20: Envelope of shear forces....................................................................... 34 Figure 3. 21: Bending moment .................................................................................. 34 Figure 3. 22: The stresses of masonry surfaces in sigma-x........................................ 35 Figure 3. 23: The stresses of masonry surfaces in sigma-y........................................ 35 Figure 4. 1: Directions of wood fibers 37 Figure 4. 2: Compression failure modes of the wood in parallel to grain (a) and perpendicular to the grain (b) (Gibson, 1997). ................................................... 39 Figure 4. 3: Typical stress–strain curves for timber loaded in compression in the longitudinal, radial and tangential directions and for tension in the longitudinal direction (Holmberg, 1999). ............................................................................... 39 Figure 4. 4: The dimension of specimens for compression test parallel to the grain . 41 Figure 4. 5: Three specimens for compression test parallel to the grain .................... 41 Figure 4. 6: Compression test parallel to the grain .................................................... 42 Figure 4. 7: Load-deformation curve for compression test parallel to the grain........ 43 Figure 4. 8: Compression failure patterns, a) Crushing, b) Wedge split, c) Shearing d) Splitting, e) Compression and shear parallel grain, f) Brooming or endrolling, (ASTM D143-14). ................................................................................. 43 Figure 4. 9: The failure patterns of compression tests parallel to the grain ............... 44 xvii Figure 4. 10: The dimension of specimens for compression test perpendicular to the grain ................................................................................................................... 45 Figure 4. 11: Three specimens for compression test perpendicular to the grain ........ 45 Figure 4. 12: Compression test perpendicular to the grain ........................................ 46 Figure 4. 13: Load-deformation curve for compression test perpendicular to the grain ............................................................................................................................ 46 Figure 4. 14: The failure patterns of compression tests perpendicular to the grain ... 47 Figure 4. 15: The dimension of specimens for bending tests ..................................... 48 Figure 4. 16: Test arrangement for measuring local modulus of elasticity in bending ............................................................................................................................ 48 Figure 4. 17: The test set up for measuring local modulus of elasticity in bending .. 50 Figure 4. 18: Load-deformation curve for the range of 0.1 Fmax-0.4 Fmax ................. 50 Figure 4. 19: Test arrangement for measuring global modulus of elasticity in bending ............................................................................................................................ 51 Figure 4. 20: The configuration of bending tests (Dimensions are presented in mm)51 Figure 4. 21: The test set up for measuring global modulus of elasticity in bending 52 Figure 4. 22: Load-deformation curve for bending test ............................................. 53 Figure 4. 23: Compression failure patterns a) Simple tension, b) Cross grain tension, c) Splintering tension, d) Brash tension, e) Compression, f) Horizontal shear (ASTM D143-14) ............................................................................................... 54 Figure 4. 24: The failure pattern of first specimen under bending ............................. 55 Figure 4. 25: The failure pattern of second specimen under bending ........................ 55 Figure 4. 26: The failure pattern of third specimen under bending ........................... 56 Figure 4. 27: Dimensions of specimens in monotonic tests (Dimensions are presented in mm) .................................................................... 57 Figure 4. 28: The set up for monotonic test ............................................................... 59 Figure 4. 29: The loading procedure of monotonic test (BS EN 26891:1991) .......... 59 Figure 4. 30: Distribution of stress of bending test under monotonic load ................ 61 Figure 4. 31: Load-deformation curve of monotonic tests ......................................... 61 Figure 4. 32: The failure pattern of two specimens under monotonic loading .......... 62 Figure 4. 33: Dimensions of reinforced specimens in monotonic test (Dimensions are presented in mm) .................................................................... 63 Figure 4. 34: The selected epoxy and carbon fiber textile for reinforcement ............ 63 Figure 4. 35: General view of reinforced specimens for monotonic test ................... 64 Figure 4. 36: General view of the set up for monotonic test on reinforced specimen 65 Figure 4. 37: Load-deformation curve of monotonic tests on reinforced specimens . 66 Figure 4. 38: Failure pattern of reinforced specimen (LJM1800AR) under monotonic loading ................................................................................................................ 66 Figure 4. 39: The failure pattern of second reinforced specimen (LJM1800BR) ...... 67 Figure 4. 40: Test set up for cyclic test on reinforced specimen................................ 68 Figure 4. 41: The loading procedure of cyclic test proposed by BS EN 12512:2001 68 Figure 4. 42: Load-deformation curve of cyclic tests on reinforced specimens ........ 69 Figure 4. 43: Graphical relationship between total strain, permanent strain and elastic strain ................................................................................................................... 70 Figure 4. 44: Load-deformation curves of specimens at each cycle .......................... 70 Figure 4. 45: Plastic strain-cycles diagram of reinforced specimens ......................... 71 Figure 4. 46: Damage affecting the modulus of elasticity in tested specimens ......... 72 Figure 4. 47: A failure mode due to tensile stress of first reinforced specimen (LPC1800AR) .................................................................................................... 73 xviii Figure 4. 48: Shear failure in second reinforced specimen (LPC1800BR) under cyclic loading ................................................................................................................ 73 Figure 4. 49: Shear failure of third reinforced specimen (LPC1800CR) under cyclic loading ................................................................................................................ 74 Figure 4. 50: Mortise tenon joint with screw ............................................................. 75 Figure 4. 51: Semi-rigid joint (a), moment-rotation curves (b) ................................. 76 Figure 4. 52: Dimensions of specimens for monotonic loading ................................ 77 Figure 4. 53: Test set up for monotonic loading ........................................................ 77 Figure 4. 54: The loading procedure of monotonic test (BS EN 26891:1991) .......... 78 Figure 4. 55: The effective centre of rotation for mortise-tenon joint (Hassan, 2008) ............................................................................................................................ 78 Figure 4. 56: Load-deformation curves of monotonic tests ....................................... 79 Figure 4. 57: The failure pattern of three specimens under monotonic loading ........ 80 Figure 4. 58: The places of CFRP in specimens ........................................................ 81 Figure 4. 59: The dimension of reinforced specimens for monotonic test (Dimensions are presented in mm) .................................................................... 81 Figure 4. 60: The selected epoxy and carbon fiber textile for reinforcement ............ 82 Figure 4. 61: The reinforced specimens for monotonic test....................................... 83 Figure 4. 62: Test set up for monotonic test on reinforced specimen ........................ 83 Figure 4. 63: Load-deformation curve of monotonic tests on reinforced specimen .. 85 Figure 4. 64: The failure pattern of three reinforced specimens under monotonic loading ................................................................................................................ 85 Figure 4. 65: The set up for cyclic test on unreinforced specimens ........................... 87 Figure 4. 66: The loading procedure of cyclic test (BS EN 12512:2001) .................. 87 Figure 4. 67: Load-deformation curve of cyclic tests on unreinforced specimens .... 88 Figure 4. 68: Complete load-deformation curves of specimens ................................ 89 Figure 4. 69: Plastic strain-cycles diagram of unreinforced specimens ..................... 90 Figure 4. 70: Degradation of modulus of elasticity in specimens .............................. 90 Figure 4. 71: Load-deformation graph within the range of elastic deformation (EN 408) .................................................................................................................... 91 Figure 4. 72: Failure pattern of three specimens under cyclic loading ...................... 92 Figure 4. 73: The set up for cyclic test on reinforced specimens ............................... 93 Figure 4. 74: Load-deformation curve of cyclic tests on reinforced specimens ........ 94 Figure 4. 75: Complete load-deformation curves of reinforced specimens ............... 95 Figure 4. 76: Plastic strain-cycles diagram of reinforced specimens ......................... 96 Figure 4. 77: Degradation of modulus of elasticity in reinforced specimens ............ 97 Figure 4. 78: The modes of failure for three reinforced specimens under cyclic loading ................................................................................................................ 97 Figure 5. 1: Linear elastic-plastic stress strain curve (Glos, 1981) 102 Figure 5. 2: Finite element model geometry ............................................................ 104 Figure 5. 3: Finite element mesh .............................................................................. 105 Figure 5. 4: SOLID 186 and SOLID 187 element types, respectively .................... 106 Figure 5. 5: Load displacement curves under monotonic loads ............................... 106 Figure 5. 6: Directional deformation ........................................................................ 107 Figure 5. 7: Normal stress distribution ..................................................................... 108 Figure 5. 8: Damage status in place of screws ......................................................... 108 Figure 5. 9: Damage status of screws ...................................................................... 109 Figure 5. 10: Finite element mesh of beam strengthened with CFRP composites .. 110 Figure 5. 11: Load displacement curves of the strengthening beams under monotonic loads ................................................................................................................. 111 xix Figure 5. 12: Normal stress distribution of the strengthening beam ........................ 112 Figure 5. 13: Damage status of the strengthening beam .......................................... 112 Figure 5. 14: The comparison of the unreinforced and reinforced specimen in FE analysis ............................................................................................................. 113 Figure 5. 15: Finite element model geometry .......................................................... 114 Figure 5. 16: Finite element mesh ............................................................................ 115 Figure 5. 17: Load-displacement curves under monotonic loads ............................ 116 Figure 5. 18: Directional deformation (Z axis) ........................................................ 117 Figure 5. 19: Normal stress distribution ................................................................... 118 Figure 5. 20: Damage status of timber ..................................................................... 118 Figure 5. 21: Damage status of screws..................................................................... 119 Figure 5. 22: Finite element model geometry of joint strengthened with CFRP ..... 120 Figure 5. 23: Finite element mesh of reinforced joint .............................................. 120 Figure 5. 24: Load-displacement curves of the strengthening joints under monotonic loads ................................................................................................................. 121 Figure 5. 25: Directional deformation (Z axis) of the strengthening joint ............... 122 Figure 5. 26: Normal stress distribution of the strengthening joint ......................... 123 Figure 5. 27: Damage status of the strengthening joint ........................................... 123 Figure 5. 28: Damage status of the tenon in tension zone ....................................... 124 Figure 5. 29: Damage status of CFRP ...................................................................... 124 Figure 5. 30: Damage status of screws..................................................................... 125 Figure 5. 31: The comparison of the unreinforced and reinforced specimen in FE analysis ............................................................................................................. 126 Figure 6. 1: The frame configuration for analysis ................................................... 127 Figure 6. 2: The non-linear curve of masonry ......................................................... 128 Figure 6. 3: The translational spring of lap joint ...................................................... 129 Figure 6. 4: The rotational spring of mortise-tenon joint ......................................... 130 Figure 6. 5: Load cases in the frame ........................................................................ 131 Figure 6. 6: Normal forces of unreinforced frame ................................................... 132 Figure 6. 7: Normal forces of reinforced frame ....................................................... 133 Figure 6. 8: Shear forces of unreinforced frame ...................................................... 133 Figure 6. 9: Shear forces of reinforced frame .......................................................... 134 Figure 6. 10: Bending moment of unreinforced frame ............................................ 134 Figure 6. 11: Bending moment of reinforced frame ................................................ 135 Figure 6. 12: Global deformation of unreinforced frame ......................................... 135 Figure 6. 13: Global deformation of reinforced frame ............................................. 136 Figure 6. 14: Global deformation-x direction of unreinforced frame ...................... 136 Figure 6. 15: Global deformation-x direction of reinforced frame .......................... 137 Figure 6. 16: Normal stresses of timber in unreinforced frame ............................... 137 Figure 6. 17: Normal stresses of timber in reinforced frame ................................... 138 Figure 6. 18: Shear stresses of timber in unreinforced frame .................................. 138 Figure 6. 19: Shear stresses of timber in reinforced frame ...................................... 139 Figure 6. 20: Normal stresses at masonry surfaces in unreinforced frame (sigma-x) .......................................................................................................................... 139 Figure 6. 21: Normal stresses at masonry surfaces in reinforced frame (sigma-x) .. 140 Figure 6. 22: Elastic strains at masonry surfaces in unreinforced frame (eps-x) ..... 140 Figure 6. 23: Elastic strains at masonry surfaces in reinforced frame (eps-x) ......... 141 Figure 6. 24: The comprasion of global analysis results between unreinforced and reinforced frames ............................................................................................. 144 Figure 6. 25: The frame configuration without infill materials for analysis ............ 144 xx Figure 6. 26: The comprasion of global analysis results between unreinforced and reinforced frames without infill material ......................................................... 145 Figure 6. 27: Global deformation of unreinforced frame without infill material ..... 146 Figure 6. 28: Global deformation of reinforced frame without infill material ......... 146 Figure 6. 29: Normal forces of unreinforced frame without infill material ............. 147 Figure 6. 30: Normal forces of reinforced frame without infill material ................. 147 Figure 6. 31: Shear forces of unreinforced frame without infill material ................ 148 Figure 6. 32: Shear forces of reinforced frame without infill material .................... 148 Figure 6. 33: Bending moment of unreinforced frame without infill material ........ 149 Figure 6. 34: Bending moment of reinforced frame without infill material ............ 149 Figure 6. 35: Normal stresses of timber in unreinforced frame without infill material .......................................................................................................................... 150 Figure 6. 36: Normal stresses of timber in reinforced frame without infill material 150 Figure 6. 37: The comprasion of all global analysis results ..................................... 151 xxi xxii INFLUENCE OF JOINTS ON THE SEISMIC RESPOND OF TRADITIONAL TIMBER FRAMES IN TURKEY ABSTRACT Timber frame structures which constitute an important cultural heritage of many countries, are well known as efficient seismic resistant structures worldwide and are worth to be preserved. Hımış is one common traditional Turkish timber system, which consists of a simple timber frame filled with masonry (such as bricks, adobes or stones with mortar), and a masonry ground floor, built on continuous stone foundations. These buildings are usually located in seismic areas. This thesis aims to make a review of the structural performance of Hımış timber system under seismic loading, with specific emphasis on joints and following strengthening of joints with CFRP (Carbon fiber reinforced polymer). Due to the seismic demands these timber structures mostly depend on connections, so that the joints have to be evaluated accurately in terms of translational and rotational stiffness and moment resistance. Subsequently, a series of experimental tests on two different types of timber joints (lap joint and mortise-tenon) which are common in Turkish timber structures have been carried out under monotonic and cyclic bending loading. The numerical analysis, FEM (the finite element method) has been performed in order to the calibrate the results from experiments. Finally, a numerical analysis considering semi-rigid joints in traditional timber connections has been performed globally. Keywords: Hımış timber frame, Traditional connections, Semi-rigid joints, Stiffness, Numerical analysis, Carbon fiber reinforced polymer. xxiii INFLUENCIA DE LAS JUNTAS EN LA RESPUESTA SÍSMICA DE ESTRUCTURAS DE MADERA TRADICIONALES EN TURQUIA RESUMEN Las estructuras de madera constituyen un importante patrimonio cultural en muchos países y son conocidas como tipologías estructurales eficientes des del punto de vista sísmico por su resistencia y ductilidad. Himis es uno de los sistemas de madera tradicional más conocidos en Turquía, consistente en marcos en forma de retícula de madera simple, relleno de mampostería en su interior (ladrillos, adobe, o piedras con mortero) en plantas piso, mientras que la planta baja es totalmente de mampostería sobre de una cimentación contínua de piedra. Esta tesis tiene como principal objetivo el de realizar una evaluación de la eficiencia estructural del sistema Himis sometido a cargas sísmicas, mediante la evaluación de la respuesta de las uniones más típicas de madera y su contribución al conjunto. Esta tesis se complementa con un análisis experimental y numérico de la contribución que aporta el refuerzo de dichas uniones mediante fibras de polímeros reforzados CFRP (polímeros reforzados con fibras de carbono). La respuesta de estas estructuras frente a sismo se debe, en parte, a la rigidez y ductilidad de las uniones entre barras de madera, por lo que éstas requieren de una especial atención en términos de rigidez traslacional y rotacional. La investigación lleva a cabo una campaña experimental, cubriendo dos tipologías básicas de uniones tradicionales entre barras de madera en Turquía (junta por solape y de espiga) bajo flexión monotónica y también cíclica. Paralelamente, se evalúan dichas uniones mediante un análisis FEM debidamente calibrado con los resultados de los ensayos experimentales, que permite reproducir el comportamiento global de estas estructuras a partir del grado de rigidez de las uniones con o sin refuerzo. Palabras Clave: Hımış Estructuras de madera, Uniones tradicionales, Uniones semirrígidas, Rigidez, Análisis numérico, Polímero reforzado con fibra de carbono. 1 1. INTRODUCTION 1.1 General Timber is one of the most frequently used building materials in both traditional and modern engineering constructions. Thanks to the low weight and high load-bearing capacity, timber frame buildings can stand horizontal forces imposed during earthquakes and thus are well suited for seismic zones. Traditional timber frame structures are characterized by a timber frame filled with an infill which is mainly masonry acting as shear walls. The masonry behaves very well under compressive stresses while wood elements act as ties resisting to tensile stresses. Timber framed constructions spread not only throughout European countries, such as Portugal (edificios pombalinos), Italy (casa baraccata), Germany (fachwerk), Greece (ksilopikti tichopiia), France (colombage), Scandinavia (bindingverk), United Kingdom (half-timber), Spain (entramados) etc., but also in India (dhaji-dewari), Turkey (hımış and bagdadi), Peru (quincha), USA (balloon frame in Chicago), Haiti (gingerbread houses) (Poletti 2014). Since this type of construction spread through out the world, it is important to point out the similarities and differences between them in order to better understand their performance. Even the evident difference in traditional construction worldwide, there is always a common idea in all them: timber resists tension, so that timber frames show better performance under seismic loading rather than masonry and provide ductility. Timber frame structures are particularly common in seismic regions, like Portugal. The adoption of timber as a structural material spread after the destruction of Lisbon derived from the strong earthquake in 1755. The typology of buildings that appeared after earthquake, called Pombalino buildings, were characterized by original external masonry walls and an adscititious internal timber structure named gaiola (cage), which is a three dimensional braced timber structure (Figure 1.1). The gaiola is formed by horizontal, vertical elements (sectional dimensions usually are 12x10, 12x15, 14x10, 10x10 cm), and diagonal (bracing) members (10x10, 10x8 cm). Timber framed walls are filled with rubble or alternatively with brick masonry or mud. The framed walls of the gaiola may have different geometries in terms of infill materials and number of timber elements (Figure 1.1a). Timber elements are connected together through various traditional joints. The most common traditional joints are: Mortise-tenon, dovetail, half- 2 lap and mitered half-lap joints (Figure 1.1b). Besides, the posts between two floors which are not continous, are usually connected to horizontal beams through a scarf joint in order to guarantee a certain continuity. Besides, once timber elements are fitted together, in order to guarantee a proper and permanent connection, a forged iron nail is usually hammered in almost every notch or interconnection (Nunes, 2017). A serious inconvenient shown by Pombalino buildings is that under seismic actions it is inevitable that heavy masonry of the façades fell down, although the timber skeleton remains almost intact, assuring the resistance of the timber floors and keeping the building in service. (a) (b) Figure 1. 1: The gaiola system (a), dovetail, half-lap, mitered half-lap joints (respectively) (b) Timber structural buildings spread widely across Europe, not only in seismic regions but also in non-seismic regions of Northern-European countries (Germany, Scandinavia) due to the easy availability and abundance of the material. Depending on the region in Germany, there are varied examples of timber framed structures (Figure 1.2a). Unlike in other countries, inclined posts are typical in German timber traditional houses. Furthermore, great variety of joints and roofs are used in traditional construction of Germany. For instance, mortise-tenon, overlapping, halving joints are used for column-beam connection, while cross-cut lap joint is used for beam-beam connection (Figure 1.2b). Dowel and pin are generally used as connectors. 9 the posts are placed every 100-140 cm. Studs are usually varied in intervals depending on infill materials. Horizontal members including the tie beam, windowsill, and knee braces are inserted between the studs in order to support the timber frame and to maintain the infill material in place. Cantilever is one of the peculiar features of the Turkish traditional timber house. Structural elements forming cantilevers are built together with the upper floor, as an extension of the joists. They are generally supported by bracings. (a) (b) (c) Figure 2. 3:Traditional Turkish wooden slabs. Double plates with one-way (a), single slab (b) and two-way slabs (c) using double plates The shape of the roof system is selected after completing the whole timber frame system. Hipped roof system with 4-side slope is usually preferred. Rafters are extended 50 to 60 cm outwards, in order to form the eaves. Tiles are generally used for roofing, by overlapping one tile on another. 2.2 Traditional Turkish Timber Joints Joints constitute usually a significant part of these structures. Many variables, such as the way of loading of timber elements, the type of fasteners, the existence of knots or even moisture content have a direct influence on timber joint design. As summarized in Table 2.1, several sources in the existing literature classify timber joints based on different criteria (Erman, 2011). From a structural point of view, joints are classified according to the type of acting forces (shear, compression, tension and bending). However, joints can also be classified depending on the type of fastener: bolted, doweled, nailed, plate components or glued connections. 10 Another important classification comes from the geometry of the components and their location within the structure. This is reason why this classification is made under constructional criteria and carpentry production. The constructional approach emphasizes direction of the components in joints, according to their grain, such as lengthening, framing, right angle (orthogonal) and diagonal joints. With a carpentry approach, joints can be classified depending on their geometry, such as plain (butt), lap and notch joints (Graham, 1951). The two classifications are regarded to be complementary. At the same time, the relative location of the joints in the structure should be also considered: right angle (orthogonal) joints, in the horizontal and the vertical plane, and diagonal joints. The classification of timber joints is complex. To summarize, a list of basic orthogonal timber joints would be: butt, cog, comb, dovetail, finger, fork, gooseneck, half lap, housed, lap, mortise-and-tenon, notch, oblique tenon, scarf and shoulder joints. Diagonal timber joints are usually used for sloped roof planes, trusses or bracings of wall and floor framings and being mainly known as bird’s mouth, bridle, butt, dovetail, lap, half lap, notch (front, back, double and tabled notch) joints, oblique tenon and step-lapped joints. Table 2. 1: Classification of timber joints (Erman, 2011) Sources Structural Approach Constructional Approach Carpentry Approach Binan 1990 –– a) End joints b) Corner joints c) Diagonal joints –– Bolshakov 1967 a) Built-up joints b) Scarf joints c) Multiple joints –– –– Götz 1989 a) Traditional joints b) Shear joints c) Dowel joints d) Nailed joints d) Glued joints –– –– Graham 1951 –– –– a) Plain joints b) Lap joints Güngör 1961 –– a) Lengthening joints b) Framing joints –– Günsoy 1967 a) Direct joints, –– –– 11 b) Connectored joints Karlsen 1989 a) Contact surface joints, b) Connectored joints. c) Glued joints –– –– Lloyd 1960 –– –– a) Notch joints b) Keyed joints c) Doweled joints d) Glued joints Kliment 1989 –– a) Right angle b) End joints c) Edge joints –– Mettem 1974 a) End joints b) Node joints c) Framing joints –– –– Schodeck 1980 a) Butt joints b) Lap joints c) Intersecting joints –– –– Ulrey 1970 –– –– a) Plain joints b) Lap joints Wood Reference Handbook 1991 a) Interlocking joints b) Fastener joints –– –– Among all these types of joints, the mortise-tenon, the preferred joint for beam-column connections (a) in Turkish traditional timber house. Another common typology is, the lap or half-lap joint (b), which is suitable for beam-beam connections (Figure 2.4-2.5). The connection between braces (diagonal) to the bottom and (angle brace of) cantilevers to the beams is achieved using tongued-grooved (c) and notch joints (d) (Figure 2.6), (Oztank, 2008). Nailed connections have been widely used in Turkish traditional timber frame structures. Nailed connections provide the ability to absorb and dissipate energy during severe earthquakes (Dogangun et al., 2015). 12 Figure 2. 4: Timber joints in frame (Turkish Timber Association, 2018) (a) (b) Figure 2. 5: Timber joints, mortise-tenon (a) and half-lap (b) joints Figure 2. 6: Timber joints, tongued-grooved (c) and notch (d) joints 13 2.3 Seismic Behaviour of Timber Structure Half-timbered structures (timber frame walls filled with masonry) are well known for their ductility-like behaviour. Several researches have been made on the timber frame structures. An important issue is to study these structures under cyclic and dynamic loading in order to understand their seismic behaviour. Previous studies show that the force-deformation response of a timber frame shear wall under cyclic or reverse loading behaves nonlinearly even at low loading levels (Pang et al., 2007). Furthermore, timber shear walls are capable of dissipating a large amount of energy through the behaviour of individual fasteners (Dinehart, 1998). The behaviour is also influenced by vertical load, as it increases the lateral stiffness and the energy dissipation, nail spacing and hold-down anchors (Johnston et al., 2006). It is commonly accepted that the load-deformation behaviour the absorption of energy of shear walls is mainly provided by joints. Therefore, the behaviour of connector under both monotonic and cyclic loading conditions related to shear walls has been investigated extensively by many researchers (Lam et al. 1997). Connections in timber frame construction are a key issue, as they control inplane behavior, particularly regarding to dissipative capacity of the timber walls. As there are very different types of connections, it is expected different dissipative behaviors. This fact justifies the extense experimental research work that has been carried out in the last years with different timber frame systems (Lukic et al., 2018). Furthermore, static cyclic tests have been performed on traditional timber framed walls, where all connections are half-lap connections, in order to study the seismic capacity in terms of strength, stiffness, ductility and energy dissipation (Poletti, 2014). It can be concluded that the predominant resisting mechanism provided by the infill is rocking, particularly in case of lower vertical pre-compression level. When the wall is excited, it may achieve large displacements without a significant loss of strength and, therefore with low damage. In case of timber frame walls with masonry infill, the fact of detaching masonry from the timber frame was evident due to the seismic excitement. During the test, masonry infill tended to move out-of-plane. Also, vertical posts and diagonals were clearly uplifting and the nail placed in the half-lap connections offered little resistance to the tearing force provided by the post (Figure 2.6a). Connections at the bottom tended to open out-of-plane, so that the post would come out, as the plastic deformation of the nail would impede the post to re-enter in its original position during unloading. Besides, notice that for masonry infill walls, most of the damage was concentrated in the lower part of the wall (Figure 2.6b), (Poletti, 2014). 14 (a) (b) Figure 2. 7: The beaviour of timber wall during test (a) and crack pattern in masonry infill (b) (Poletti, 2014) After a general review of the timber frame structure, especially in order to describe the seismic behavior of traditional Turkish timber houses (hımış), the different ways of damage affecting timber joints have been identified. The most relevant modes of collapse have been classified according to joints location within the structure (Table 2.2). Damage of timber joints during an earthquake usually leads to partial collapse of the building; This is due to their crucial role in the integrity of the entire building. Also, as stated previously, timber connections in traditional Turkish structures are always complemented with nails to provide resistance against tensile or shear forces. Nails contribute to provide ductility and the ability of dissipating energy, especially if they are "semi-rigid" through nails or other metal elements, instead of being “perfectly rigid” (Palma, 2012). Connections within the roof, floors, wall frames and bracing elements are not especially rigid in Turkish traditional timber houses, so during a seismic event, the failure of the infill usually leads to the collapse of joints by separating structural members. In case that the failure comes from a column or bracing members, large lateral displacements may lead to partial collapse. 15 Table 2. 2: Classification of joint damages in traditional Turkish house Location Mode of collapse Images Excess of movement at top Tension splitting of nailed joints due to collapse of the heavy roof. (Korkmaz et al., 2010). Excess of tension Failure of specific joints due to pull out of nailed connections. (Dogangun et al., 2015). Failure of infill material Failure of diagonal joints due to the failure of the material infill. (Aksoy et al., 2005). Excess of global base shear Base shear is enhanced by the weight of infill. (Dogangun et al., 2015). 16 2.4 Strengthening of Timber Structures with FRP Fiber reinforced polymer materials formed by high strength fibers and a resin matrix have a wide variety of industrial applications due to their high strength-to-weight ratio and ease of handling. FRP materials are composites comprising fibers that provide the load-bearing capacity and stiffness, embedded in a polymeric resin that transfers loads between the fibers and provides protection to the fibers. They are available in a wide variety of forms, and have properties that vary considerably depending on the fibre material, volume fraction and orientation. Typical properties of the common fibers and polymers are given in Table 2.3. For structural reinforcement, two main forms of FRP are generally used, namely, pultruded rods or plates and fabrics. For internal reinforcement, pultruded rods and plates are bonded into slots or grooves formed in the timber element. For external reinforcement, FRP plates or fabric materials are used. The reinforcement of timber with FRP is normally implemented by adhesive bonding. An important aspect of the behaviour of the composite material is the bond between wood and the fiber reinforced plastic. Besides, mechanical properties of FRP strongly depend on the fiber content in each direction and on the fiber itself. Unidirectional fiber-reinforced polymers (FRP) are highly orthotropic. Table 2. 3: Fiber and polymer properties (Schober, 2015). Material Modulus of elasticity Tensile strength Failure strain Density (GPa) (MPa) (%) (g/cm3) E-glass 70-80 2000-4800 3.5-4.5 2.5-2.6 Carbon (HM) 390-760 2400-3400 0.5-0.8 1.85-1.90 Carbon (HS) 240-280 4100-5100 1.6-1.73 1.75 Aramid 62-180 3600-3800 1.9-5.5 1.44-1.47 Basalt 82-110 860-3450 5.5 1.52-2.7 Polymer 2.7-3.6 40-82 1.4-5.2 1.10-1.25 (HM: High modulus, HS: High tensile strength) In timber beams subjected to bending, the predominant failure occurs due to tensile stress, frequently by failure at the lower beam side. FRP have a linear elastic behavior until the yield stress, showing excellent mechanical properties, with high elasticity module and tensile strength values, in comparison to the weight and volume (Garcia, 2016). Research studies have progressively increased in order to expand the knowledge on this matter. These studies on 17 reinforcement of timber aiming to improve behaviour with respect to flexural capacity, stiffness, and ductility yielded results in quite a wide range. Several research projects and applications are summarized in Table 2.4. Buell and Saadatmanesh tested the beams, reinforced by wrapping CFRP (Carbon fiber reinforced polymer) fabrics under bending. The enhancement of bending strength of these wrapped beams, based on one single control, was between 40 and 53%. The stiffness increased from 17 to 27% (Buell and Saadatmanesh, 2005). Blass et al. investigated the influence of FRP reinforcement on bending stiffness and load bearing capability of glulam, testing different reinforcement layouts, qualities of timber and adhesives. The CFRP lamellas were bonded at the bottom of some specimens, and in another set, CFRP-lamellas were vertically slotted into the bottom part of the timber beams. They concluded that the beams with slotted in CRFP lamellas showed linear elastic behaviour nearly up to the loadbearing capacity (Blass et al., 2002). Borri, Corradi and Grazini investigated timber beams that were reinforced using CFRP fabrics or bars. Compared with the control beams, these reinforced beams showed an increase of about 30% in stiffness up to 60% for the CFRP fabric layers reinforced beams. The beams with slotted-in CFRP bars all had a lower stiffness and capacity than the ones reinforced using fabrics (Borri et al., 2005). Besides, Issa Camille also covered the glulam wood beams with CFRP fabrics. Obtained results indicate that the behavior of reinforced beams is totally different from the un-reinforced ones. The reinforcement changed the mode of failure from brittle to ductile and increased the loadcarrying capacity of specimens (Issa Camille, 2005). Gezer and Aydemir observed the strength ratio of the wrapped and non-wrapped wood material with CFRP. Compression and bending strength of the specially wrapped wood materials was investigated. At the same time, two types of woods were compared in terms of strength ratios. As a result of this study, the increment of compression and three-point bending strengths were determined for wrapped CFRP wood materials. Bending tests showed that samples exhibited an improvement of 65% in smaller cross sections and 15% in larger cross sections. Also, the variation of elastic modulus was analyzed, it was seen that the CFRP material leads to an increase in elastic modulus (E) of the material for both tree types. The wrapped specimen became a very rigid structure under bending compared to the non-wrapped specimen (Gezer, 2010). Furthermore, a series of different experiments were conducted on timber beams reinforced with different amounts of CFRP sheets (carbon fiber reinforced plastic) followed by a statistical 18 analysis. A moderate increase of load-bearing capacity and ductility (approx. 30%) and a small increase of elastic stiffness (approx. 16%) may be achieved (Andor et al., 2015). Also, another study showed that the externally strengthening systems of pine timber LVL beams with different grammage basalt and carbon FRP gave rise to structures having higher stiffness and carrying capacity than the initial ones. However, the ultimate displacement experienced was not increased in the reinforced beams. By comparing the three unidirectional fabrics, the best results of ultimate load were obtained with FB280 (280 g/m2 basalt), followed by FB600 (600 g/m2 basalt), and finally FC300 (300 g/m2 carbon), while the ultimate stress of FC300 was higher than that of FB280 and FC210 (De la Rosa Garcia et al., 2013). Another research investigated an experimental programme based on strengthening laminated wood beams by using two different types of FRP -carbon fiber reinforced polymer (CFRP) and glass fiber reinforced polymer (GFRP) compositesheets. The results of the study are encouraging with an increment of flexural stiffness up to 45.76% for 5% addition of GFRP composite sheet to the tension side of the beam. For the same percentage each of GFRP the flexural strength increase was 40%, compared to the same unstrengthened beam. For 3.33% addition of CFRP composite sheet to the tension side of the beam the increment in flexural stiffness was 64.12%. The gain in flexural strength for the corresponding percentage addition of CFRP was 50.62%. Thus, carbon fiber reinforced polymer increased the flexural stiffness and strength of timber beams more than glass fiber reinforced polymer (Nadir et al., 2016). Besides, flexural behaviour of wood beams strengthened with hybrid FRP (HFRP) was carried out by Yang et al., 2013. Strengthening technique consists of adding carbon fiber (CF), highstrength glass fiber (SGF), hybrid CF/glass fiber, and hybrid CF/SGF to elements. Test results indicated that hybrid CF/SGF strengthening showed better ductility and strength compared with other strengthening schemes. When it comes to shear reinforcement, some authors have analyzed the behavior of reinforced beams to shear stress through sheets arranged transversally and longitudinally to the direction of the wood fiber on the lateral beam sides (Greenland et al., 1999). Another form of shear reinforcement has been carried out with FRP pultruded rods embedded in epoxy resin into holes in the lower beam face (Radford, 2002). This application of the reinforcement is intended to diminish the possible early failure to shear effect that the drying splits may cause on beams subjected to bending. In joint scale, Silva et al. conducted four-point bending tests on timber lap joints with the CFRP strengthening techniques of near-surface mounted (NSM) and externally bonded (EBR) reinforcement which was sufficient to get the CFRP strain distribution, shear stress distribution 25 Figure 3. 4: Acceleration Timber members have been modelled by using bar elements and the masonry infill has been modelled with 2D planer elements. The masonry base has been chosen as a fixed support (Figure 3.5). Two load cases have been defined: dead loads and live loads. Dead loads come from permanent construction material loads comprising the roof, floor, wall, and masonry. Live loads come from use and occupancy of a building. These loads have been calculated as; Dead Load = 2 kN/m2 (assumed) x 4 m width of floor x 2 floors = 16 kN/m Live Load = 2 kN/m2 (assumed) x 4 m width of floor= 8 kN/m (Figure 3.6-3.7). Figure 3. 5: General model of frame Z XY In Y-direction RF-DYNAM Pro Max u: 1.00000, Min u: 0.00000 - 26 Figure 3. 6: Definition of dead loads Figure 3. 7: Definition of live loads 3.3 Analysis of Results After solving a dynamic non-linear analysis, the envelope of normal and shear forces and bending moment, together with the elastic deformed shape, is shown in Figure 3.8-3.10. 8.000 8.000 X 8.000 8.000 8.000 8.000 8.000 Z Y 8.000 8.000 8.000 In Y-direction LC 1: Dead Loads Loads [kN/m] 8.000 Y 8.000 8.000 Z 8.000 X In Y-direction LC 3: Live Loads Loads [kN/m] 27 Figure 3. 8: Normal forces Figure 3. 9: Shear forces 2.925 1.394 5.068 0.393 0.697 3.494 1.073 5.672 0.160 0.411 0.099 5.950 1.611 0.332 0.2251.778 3.140 1.132 0.109 0.171 0.818 2.882 3.598 7.125 0.444 4.120 4.477 0.285 3.378 0.139 0.349 1.216 3.714 0.305 3.298 0.288 0.011 0.054 0.366 1.247 0.328 3.057 4.616 1.059 0.190 X 0.857 Z Y 4.365 0.976 0.221 4.843 0.013 0.523 1.073 0.164 0.057 1.997 1.484 0.053 5.189 0.762 0.397 0.414 0.049 4.846 0.320 0.149 3.858 2.681 In Y-direction RC 5: DLC2 - Result Envelope Members Internal Forces N Result Combinations: Max Values Max N: 7.125, Min N: 0.007 kN 0.030 0.361 0.271 0.1350.071 0.156 0.074 0.010 0.004 0.002 1.081 0.007 0.401 1.290 0.022 0.115 1.622 0.182 0.499 0.235 0.068 0.233 0.785 0.036 0.003 0.860 0.019 0.008 0.012 1.649 0.190 0.058 0.264 0.009 1.028 0.091 0.096 0.054 0.022 0.179 0.065 0.135 0.029 0.072 0.502 0.046 0.105 0.712 0.020 0.339 0.954 0.113 0.068 0.173 0.066 X Z Y 0.150 0.248 0.006 0.245 0.652 0.196 0.050 1.067 0.102 0.096 0.010 0.061 0.012 0.056 0.117 0.002 0.152 0.0110.074 0.113 0.546 0.038 0.010 0.026 0.854 0.096 In Y-direction RC 5: DLC2 - Result Envelope Members Internal Forces V-z Result Combinations: Max Values Max V-z: 1.649, Min V-z: 0.002 kN 28 Figure 3. 10: Bending moment Figure 3. 11: Global deformations 0.0050.040 0.042 0.010 0.014 0.011 0.012 0.066 0.007 0.005 0.001 0.009 0.101 0.005 0.003 0.030 0.019 0.129 0.004 0.037 0.002 0.008 0.009 0.006 0.005 0.118 0.039 0.001 0.093 0.003 0.011 0.009 0.004 0.082 0.029 0.014 0.002 0.003 0.029 0.046 0.011 0.009 0.009 0.008 0.098 0.099 0.008 0.052 0.080 0.063 0.031 X 0.002 Z Y 0.044 0.005 0.011 0.021 0.080 0.114 0.002 0.006 0.006 0.002 0.008 0.001 0.010 0.001 0.037 0.074 0.008 0.001 0.083 0.005 0.007 0.004 0.003 0.034 0.008 In Y-direction RC 5: DLC2 - Result Envelope Members Internal Forces M-y Result Combinations: Max Values Max M-y: 0.129, Min M-y: 0.000 kNm X 0.3 Z Y Global Deformations |u| [mm] 0.3 0.2 0.2 0.2 0.2 0.1 0.1 0.1 0.1 0.0 0.0 0.0 Max : 0.3 Min : 0.0 In Y-direction RC 5: DLC2 - Result Envelope Global Deformations u Result Combinations: Max Values Factor of deformations: 920.00 Max u: 0.3, Min u: 0.0 mm 29 Results from analysis have proved that the acceleration dependence of connections significantly affects the global response of the system, since they represent important dissipative elements in the frame. Connections have been considered rigid in order to reproduce an ideal behaviour. Joints are crucial during a seismic episode, particularly when infill is not present at opening parts. A first review of results shows that the Hımış timber frame is significantly rigid and low deformable, with deformations of 0.3 mm (Figure 3.11). The upper part suffers the highest deformation, where the force is applied in the opposite direction. In general, the frame behaves in a rocking mode on further increasing accelaration in the lateral drift demand whereby the lateral capacity has been largely dependent on tension capacity of vertical posts. The upper beam of opening parts (window, door) are subjected to the maximum compressive forces, as 7.12 kN. Moreover, different behavior is observed at the braces depending on the location. The diagonal element inclined against the applied displacement is under compression while the another one under tension. This diagonal brace (at the right) suffers from pure tension (4 kN) leading to physical separation of elements. Finally, during severe cycles, lower parts of the frame tend to slide off the masonry basement due to heavy shear forces, approximately 1.64 kN. Besides, these nodes allow certain rotation of the post, moment is about 0.052 kN.m. Highlighted areas in Figure 3.10 show higher bending moments in the frame. Maximum bending moments, like 0.129 kN.m can be observed on the window and door openings, where the lap joints are located, thus these parts show the highest bending moments. Furthermore, another high bending moment is seen at the lower part of frame, where the column and beam are usually connected with mortise-tenon joints. Thereafter, the whole structure deforms significantly. Depending on the obtained results of the analysis, internal forces of members and transferring of loads to each node have been detected, thereby the efficiency of the structural connections can be evaluated. Moreover, the distribution of stresses in sigma x and sigma y on masonry surfaces are given in Figure 3.12-3.13. It is evident that the value of the tensile municipal stresses are over the above limit slightly at the corner of the openings, the limit tension of masonry was defined as 0.1 MPa. 30 Figure 3. 12: Axial stresses in sigma-x at masonry surface Figure 3. 13: Axial stresses in sigma-y at masonry surface Z XY Axial Stresses x,+ [kN/cm2] 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.00 0.00 0.00 Max : 0.02 Min : 0.00 In Y-direction RC 5: DLC2 - Result Envelope Surfaces Stresses Sigma-x,+ Result Combinations: Max Values Max Sigma-x,+: 0.02, Min Sigma-x,+: 0.00 kN/cm2 Z XY Axial Stresses y,+ [kN/cm2] 0.04 0.03 0.03 0.03 0.02 0.02 0.02 0.01 0.01 0.01 0.00 0.00 Max : 0.04 Min : 0.00 In Y-direction RC 5: DLC2 - Result Envelope Surfaces Stresses Sigma-y,+ Result Combinations: Max Values Max Sigma-y,+: 0.04, Min Sigma-y,+: 0.00 kN/cm2 31 3.4 Response Spectrum Analysis In order to indicate the likely maximum seismic response of an essentially elastic structure, a respond spectrum analysis has been carried out. Response spectrum analysis provides insight into dynamic behaviour of structure. It was carried out with contribution from each natural modes. Modes are inherent properties of a specific structure, and are determined by material properties (mass, damping, and stiffness) and boundary conditions. Each mode is defined by a natural (modal or resonant) frequency, modal damping, and a mode shape. The natural and angular frequencies ω and f as well as the periods T are listed in table 3.4. Table 3. 4: The parameters of mode shapes Mode Eigenvalue Angular Frequency Natural frequency Natural period No. v [1/s2]  [rad/s] f [Hz] T [s] 1 11247.050 106.052 16.879 0.059 2 94582.086 307.542 48.947 0.020 3 180791.094 425.195 67.672 0.015 4 243747.516 493.708 78.576 0.013 5 329508.000 574.028 91.359 0.011 6 416756.125 645.567 102.745 0.010 The assigned response spectrum (derived from 1999 Kocaeli Earthquake) is illustrated in the graphic (Figure 3.14). Figure 3. 14: Acceleration spectrum The frame has been modelled two-dimensional and the acceleration has been recorded at each node of the structure. Six modes in lateral direction have been analyzed (Figure 3.15-3.17). Figure 3.18 that the first mode correspond to large displacements associated with strong excitations, shows 0.8 mm. Considering the peak, period of 0.059 seconds is reached in the first mode of the structure. Internal forces and the distribution of stresses in masonry surfaces are 32 given in Figure 3.19-3.23. Resulting values are higher than the results from dynamic non-linear time history analysis. Figure 3. 15: Mode shape 1 and mode shape 2 Figure 3. 16: Mode Shape 3 and mode shape 4 Figure 3. 17: Mode Shape 5 and mode shape 6 X 1.00000 Z Y Natural vibration u [-] 1.00000 0.90909 0.81818 0.72727 0.63636 0.54545 0.45455 0.36364 0.27273 0.18182 0.09091 0.00000 Max : 1.00000 Min : 0.00000 In Y-direction RF-DYNAM Pro Natural vibration u Mode shape No. 1 - 16.879 Hz Factor of deformations: 0.52 Max u: 1.00000, Min u: 0.00000 - X 1.00000 Z Y Natural vibration u [-] 1.00000 0.90909 0.81818 0.72727 0.63636 0.54545 0.45455 0.36364 0.27273 0.18182 0.09091 0.00000 Max : 1.00000 Min : 0.00000 In Y-direction RF-DYNAM Pro Natural vibration u Mode shape No. 2 - 48.947 Hz Factor of deformations: 0.52 Max u: 1.00000, Min u: 0.00000 - X 1.00000 Z Y Natural vibration u [-] 1.00000 0.90909 0.81818 0.72727 0.63636 0.54545 0.45455 0.36364 0.27273 0.18182 0.09091 0.00000 Max : 1.00000 Min : 0.00000 In Y-direction RF-DYNAM Pro Natural vibration u Mode shape No. 3 - 67.672 Hz Factor of deformations: 0.52 Max u: 1.00000, Min u: 0.00000 - X 1.00000 Z Y Natural vibration u [-] 1.00000 0.90909 0.81818 0.72727 0.63636 0.54545 0.45455 0.36364 0.27273 0.18182 0.09091 0.00000 Max : 1.00000 Min : 0.00000 In Y-direction RF-DYNAM Pro Natural vibration u Mode shape No. 4 - 78.576 Hz Factor of deformations: 0.52 Max u: 1.00000, Min u: 0.00000 - X 1.00000 Z Y Natural vibration u [-] 1.00000 0.90909 0.81818 0.72727 0.63636 0.54545 0.45455 0.36364 0.27273 0.18182 0.09091 0.00000 Max : 1.00000 Min : 0.00000 In Y-direction RF-DYNAM Pro Natural vibration u Mode shape No. 5 - 91.359 Hz Factor of deformations: 0.52 Max u: 1.00000, Min u: 0.00000 - X 1.00000 Z Y Natural vibration u [-] 1.00000 0.90909 0.81818 0.72727 0.63636 0.54545 0.45455 0.36364 0.27273 0.18182 0.09091 0.00000 Max : 1.00000 Min : 0.00000 In Y-direction RF-DYNAM Pro Natural vibration u Mode shape No. 6 - 102.745 Hz Factor of deformations: 0.52 Max u: 1.00000, Min u: 0.00000 - 33 Figure 3. 18: The deformation in mode shape 1 Figure 3. 19: Envelope of normal forces X 0.8 Z Y Global Deformations |u| [mm] 0.8 0.8 0.7 0.6 0.5 0.5 0.4 0.3 0.2 0.2 0.1 0.0 Max : 0.8 Min : 0.0 In Y-direction RC 6: DLC1, Result Envelope Global Deformations u Result Combinations: Max Values Factor of deformations: 300.00 Max u: 0.8, Min u: 0.0 mm 3.302 10.23611.315 7.024 10.135 10.792 15.915 0.682 13.983 8.383 21.109 19.386 1.339 3.592 3.898 0.932 3.924 11.135 20.058 9.681 10.308 20.361 16.917 9.455 17.279 4.582 15.710 1.948 8.489 17.929 10.144 12.450 0.698 2.555 16.305 4.368 15.519 21.565 5.897 11.535 15.17310.339 6.476 13.078 20.646 12.691 4.959 14.193 2.337 12.001 15.061 X 0.609 Z Y 10.934 9.693 14.115 3.850 9.087 1.056 1.710 4.288 5.931 6.513 19.313 20.002 2.765 7.783 7.531 18.759 5.885 12.01124.774 16.618 1.006 9.898 2.627 22.827 7.751 In Y-direction RC 6: DLC1, Result Envelope Members Internal Forces N Result Combinations: Max Values Max N: 24.774, Min N: 0.413 kN 34 Figure 3. 20: Envelope of shear forces Figure 3. 21: Bending moment 0.947 0.999 3.475 0.165 0.612 0.561 1.340 1.117 3.029 0.404 3.029 0.069 0.169 4.551 0.111 0.166 0.341 0.676 0.603 1.312 4.172 0.389 0.071 1.389 1.332 2.857 0.092 0.494 2.901 0.312 1.419 1.291 0.583 0.653 5.779 0.534 4.092 2.876 0.644 2.730 0.843 0.697 0.404 4.398 0.607 0.259 0.702 2.867 1.541 3.071 0.357 0.513 0.407 6.201 1.098 0.620 0.719 1.634 3.288 0.099 0.671 X Z Y 2.831 1.235 1.356 0.916 0.726 0.636 0.704 1.071 0.785 5.218 0.509 3.057 1.004 0.439 4.618 0.085 0.075 4.812 2.965 2.495 0.400 2.506 0.075 0.783 0.088 0.654 1.751 0.172 3.858 In Y-direction RC 6: DLC1, Result Envelope Members Internal Forces V-z Result Combinations: Max Values Max V-z: 6.201, Min V-z: 0.058 kN 0.114 0.049 0.190 0.119 0.298 0.024 0.019 0.502 0.210 0.267 0.223 0.353 0.034 0.034 0.050 0.043 0.494 0.027 0.063 0.029 0.007 0.063 0.205 0.026 0.299 0.102 0.122 0.1900.128 0.160 0.083 0.053 0.045 0.080 0.128 0.389 0.054 0.238 0.022 0.290 0.195 0.052 0.216 0.272 0.429 0.350 0.272 0.278 0.857 0.151 0.199 0.388 0.130 0.137 0.054 0.040 0.451 0.030 X Z Y 0.263 0.166 0.033 0.174 0.110 0.143 0.033 0.032 0.020 0.093 0.029 0.118 0.031 0.361 0.168 0.134 0.084 0.024 0.457 0.349 0.152 0.030 0.028 0.110 0.220 0.138 0.129 In Y-direction RC 6: DLC1, Result Envelope Members Internal Forces M-y Result Combinations: Max Values Max M-y: 0.857, Min M-y: 0.005 kNm 41 Figure 4. 4: The dimension of specimens for compression test parallel to the grain (Dimensions are presented in mm) Figure 4. 5: Three specimens for compression test parallel to the grain Experimental tests have been performed with an automatic press with maximum capacity of 300 kN and strain rate sensitivity about 0,01 mm. According to EN 408:2010, load was applied at a constant loading-head movement so that the maximum load is reached within (300±120) s. The load was applied continuously with constant velocity 3kN/min until failure (Figure 4.6). Load and deformation were recorded in a computer and later stored as excel files. 42 Figure 4. 6: Compression test parallel to the grain Load-deformation curve for compression parallel to the grain for three specimens are given in Figure 4.7. The stress-strain behavior under compression parallel to the grain is characterized by a decrease after reaching the ultimate load. At the ultimate load, the weakest cell starts to collapse followed by the adjacent cells to guarantee that crushing band occurs. The collapse, which is a stability failure of the cell walls, leads to the loss of the load capacity of the cell. The load drops down to a level between 40 and 50% of the ultimate load. On average, the load slightly decreases in a ductile manner followed by a significant softening. Specimens 1 and 3 show less load capacity than specimen 2 due to the existence of knot in timber (Figure 4.7). 43 Figure 4. 7: Load-deformation curve for compression test parallel to the grain In ASTM D143-14, compression failure patterns of wood were classified according to its shape as shown in Fig 4.8. The results of failure modes were evaluated with this standard. a b c d e f Figure 4. 8: Compression failure patterns, a) Crushing, b) Wedge split, c) Shearing d) Splitting, e) Compression and shear parallel grain, f) Brooming or end-rolling, (ASTM D143-14). After tests, the failure patterns were examined based on the classification of failure patterns in standard ASTM D143-14. In the first and third specimens, ´crushing and oblique shearing´ can be detected, while in the second specimen ´crushing´ at end local is clearly seen. When the plane of rupture is horizontal, crushing occurred. Besides this, because of the initial eccentricity the plane rupture makes an angle of aprx. 450, the oblique shear formed in the middle of specimens-1 and 3 (Figure 4.9a-c). At the end of specimen-2 within 30mm, end local pressure 44 occurred and wood fibers were instable due to compression and a transverse fold formed in the specimen surface (Figure 4.9b). a b c Figure 4. 9: The failure patterns of compression tests parallel to the grain In order to determine the compressive strength of timber, the equation 4.1 from EN 408:2010 was used. The average compressive strength of three specimens was calculated as 39 N/mm2. max ,0cF fA  (4.1) fc,0 compressive strength parallel to the grain, in newtons per square millimetre; A cross-sectional area, in square millimetres; Fmax maximum load, in newtons 2 ,0 123 1000 39 45 70 c kN f N mm mm    4.2.2. Compression perpendicular to the grain The dimensions of timber specimens have been considered according to Table 2 in standard EN 408:2010 (Figure 4.10). A total of 3 specimens with a cross section of 45x70 mm (bxl) and 90 mm height (h), have been subjected under compression test perpendicular to the grain (Figure 4.11). Before characterization, the moisture content of the three specimens has been accurately measured, approximately 10.5%, 10.3% and 9.5% respectively. 45 Figure 4. 10: The dimension of specimens for compression test perpendicular to the grain (Dimensions are presented in mm) Figure 4. 11: Three specimens for compression test perpendicular to the grain According to EN 408:2010, the load has been applied at a constant loading-head movement adjusted in order that maximum load is reached within (300±120) s. The load has been applied continuously with constant velocity 3kN/min until failure (Figure 4.12). When a load is perpendicularly applied to the cells (grains), the thin walled tubes are affected laterally and become squeezed together with the increase of compression stresses, this leads to the collapse. This behavior continues until all the fibers are fully crushed. When all fibers are crushed together it is possible to once again increase the loads and it is difficult to define a failure level. Load-deformation curve for compression perpendicular to the grain for three specimens are given in Figure 4.13. Timber is markedly ductile with a continual increment of load after yielding and an additional hardening after 45 to 55% deformation. Specimen 3 shows less load capacity than specimens 1 and 2, due to the existence of a knot in timber (Figure 4.13). 46 Figure 4. 12: Compression test perpendicular to the grain Figure 4. 13: Load-deformation curve for compression test perpendicular to the grain It is important to note that the compressive strength in the direction perpendicular to the grain is less than 10% of the strength in the direction parallel to the grain. After the tests, the failure patterns have been examined. In the first specimen, ´rolling shear´ can be detected, while in the second and third specimens, ´densification and buckling´ is 47 observed. When a load is applied perpendicular to annual rings, tend to buckle total by leading to rolling shear failure (Figure 4.14a) and also exhibits crushing of annual rings with corresponding densification (Figure 4.14b-c). a b c Figure 4. 14: The failure patterns of compression tests perpendicular to the grain In order to calculate the compressive strength of timber , the equation 4.2 from EN 408:2010 has been used. The average compressive strength perpendicular to the grain of three specimens has been calculated as 4,12 N/mm2 which is 9 times less than compressive strength parallel to the grain. max 90 F fA  (4.2) fc,90 compressive strength perpendicular to the grain, in newtons per square millimetre; A cross-sectional area, in square millimetres; Fmax maximum load, in newtons; 2 13 1000 90 4,12 45 70 kN f N mm mm    48 4.2.3. Bending To determine the local modulus of elasticity, global modulus of elasticity and static bending (flexural) strength of wood, four point bending tests have been carried out according to EN 408:2010. According to the standard, the specimen shall have a minimum length of 19 times the depth of the section. A total of 3 specimens were tested, with a cross section of 90 mm × 90 mm (b × h) and 1800 mm length (Figure 4.15). The specimen shall be symmetrically loaded under two bending points with a span of 18 times the depth as shown in configuration of experiment. Figure 4. 15: The dimension of specimens for bending tests (Dimensions are presented in mm) The specimen has been simply supported. Small steel plates of length not greater than one-half of the depth of the specimen have been inserted between the piece and the loading heads or supports to minimize local indentation (Figure 4.16). Figure 4. 16: Test arrangement for measuring local modulus of elasticity in bending (EN 408). 49 Previously, in order to determine the local modulus of elasticity, timber beams have been subjected to four point flexural loading, by using a 1000 kN displacement control hydraulic jack. Constant velocity of load application was imposed to 12 mm/min. Two Linear Variable Differential Transformers (LVDT) having a resolution of 0.1 mm, were used for monitoring the vertical deflections at the mid-span under the mid points of two side faces of the beam (Figure 4.17). The deformation (w) shall be taken as the average of measurements on both side faces at the neutral axis and shall be measured at the centre of a central gauge length of five times the depth of the section. According to EN 408:2010, to determine the local modulus of elasticity, the maximum load applied shall not exceed 0,4 Fmax. Besides, at the load/deformation graph within the range of elastic deformation, the section 0,1 Fmax and 0,4 Fmax is used for a regression analysis. In order to calculate the local modulus of elasticity, the equation 4.3 was used.     2 , 21 2116 ml FF EWW al I     (4.3) Em,1 local modulus of elasticity, in newtons per square millimetres; F2-F1 an increment load, in newtons on the regression line; W2-W1 increment of deformation, in mm corressponding to F2-F1; a distance between a loading position and the nearest support, in millimetres; l length (aprx. 5h) for the determination of modulus of elasticity, in millimetres; I moment of inertia, in millimetres to the fourth. The load-deformation graph has been obtained from test results (Figure 4.18). The regression line has been obtained between 0.1 Fmax- 0.4 Fmax loads and the deformations corresponding to them. The regression value was calculated with 988,96 N/mm from diagram, hereby the average local modulus of elasticity of three specimens has been calculated as 1236,2 N/mm2. 22 , 988,96 1236,2 540 450 16 5467500 ml E N mm    50 Figure 4. 17: The test set up for measuring local modulus of elasticity in bending Figure 4. 18: Load-deformation curve for the range of 0.1 Fmax-0.4 Fmax 57 4.3 Lap Joint Tests A set of monotonic tests of unreinforced specimens have been performed in order to describe the behaviour of lap joints which is used for beam-beam connections in Turkish timber house (The details are given in part 2.2). Subsequently, joints locally strengthened with carbon fiber textile have been tested under monotonic and cyclic loadings. The purpose of these tests is to increase the flexural strength and load-bearing capacity of the joint. Besides, carbon fiber textile may prevent the premature seperation of components of the joint under loading. 4.3.1. Monotonic tests on unreinforced specimens To determine the load-deformation behaviour of lap joint under monotonic loading, four point bending tests have been carried out according to BS EN 26891 (Timber structures- Joints made with mechanical fasteners- General principles for the determination of strength and deformation characteristics). A total of 2 lap joint specimens have been tested, by using the dimensions of 90 mm × 90 mm in cross-section (b × h) and 1800 mm length (Figure 4.27). Two screws with dimensions of Ø4.5 mmx h:80 mm have been used for connection introduced with an angle of 450 angles (Figure 4.27). In table 4.2, the specimens are abbreviated with codes. In the code, ´LP` shows the type of joint, such as; lap joint. Besides, ´M` and ´C` indicate monotonic and cyclic loads. Following, the three specimens are shown as ´A`, ´B` and ´C`. Last letter ´R` in the code indicates that CFRP reinforcement exists on the specimen. The specimens have been loaded under two point bending over a span of 18 times the depth according to the loading procedure in standard (Figure 4.28) Figure 4. 27: Dimensions of specimens in monotonic tests (Dimensions are presented in mm) 58 Table 4. 2: The codes of specimens for tests Codes Height (cm) Width (cm) Length (cm) Bending Strength (MPa) Reinforcement with CFRP LJM1800A 90 90 1800 72 - LJM1800B 90 90 1800 72 - LJM1800AR LJM1800BR LJC1800AR LJC1800BR LJC1800CR 90 90 90 90 90 90 90 90 90 90 1800 1800 1800 1800 1800 72 72 72 72 72 + + + + + Two LVDT with sensitivity of 0.1 mm, have been used for monitoring the vertical deflections at mid-span under the mid points of two side faces of the beam. The load has been distributed in 2 points with two cylinders of a diameter of Ø4 mm. At the same time, two metal semicylinders at the supports have been used. The loading procedure of test has been obtained from BS EN 26891:1991 (Figure 4.29). According to the standard, the load is applied up to 0,4 Fest and maintained for 30 s. Then the load is reduced to 0,1 Fest and maintained for 30 s. Thereafter, the load is increased until the ultimate load. The test is stopped when the ultimate load is reached. Following, the estimated maximum load, Fest, 1000 N has been taken on the basis of previous bending experiments. The load has been applied up to 0.4 Fest, which corresponds to 400 N, maintaned for 30 sec. Then, the load has been reduced to 0.1 Fest, which corresponds to 100 N, maintaned for 30 sec at this value. Thereafter, the load has been increased until the ultimate load (Figure 4.29). Constant velocity of load application has been imposed to 10 mm/min. 59 Figure 4. 28: The set up for monotonic test Figure 4. 29: The loading procedure of monotonic test (BS EN 26891:1991) 60 When the timber beam deforms under bending, the deformation has been caused by moment and changing geometry (Figure 4.30). In addition to tensile and compressive stresses, shear also take place. The deformation is a result of normal and shear stresses in the beam cross section. The load-deformation diagram has been obtained directly from test results (Figure 4.31). The curves which are presented in Figure 4.31 show a quite elastic behaviour until the maximum strength; this point when a slip occurs, there is loss of friction by inducing a rapid decrease of resistance. Thus, the brittle behaviour is followed by an inelastic phase. Finally, a total loss of friction occurs with the failure of the connection. The screws have bent under stresses and at the same time have been pulled out from the wood. By comparing the force-displacement curves obtained from the two tests of specimens, under the same loading conditions, only an increment of the maximum force and corresponding elastic limit displacement can be pointed out (Figure 4.31). Regarding the stiffness, it remains constant and similar in both cases. The yield point after the elastic limit for the first specimen (LJM1800A) is higher than for the second specimen (LJM1800B). Second specimen ((LJM1800B) shows more ductile behaviour due to local compression of wood and the behavioral difference of screws. Screws pressed the timber locally and finally has led to 23% less load-bearing capacity than in specimen 1 (LJM1800A). The mode of failure of joints with screw connectors has been relatively brittle; the brittleness of the failure mode becomes evident due to the sudden drop of the load after the peak load. The failure mode in specimens with screws has been associated with localised crushing of the timber and bending of the screws. The components of joint have separated in tension zone of timber with a depth of approximately 5 mm (Figure 4.32). 61 Figure 4. 30: Distribution of stress of bending test under monotonic load Figure 4. 31: Load-deformation curve of monotonic tests 62 Figure 4. 32: The failure pattern of two specimens under monotonic loading 4.3.2. Monotonic tests on reinforced specimens A total of 2 lap joint specimens, with dimensions of 90 mm × 90 mm of cross-section (b × h) and 1800 mm length, have been reinforced with undirectional carbon fiber textile. High strength carbon fiber textile has been wrapped all around the timber specimen, as 400 mm wideness (Figure 4.33). It has been bonded parallel to the longitudinal direction of the beam in one layer with two component epoxy. The epoxy, named MasterBrace P 3500, is composed by two parts A (resin) and B (hardener). The mix ratio is 3:1 (Part A to Part B) by volume. Each component were carefully measured and then added part B (hardener) to part A (resin) (Figure 4.34). Technical data of carbon fiber textile and epoxy are given in Table 4.2 and 4.3. Firstly, the initial resin coat with a thickness of 1 mm has been applied with a brush. Subsequently, the unidirectional CFRP fabric reinforcement with a thickness of 5 mm has been placed parallel to the longitudinal direction of the beam and finally a finishing layer of the same epoxy resin with a thickness 1 mm has been applied again (Figure 4.35). The curing time is 48 hours. 63 Figure 4. 33: Dimensions of reinforced specimens in monotonic test (Dimensions are presented in mm) Figure 4. 34: The selected epoxy and carbon fiber textile for reinforcement Table 4. 3: The technical data of carbon fiber textile (Ticem) Property Average value ASTM test method Tensile strength 630 MPa D3039 Tensile modulus 42000 MPa D3039 Elongation at break 1.5 % D3039 Nominal layer thickness 0,5 mm - 64 Table 4. 4: The technical data of epoxy (BASF) Property Average value ASTM test method Adhesion strength on carbon 2.87 MPa D4541:95e1 Tensile strength 35 MPa D638:00 Tension strain at yield 2.0 % D638:00 Tension elastic modulus Flexural strength Flexural modulus Compressive strength Compressive modulus 717 MPa 24.1 MPa 595 MPa 28.3 MPa 670 MPa D638:00 D790:01 D790:01 D695:96 D695:96 Figure 4. 35: General view of reinforced specimens for monotonic test Timber reinforced specimens have been subjected to four point flexural loading, using a 1000 kN displacement control hydraulic jack. Two LVDT with a sensitivity of 0.1 mm, have been mounted for monitoring the vertical deflections at the mid-span under the mid points of two side faces of the beam (Figure 4.36). The loading steps were similar to monotonic tests on unreinforced specimens. The estimated maximum load, Fest, 1000 N has been taken on the 65 basis of previous bending experiments. The load has been applied up to 0.4 Fest, which corresponds to 400 N, maintaned for 30 sec. Then, the load has been reduced to 0.1 Fest, which corresponds to 100 N, maintaned for 30 sec at this value. Thereafter the load was increased until the ultimate load. Constant velocity in the load application has been imposed to 10 mm/min. Figure 4. 36: General view of the set up for monotonic test on reinforced specimen The load-deformation diagram has been obtained from test results (Figure 4.37). The curves show that beam exhibited more essentially linear elastic behaviour up to the failure. The averaged maximum force is approximately 9 times more than in the unreinforced specimen. After a loss of friction, a rapid decrease of the resistance takes place. Thus, the brittle behaviour is replaced by an inelastic phase. Finally, a total loss of friction occurs with the general failure of the connection. Carbon fiber textile reinforced beams revealed less ductile behavior compared to the un-reinforced beams. The CFRP reinforcement derived into a global increment of the maximum load at failure, from 1400 N to 13000 N, which represents an increase of 830 percent. It should be noticed that the rupture of the strengthened timber beams occurred due to the first crack of the solid timber in the tensile region. Failure has been initiated at the joint in tension zone due to screw buckling. Even though the components of the joint have been separated in tension zone, the joint still resisted to the increment of load by means of high tensile strength 66 of carbon fiber textile. When the specimen has reached the peak of load, the complete failure has occurred. Carbon fiber textile has separated from the timber surface (Figure 4.37-4.38). In addition, crack has been formed as horizontal in the first specimen (LJM1800AR) due to shear stress (Figure 4.38). Figure 4. 37: Load-deformation curve of monotonic tests on reinforced specimens Figure 4. 38: Failure pattern of reinforced specimen (LJM1800AR) under monotonic loading 0 2000 4000 6000 8000 10000 12000 14000 16000 -10,000 0,000 10,000 20,000 30,000 40,000 50,000 60,000 70,000 Load (N) Displacement (mm) Reinforced Lap Joint Test 1 (LJM1800AR) Reinforced Lap Joint Test 2 (LJM1800BR) 73 Figure 4. 47: A failure mode due to tensile stress of first reinforced specimen (LPC1800AR) Figure 4. 48: Shear failure in second reinforced specimen (LPC1800BR) under cyclic loading 74 Figure 4. 49: Shear failure of third reinforced specimen (LPC1800CR) under cyclic loading 4.3.4. Discussion of test results The test results show that monotonic tests on unreinforced specimens were mainly influenced by the screws, which would increase the load carrying capacity of the connection. Furthermore, the strengthening of timber joint under bending with CFRP had a beneficial effect on the loadbearing capacity and on the rigidity of the reinforced specimens. The comparison between the reinforced and unreinforced specimens under monotonic loading, confirms that carbon fiber textile led to higher stiffness and strength in the joints. It can be observed that load and deformation capacities of reinforced specimens are higher than unreinforced specimens. The CFRP reinforcement caused an increase in the average maximum load at failure from 1400 N to 13000 N, which represents an increase of 830 percent. Besides, at load-deformation curve of reinforced specimen, sudden drop after the ultimate load is seen, which shows brittle behavior. It should be pointed out the maximum load is approximately similar in monotonic and unidirectional cyclic tests on reinforced specimens. The average of maximum load and top displacement of reinforced specimens are 12000 N and 35 mm under monotonic loading, while 11713 N, 36 mm under cyclic loading. In other words, even if the types of loading are different, the maximum load and displacements of the specimens with CFRP are similar. 75 4.4 Mortise Tenon Joint Tests A set of monotonic tests of unreinforced specimens have been performed in order to describe the behaviour of mortise tenon joint which is commonly used between beam and column members in Turkish traditional timber house. There are numerous examples of this type of joint. Tenon joints members that usually form an "L" or "T" type configuration. The joint comprises two components: the mortise hole and the tenon tongue (Figure 4.50). These joints were implemented with metal fasteners such as; nails, screws or bolts and their ability to carry the loads was achieved through friction. Moreover, various reinforcement techniques such as; metal plates (strips, stirrup), glued composites (glass or carbon fibres textiles) and glued-in rods are used. In this study, joints which are locally strengthened with carbon fiber textile have been tested under monotonic and cyclic loading. The purpose of these tests is to increase the flexural strength and load-bearing capacity of these joints. Besides, carbon fiber textile may prevent premature seperation of joint components under loading. Figure 4. 50: Mortise tenon joint with screw Joints are assumed to be ideally rigid or pinned in some simplified analysis. It is quite obvious that the assumption of pinned joints is conservative, provided that the joints have enough ductility, in a way their rotation may develop. In fact, most joints in real wood structures are more or less flexible or semi-rigid. The slope of the moment-rotation curves to the elastic curves 76 change as the joint is loaded (Figure 4.51a). The moment is dependent on the function of relative rotation between structural elements which are loaded (Figure 4.51b). a b Figure 4. 51: Semi-rigid joint (a), moment-rotation curves (b) 4.4.1. Monotonic tests on unreinforced specimens A total of 3 pine specimens with 10% moisture content, have been used in T-type mortise-tenon joint, with the dimension of 90x90x500 mm beam and 90x90x1000 mm post member (bxhxl) (Figure 4.52). The dimension of 30x40x50 mm tenon (bxhxl) is connected to mortise hole with 2 lateral screws (Ø4.5 mmx h: 80 mm). The post of each specimens has been horizontally placed and bolted to steel reaction wall which has dimensions of 1000x1000 mm triangle shape using HEB 180 profile. Thus, the post has kept the original vertical position. The load has been concentrated in one point through a rectangle metal plate which has the dimension of 1x4x7 mm (bxhxl). Loaded end of the beam has been at a distance of 480 mm from the face of the post. One LVDT having a resolution of 0.1 mm, has been installed for monitoring the vertical deflections at the corner points of lower side of the beam. The tests of specimens have been carried out under monotonic loading and test set up is shown in Figure 4.53. The loading procedure of test has been obtained from BS EN 26891:1991 (Figure 4.54). According to the standard, the load is applied up to 0,4 Fest and maintained for 30 s. Then the load is reduced to 0,1 Fest and maintained for 30 s. Thereafter, the load is increased until the ultimate load. The test is stopped when the ultimate load is reached. Following, the estimated maximum load, Fest, 1000 N was taken on the basis of previous bending experiments. The load has been applied up to 0.4 Fest, which corresponds to 400 N, maintaned for 30 sec. Then, the load was reduced to 0.1 Fest, which corresponds to 100 N, maintaned for 30 sec at this value. 77 Thereafter the load has been increased until reading the ultimate load. Constant velocity in the application of load has been imposed to 10 mm/min. Figure 4. 52: Dimensions of specimens for monotonic loading (Dimensions are presented in mm) Figure 4. 53: Test set up for monotonic loading 78 As a pin-jointed connection, the tenon member rotated around the corner of the tenon shoulder once a bending moment is applied to a single screw connection. Resistance to bending is provided by lateral strength and stiffness of the screw. The effective centre of rotation is at the corner of the tenon shoulder creating an effectively solid hinge point (Hassan, 2008). Moment rotation is the value of force at the load (P1) times the distance of d1 and equals to force at screw (P2) times the distance of d2 (Figure 4.55). 1 1 2 2MP d P d    (4.7) Figure 4. 54: The loading procedure of monotonic test (BS EN 26891:1991) Figure 4. 55: The effective centre of rotation for mortise-tenon joint (Hassan, 2008) 79 Typical load-displacement curves in the three tested specimens measured is shown in Figure 4.56. Initially, the response is linear and elastic, where a linear increment of displacement corresponds to a linear increment of load. When the load reached 800 N, yielding took place in three specimens. Later, a nonlinear load-displacement curve has occurred and smooth plateau associated with tenon end crushing of mortise. The seperation between tenon and mortise has gradually seen and screws have bent under combined stresses. All specimens showed ductile behaviour under bending loading. Finally, a total loss of friction occurred with the global failure of the connection. One of the conclusions is that the maximum bending load and displacement are similar for two specimens, corresponding to approximately 1200 N and 95 mm. The third one showed the maximum bending load and displacement, 1160 N and 90 mm with lower rotational stiffness than others. When considering the distance from the load point to the centre of rotation is 480 mm, bending moment has been calculated as 595.2 N.m. Figure 4. 56: Load-deformation curves of monotonic tests The experiments revealed that at early stages of loading, tenon and mortise have squeezed each other on contact surfaces, thus the specimens squeaked. Then, the tenon member has started to rotate. The upper tenon surface has slipped outside the mortise and has moved downwards, 0 200 400 600 800 1000 1200 1400 020 40 60 80 100 120 Force (N) Displacement (mm) Mortise Tenon Joint Test 1 Mortise Tenon Joint Test 2 Mortise Tenon Joint Test 3 80 while the bottom of tenon surface has slipped inside of the mortise. With the increase of loading, the nonlinear compressive deformation has been observed on the interfaces. Finally, the tenon has been partially pulled out (aprx. 25 mm) and the joint failed since the vertical displacement is excessively large (Figure 4.57). Figure 4. 57: The failure pattern of three specimens under monotonic loading 4.4.2. Monotonic tests on reinforced specimens A total of 3 mortise tenon joint specimens, with the same dimension of 90x90x500 mm beam, 90x90x1000 mm post member (bxhxl) have been subjected to monotonic vertical loads (see Figure 4.58). The dimension of 30x40x50 mm tenon (bxhxl) is connected to mortise hole with 2 lateral screws (Ø4.5 mmx h:80 mm). They have been reinforced with undirectional carbon fiber textile. High strength carbon fiber textiles have been bonded to the upper surface of joint, as 90x200 mm with L shaped. It has been bonded parallel to the longitudinal direction of the beam in single layer with two components epoxy. Furthermore, two CFRP textiles have been bonded with 450 angle to two lateral surfaces of the specimens with a dimension of 100x200 mm (Figure 4.58-4.59). The epoxy, which name is MasterBrace P 3500, is composed by two parts: A and B. The component A is basically an epoxy resin, and Component B is a hardener. By mixing both components, the reaction starts, which is the responsible for hardening. Components A and B have been mixed in the ratio prescribed by the manufacturer that is 3:1 (Part A to Part B) by volume. Each component were carefully measured and then added part B (hardener) to part A (resin) (Figure 4.60). The viscosity of the adhesive plays a very important 81 role in the workability which in turn affects the overall quality of the process. Technical data of carbon fiber textile and epoxy are given in Table 4.3 and 4.4 (in previous part of lap joint). Figure 4. 58: The places of CFRP in specimens Figure 4. 59: The dimension of reinforced specimens for monotonic test (Dimensions are presented in mm) 82 Figure 4. 60: The selected epoxy and carbon fiber textile for reinforcement First of all, the initial resin coat of thickness 1 mm has been applied with a brush on upper surface which is the tension zone of timber specimen. Subsequently, the unidirectional fabric reinforcement with a thickness of 5 mm has been placed parallel to the longitudinal direction of the joint and finally a finishing layer of the same epoxy resin of thickness 1 mm has been applied again. Besides, two CFRP textiles were bonded in 450 angle to each others on lateral surfaces of specimens (Figure 4.61). The curing time was 48 hours. After the preparation of specimens, the post of each specimens was vertically placed and bolted to steel reaction wall which has the dimension of 1000x1000 mm triangle shape using HEB 180 profile. The loads concentrated on one point with rectangle metal plate with a dimension of 1x4x7 mm (bxhxl). Timber reinforced specimens have been subjected to one point flexural loading, at the end of beam the distance of 480 mm from the post. The load has been applied by one loading cell, powered by a maximum capacity of 1000kN hydraulic jack. One LVDT with sensitivity of 0.1 mm, has been used for monitoring the vertical deflections at the corner points of lower side of the beam (Figure 4.62). 89 3 3 4 F bh d L E (4.6) L length of span, in millimetre; h height of beam, in millimetre; b width of beam, in millimetre; F load, in newton; d deflection, in millimetre; For the specimens, modulus of elasticity has been calculated as E1=19.39 N/mm2; E2=28 N/mm2; E3,4,5=26 N/mm2; E6,7,8=21 N/mm2 at cycle 1 (Vy=20 mm), cycle 2 (Vy=40 mm), cycles 3,4,5 (Vy=60 mm), cycles 6,7,8 (Vy=80 mm) respectively (Figure 4.70). Figure 4. 68: Complete load-deformation curves of specimens 90 Figure 4. 69: Plastic strain-cycles diagram of unreinforced specimens Figure 4. 70: Degradation of modulus of elasticity in specimens 91 In order to determine the global modulus of elasticity of timber under bending loads, the section of the graph between 0,1 Fmax and 0,4 Fmax for a regression analysis is used. The slope of graph gives the global modulus of elasticity (EN 408). Generally, for the numerical calculations of timbers, the value of Em is use rather than initial modulus of elasticity, E0 (Figure 4.71). In the graph; F2-F1: is an increment of load in newtons on the regression line with a correlation coefficient of 0,99 and w2-w1 is the increment of deformation in millimetres corresponding to F2-F1. Figure 4. 71: Load-deformation graph within the range of elastic deformation (EN 408) When joints have been subjected to cyclic loading, the beam has been pushed outwards the actuator. The lower contact surface of the tenon and the mortise has squeezed each other simultaneously. A sound of creaking has occurred between the compression contact surface continuously during the tests due to squeezing of timber fibers and interface friction as the rotation increased. With the increased displacements, significant plastic compression deformation has occurred on the contact surfaces and could not restore and turn back to the first state after unloading. It resulted in a gap between tenon and mortise. For all specimens, the failure were seen around 91 mm, when the specimens have been pushed to cycle (2Vy) 160mm. Failure patterns are detected as tenon pull out from mortise hole (Figure 4.72). 92 Figure 4. 72: Failure pattern of three specimens under cyclic loading 4.4.4. Cyclic tests on reinforced specimens A total of 3 mortise tenon joint specimens, with the same dimension of 90x90x500 mm beam, 90x90x1000 mm post member (bxhxl) have been subjected to unidirectional cyclic vertical loads. The dimension of 30x40x50 mm tenon (bxhxl) is connected to mortise hole with 2 lateral screws (Ø4.5 mmx h:80 mm). They were reinforced with undirectional carbon fiber textile. High strength carbon fiber textiles has been bonded to the upper surface of joint, as 90x200 mm with L shaped. It has been bonded parallel to the longitudinal direction of the beam in one layer with two component epoxy. Furthermore, two CFRP textiles have been bonded at 450 angle to two lateral surfaces of specimens with a dimension of 100x200 mm. The technical data of carbon fiber textile and epoxy are given in Table 4.3 and 4.4 (in previous part of lap joint). After the preparation of specimens, the post of each specimens was vertically placed and bolted to steel reaction wall which has the dimension of 1000x1000 mm triangle shape using HEB 180 profile. The loads, which distributed in one point with rectangle metal plate, has the dimension of 1x4x7 mm (bxhxl). Timber reinforced specimens have been subjected to one point flexural loading, at the end of beam the distance of 480 mm from the post. The loading has been applied by one loading cell which powered by maximum capacity of 1000kN hydraulic jack. One LVDT with sensitivity of 0.1 mm, has been mounted for monitoring the vertical deflections at the corner points of lower side of the beam (Figure 4.73). 93 Figure 4. 73: The set up for cyclic test on reinforced specimens The loading protocol is given in Figure 4.66 (at previous part, 4.4.3. Cyclic tests on unreinforced specimens). Estimated yield slip has been determined as Vy=16 mm and target displacements have been calculated by using this value (Table 4.7). The first two target displacements were applied for only one cycle. Further target displacements have been applied as three sets of cycles. First, at 1st cycle, the load applied in compression, until a slip of 25% of the estimated yield slip Vy is reached. The value of Vy has been evaluated by calculation as 0.25Vy=4 mm. Then, the specimen has been unloaded. At the 2nd cycle, the load has been applied in compression up to a slip of 50% of Vy which corresponded to 8 mm and it unloaded to zeroslip. At 3th, 4th, 5th cycles, it has been loaded in compression up to a slip of 100% of Vy, 16 mm. At following set of three cycles the load has been applied three times, 200% of Vy, 32 mm. The failure of all three specimens have occured at 4,00Vy load level (9th step). Typical load-displacement curves for three specimens which have been measured during cyclic loading is shown in Figure 4.74. The average of maximum load and top displacements of three specimens is 3625 N and 55.8 mm. For each of cycles, load-top displacement diagram and plastic strains are given in Figure 4.75 and Figure 4.76. Plastic strain has resulted in permanent 94 deformation under loading and it has not recovered upon unloading (Figure 4.75). The total strain (Tis composed two components: an elastic strain (eand plastic strain (p. Table 4. 7: Loading steps of cyclic tests for reinforced specimens Load steps Number of cycles Target displacement ratio (Vy=16 mm) Target displacement (mm) 1 1 0,25Vy 4 2 1 0,50Vy 8 3-4-5 3 1,00Vy 16 6-7-8 3 2,00Vy 32 9-10-11 3 4,00Vy 64 Figure 4. 74: Load-deformation curve of cyclic tests on reinforced specimens 0 500 1000 1500 2000 2500 3000 3500 4000 4500 0,000 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 Force (N) Displacement (mm) Reinforced Mortise tenon Joint -Cyclic Test 1 Reinforced Mortise tenon Joint -Cyclic Test 2 Reinforced Mortise tenon Joint -Cyclic Test 3 95 Figure 4. 75: Complete load-deformation curves of reinforced specimens When the distance from the load point to the centre of rotation is 480 mm, the moment resistance of joint has been calculated as M (kN·m) = 0.48F. One displacement transducer (D1, m) has been installed below the face of the beam to measure the vertical displacement of the beam and to measure the rotation of the joint, rotation (rad) = D1/0.48. Then, the rotational stiffness has been determined M (kN·m) / rotation (rad) for each cycle. For reinforced specimens, rotational stiffness has been calculated as k1=787 kN·m/rad; k2=324 kN·m/rad; k3,4,5=190 kN·m/rad; k6,7,8=23.78 kN·m/rad, k9=6.95 kN·m/rad at cycle 1 (Vy=4 mm), cycle 2 (Vy=8 mm), cycles 3,4,5 (Vy=16 mm), cycles 6,7,8 (Vy=32 mm), cycle 9 (Vy=64 mm) 96 respectively. The equivalent rotational stiffness decreases gradually in all specimens as rotation increases. Besides, decreasing modulus of elasticity has been calculated by using the equation 4.6 for each cycle. For reinforced specimens, modulus of elasticities were calculated as E1=1569 N/mm2; E2=938 N/mm2; E3,4,5=558 N/mm2; E6,7,8=58.25 N/mm2 ; E9=21.25 N/mm2 at cycle 1 (Vy=4 mm), cycle 2 (Vy=8 mm), cycles 3,4,5 (Vy=16 mm), cycles 6,7,8 (Vy=32 mm), cycle 9 (Vy=64 mm) respectively (Figure 4.77). Figure 4. 76: Plastic strain-cycles diagram of reinforced specimens When reinforced specimens subjected to cyclic loading under pushing, the lower contact surface of the tenon and the mortise has squeezed each other simultaneously. The significant plastic compression deformation occurred on contact surfaces. Failure has been initiated at the joint in tension zone due to rotation of tenon member. Tenon and the mortise members started to separate from each other. Even though the joint has been separated in tension zone, it still resisted to the increment of load by help of high tensile strength of carbon fiber textile. When the specimen has reached the maximum load, carbon fiber textile has separated from the timber surface. CFRP has provided continuity of timber members together until the failure. In other 97 words, CFRP reinforcements lead to progressive/gradual failure of joint rather than abrupt failure. For all specimens, the ultimate failure has been seen detected at 80 mm. The irrecoverable deformation of specimens is 55.80 mm. The modes of failure for reinforced specimens are detected as the rupture of CFRP sheet after the rotation of timber tenon (Figure 4.78). Figure 4. 77: Degradation of modulus of elasticity in reinforced specimens Figure 4. 78: The modes of failure for three reinforced specimens under cyclic loading 98 4.4.5. Discussion of test results Test results show that unreinforced specimens have been mainly influenced by the quality of interlocking of the connection, which has increased the load carrying capacity of the screw. Furthermore, the strengthening of timber joints under bending with CFRP has had a beneficial effect on the load-bearing capacity and on the rigidity of the reinforced specimens. The comparison between reinforced and unreinforced specimens under loading, confirms that carbon fiber textile has led to higher stiffness and strength of the joints. In monotonic tests, the CFRP reinforcement has allowed an increment of the average maximum load from 1200 N to 3600 N, which represents 300 percent more. Besides, at load-deformation curve of reinforced specimen, a sudden drop after the ultimate load is seen, which shows a clear brittle behavior. Even though the failure of CFRP, the joint system still worked and loaded until the ultimate strain. After the peak point of load, with or without CFRP in both conditions, specimens have resisted to load until the deformation, 90 mm. However, in the plateau part of curves the value of load is 1700 N with CFRP while it is 1000 N without CFRP. In other words, under monotonic loading, CFRP provides an increment of load bearing capacity of the joint. Also, the irrecoverable deformations are similar, 90 mm, for the both of unreinforced and reinforced specimens under monotonic loading. In cyclic tests, the CFRP reinforcement provided the increment of the average maximum load from 1115 N to 3625 N. Furthermore, the rotational stiffness of unreinforced specimens were calculated as k1=9.26 kN·m/rad; k2=7.11 kN·m/rad; k3,4,5=5.16 kN·m/rad; k6,7,8=3.79 kN·m/rad at cycle 1 (Vy=20 mm), cycle 2 (Vy=40 mm), cycles 3,4,5 (Vy=60 mm), cycles 6,7,8 (Vy=80 mm) respectively. For the reinforced specimens, rotational stiffness have been calculated as k1=787 kN·m/rad; k2=324 kN·m/rad; k3,4,5=190 kN·m/rad; k6,7,8=23.78 kN·m/rad, k9=6.95 kN·m/rad at cycle 1 (Vy=4 mm), cycle 2 (Vy=8 mm), cycles 3,4,5 (Vy=16 mm), cycles 6,7,8 (Vy=32 mm), cycle 9 (Vy=64 mm) respectively. The rotational stiffness of reinforced specimens is higher than the unreinforced specimens which has led to more brittle behaviour. The rotational behavior of the damaged mortise tenon joints is semi-rigid. When the reinforced specimens have been pushed to 32 mm (cycle 6), it reached the maximum load, CFRP ruptured and sudden drop from 3625 N to 1800 N has been seen. Then, after three cycles more, the reinforced specimens showed ductile behaviour. It continued with constant load (1780 N) until the ultimate strain, 80 mm. Permanent deformation is detected as 71 mm in unreinforced specimen, while it is 55 mm in reinforced specimen. It can be observed that CFRP provided the reduction of deformation as 23% and the increasement of load bearing capacity 105 The adopted finite element mesh has been created using total 24525 solid elements and 41568 nodes. (Figure 5.3). In the analysis SOLID 186 element and SOLID 187 element are used (Figure 5.4). SOLID 186 is a higher order 3D 20-node solid element with quadratic displacement behavior. The element is defined by 20 nodes having three degrees of freedom per node: translations in the nodal x, y, and z directions. SOLID 187 element is a higher order 3D, 10-node element. It has a quadratic displacement behavior and is well suited to modeling irregular meshes. The element has been defined by 10 nodes having three degrees of freedom at each node: translations in the nodal x, y, and z directions. The element supports plasticity, hyperelasticity, creep, stress stiffening, large deflection, and large strain capabilities. Contacts between timber-timber and timber-steel elements have been defined using 3-D contact surface elements (CONTA174) associated with the 3-D target segment elements (TARGE170). CONTA 174 is used to represent contact and sliding between 3D target surfaces and a deformable surface defined by this element. Figure 5. 3: Finite element mesh 106 Figure 5. 4: SOLID 186 and SOLID 187 element types, respectively Note that the stiffness of the frame wall mostly depends on the contact status (both faces touching or not). Therefore, at each contact surface, isotropic Coulomb friction is considered by using coefficients of friction of 0.2 and 0.4 for timber-steel and timber-timber contact, respectively (BS 5975 1996). 5.3.2. Analysis of the unreinforced model under monotonic loading Assuming the values above, the model of the timber beam with lap joint has been calibrated, by applying a monotonic load to the top of the model in displacement control. Figure 5.5 shows the numerical load-displacement curve with the experimental monotonic results. Figure 5. 5: Load displacement curves under monotonic loads 107 A significant fitting between experimental and numerical results is observed both in terms of stiffness and lateral resistance. When the load reaches the peak of load, the joint starts to separate by showing ductile behaviour until the whole collapse. The maximum displacement has been about 25 mm in average. Figure 5.6 presents the directional deformation (Z axis) and maximum deformation of the test. Concentrated normal stresses parallel to grain are shown in Figure 5.7. In particular, the maximum compressive stress take place around screws, which led to local crushing in timber. Besides, the supports and load introduction are exposed to high compressive stress levels. The maximum tension stress is observed at the lower part of the joint. The lap joint is a weak joint type under bending. At first stage of loading at lower bending moment levels, the joint starts rotating by finishing with plastic deformations within the screws and local damage in timber in direct vicinity of screws. Load-deflection behavior of the beam with lap joint is majorly affected by the interaction of the screws with timber; in other words: friction coefficient between the two materials. In the analysis, a value of 0.2 for the friction coefficient was considered for timber-steel contact. The failure mode in specimens with screws is clearly associated with local crushing of timber derived from bending of the screws (Figure 5.8-5.9). Figure 5. 6: Directional deformation 108 Figure 5. 7: Normal stress distribution Figure 5. 8: Damage status in place of screws 109 Figure 5. 9: Damage status of screws 5.3.3. Analysis of the reinforced model under monotonic loading A numerical approach of timber beams strengthened with carbon fiber reinforced polymer (CFRP) composites is shown here. To predict the behavior of timber beam strengthened with CFRP composites, a three dimensional computational model was developed using the generalpurpose FEA program ANSYS. Timber and CFRP composites were modeled as an elastic orthotropic constitutive model until failure. FRP composites are supposed to be bonded on the joint of timber beams, as 400 mm wideness in order to enhance load-carrying capacity. As a solid element types SOLID 186 element and SOLID 187 element were used for timber. The CFRP layer, thickness 0,5 mm, is also meshed with the same element type, SOLID 186 (Figure 5.10). The elastic properties of carbon fiber reinforced polymer (CFRP) is given in table 5.5. 110 Figure 5. 10: Finite element mesh of beam strengthened with CFRP composites Table 5. 5: Elastic parameters of CFRP (Ticem) Young´s modulus X direction: 42000 MPa Young´s modulus Y,Z directions: 8600 MPa Tensile strength (RT,0): 630 MPa Tensile strength (RT,90): 29 MPa Compression strength (RC,0): 1082 MPa Compression strength (RC,90): 100 MPa Between each contact surface, the coefficients of friction have been considered as 0.2 and 0.4 for timber-steel and timber-timber contact. As an adhesive layer, between timber and carbon fiber reinforced polymer, bonded contact was defined. A vertical displacement of 60 mm with a constant amplitude, has been applied to the beam model until the established failure criteria were satisfied. The CFRP element has been modelled as it ruptures when the maximum axial stress exceeds the tensile and compressive strength of the composite. As the progressive decreasing stiffness, tensile matrix stiffness reduction and compressive matrix stiffness reduction were taken as the value of 1 (which means complete stiffness loss). Figure 5.11 shows 111 the numerical load-displacement curve after reinforcement with the experimental monotonic results. Figure 5. 11: Load displacement curves of the strengthening beams under monotonic loads It is observed that the results of FE analysis correlate well with those experimental results. The model captures the behaviours of the strengthened beams of experiments. Particularly with specimen 2, they have similar responses of initially elastic before undergoing a non-linear softening phase and the maximum load point. The response of the beams was essentially linear until the failure. After that, due to the high stiffness of CFRP material, sharp crack and brittle behaviour has been seen in the strengthening beams. The strengthened beam reaches a maximum load of 12 kN at a displacement of about 40 mm. Figure 5.12 shows the stress distribution in the strengthening beam. It is clearly seen that stress is mainly concentrated along the CFRP sheet, due to a higher stiffness. Maximum compressive stress, approximately 238 MPa, is concentrated at the upper part of the beam which exposed the loads. Maximum tension stress, approximately 788 MPa, is concentrated at the lower part of the beam. When the maximum axial stress exceeded the tensile strength of the composite, the strengthened beam collapsed by the tension side. However with the CFRP sheet, the whole 112 timber beam could still provide a certain bending capacity. Finally, CFRP ruptured and separated from timber beam. Slip between the wood and the adhesive did not take place. Figure 5. 12: Normal stress distribution of the strengthening beam Figure 5. 13: Damage status (0-1) of the strengthening beam 113 Two distinct collapse modes of the strengthened timber beams are observed, namely, the timber fracture at a flexure-critical region (near midspan) and at the CFRP composites where stress concentrations occurred (Figure 5.13). The damage status is defined with values of 0,1 and 2. (0: undamaged, 1:partially damaged, 2:completely). The fiber tensile damage in the range of 0- 1 has been detected under part of beam. Besides, the elastic modulus of the CFRP material increased the load-carrying capacity of the strengthened timber beams and also governed the failure mode of the beams. 5.3.4. Comparison of analysis results It can be clearly seen that the strengthening enhances the load carrying ability of the beam with lap-joint. The maximum load and top displacement of unstrengthened specimen are 14000 N and 25 mm, while 12000 N, 40 mm for strengthened specimen. The CFRP shows an increase of the ultimate load by about 800% and a change of failure mode is observed with greater ductility. The tension failure in wood in bending is brittle, for this reason, CFRP layers bonded on the tension side of the beam. The overall aim is then to increase the flexural strength and stiffness, and achieve a ductile compression failure mode. CFRP sheet improves the flexural capacity and rigidity of timber beam. Figure 5. 14: The comparison of the unreinforced and reinforced specimen in FE analysis 114 Ductility index obtained from energy method which is calculated by the area under the curves. Reinforced beam exhibits high ductility index and the main reason was due to higher range of inelastic region in the compression zone (Figure 5.14). It is very obvious that the reinforced beam reachs high ultimate load compare to the unreinforced beam, even though it yields to low ultimate deflection, it has high total energy and ductility. 5.4 Numerical Modelling of Mortise Tenon Joint Timber material has been modelled with a constitutive model based on elastic orthotrophy with maximum stress failure criterion. The mortise tenon joint is connected via two screws in the model. A basic elastic-plastic material model with bilinear isotropic hardening is used for steel of screws, assuming the von Mises yield criterion. Material parameters used for this model are given in previous chapter in Table 4.4, (ASME BPV Code ,1998). 5.4.1. Geometric constraints, mesh and loading The geometry of finite element model, loading, boundry conditions and material axes are shown in Figure 5.15. Subsequently, a vertical displacement of 60 mm is applied, with a constant movement rate, at the one loading point. It is noted that the self-weight of the wood element and standard earth gravity are considered in the analyses as well. The back side of column member is selected as a fixed support which restrain both rotation and translation. Figure 5. 15: Finite element model geometry 121 the progressive decreasing stiffness, tensile matrix stiffness reduction and compressive matrix stiffness reduction have been taken as 1 (which means complete stiffness loss). Figure 5.24 shows the numerical load-displacement curves after reinforcement with the experimental monotonic results and finite element analysis result. Figure 5. 24: Load-displacement curves of the strengthening joints under monotonic loads It is observed that the results of FE analysis correlate well with those experimental results. The developed model captures the behaviours of the strengthened specimens of experiments. Particularly with specimen 3, they have similar inelastic behaviour and the maximum load point. The response of the joints has been essentially linear until failure has occurred. After that, due to the high stiffness of CFRP material, sharp crack and brittle behaviour has seen in the strengthened beams. The strengthened joint reaches a maximum of 3.2 kN at a displacement of about 80 mm. The directional deformation (Z axis) is given in Figure 5.25. Figure 5.26 shows the stress distribution in the strengthened beam. It is clearly seen the stress is mainly concentrated along the CFRP sheet, due to the higher stiffness. Maximum compressive stress, approximately 155 MPa, is concentrated at the upper part of the beam which exposed the loads. Maximum tension stress, approximately 625 MPa, is concentrated at the lower part of the beam. 122 It should be noticed that the failure of the strengthened timber joints occurred due to the seperation of joint components in the tensile region. Failure was initiated at the joint in tension zone due to rotation of tenon member. Even though the joint components have been separated in tension zone, it still resisted to the increase load by help of high tensile strength of carbon fiber textile. When the specimen reached the maximum load, carbon fiber textile pulled out from the timber surface. CFRP has worked as a binder holding two timber members and has provided continuity together until the ultimate deformation. A distinct collapse mode of joint is detected as the rupture of CFRP sheet at a lateral surface of joint after the rotation of timber tenon (Figure 5.27-5.29). Besides, when the seperation of tenon from the mortise member was occured, the screws bent under stresses (Figure 5.30). The damage status is defined with values of 0,1 and 2. (0: undamaged, 1:partially damaged, 2:completely). The fiber tensile damage in the range of 0-1 has been detected in tenon member and on surface of CFRP sheets (Figure 5.27-29). The compressive damage in the range of 0-1 has been detected at screws (Figure 5.30). Figure 5. 25: Directional deformation (Z axis) of the strengthening joint 123 Figure 5. 26: Normal stress distribution of the strengthening joint Figure 5. 27: Damage status (0-1) of the strengthening joint 124 Figure 5. 28: Damage status (0-1) of the tenon in tension zone Figure 5. 29: Damage status (0-1) of CFRP 125 Figure 5. 30: Damage status (0-1) of screws 5.4.4. Comparison of analysis results It can be clearly seen that the strengthening enhances the load carrying ability of the mortise tenon joint. The maximum load and top displacement of unstrengthened specimen are 1020 N and 90 mm, while 3200 N, 75 mm for strengthened specimen. The CFRP shows an increase of the ultimate load by about 313% and a change of failure mode is observed with greater ductility. The tension failure of wood under bending is brittle, for this reason, CFRP layers bonded on the tension side of the joint. The overall aim is then to increase the flexural strength and stiffness, and achieve a ductile compression failure mode. CFRP sheet improves the flexural capacity and rigidity of the timber beam (Figure 5.31). Ductility index obtained from energy method which is calculated by the area under the curves. Reinforced joint exhibits high ductility index and the main reason is due to higher range of inelastic region in the compression zone. It is very obvious that the reinforced joint reaches high ultimate load compare to other joint, eventhough yields to low ultimate deflection, it has high total energy and ductility. 126 Figure 5. 31: The comparison of the unreinforced and reinforced specimen in FE analysis 6. GLOBAL SEMI-RIGID ANALYSIS In the case of existing timber structures to be rehabilitated and reinforced, a realistic interpretation of the global structural behaviour is a primary need. In typical structural configuration of timber construction, the commonly used hinge models are inadequate; because in real structures, where joints have moment resisting capability, the equilibrium conditions may not be reached analytically. The semi-rigid modelling of timber connections, using nonlinear moment-rotation laws and hysteretic rules, intends to represent the behaviour of timber structures with a comparable level of detail for all the structural components. The original and strengthened traditional timber connections are modelled using a nonlinear spring element available in a structural frame analysis software, RSTAB in order to analyze the internal forces, deformations and support reactions of frame in terms of global scale. 127 6.1 Definition of Model Parameters In order to evaluate the internal forces of members and deformation in existing structure and reinforced structure, a selected frame configuration has been analysed under lateral loads (Figure 6.1). The frame configuration is based on the work (Aktas, 2007). The timber posts and beams have been modelled as isotropic linear elastic bars. The mechanical characterization of the materials is obtained through experimental testing on two types of joint. The numerical values for the elastic parameters that have been used are given in Table 6.1. The reference material properties were obtained from experimental tests. The additional information necessary which was not derived from experimental results, was obtained using the Joint Committee on Structural Safety probabilistic model code (JCSS, 2006). Other material properties are estimated based on this model code, given in Table 6.2. Besides, for the infill material, masonry with standard mortar from the material libraries of program, that creates masonry according to EN 1996-1-1:Eurocode 6, has been modelled as isotropic plastic (Table 6.3). The mechanical behaviour of masonry under compression is non-linear (Figure 6.2). Limit compression strength is considered as 10 MPa then yielding, while limit tension is 0.1 MPa (made reference to Chapter 3, Figure 3.3). Figure 6. 1: The frame configuration for analysis Z XY In Y-direction 128 Table 6. 1: Elastic parameters of timber from experimental tests Bending strength (Rm): 72.97 MPa Bending MOE (Em): 13648 MPa Compression strength (RC,0): 39 MPa Compression strength (RC,90): 4.12 MPa Density ( ρ den): 500 kg/m3 Table 6. 2: Elastic parameters of timber from JCSS (2006) Tension strength (Rt.0): 43.782 MPa Tension strength (Rt.90): 7.5 MPa MOE tension (ET,0): 13648 MPa MOE tension (ET,90): 454.93 MPa Shear modulus (Gv): 853 MPa Shear strength (Rv): 6.18 MPa Table 6. 3: Elastic parameters of masonry Modulus of elasticity (E): 1000 MPa Shear modulus (G): 416.5 MPa Poisson´s ratio (v): 0.2 Specific weight (  ): 24.52 kN/m3 Specific weight (  m): 1 Figure 6. 2: The non-linear curve of masonry 129 Plastic deformation capability, ductility and energy dissipation are important concepts that determine seismic behaviour of structure. Selected spring model must reflect the ductility and energy dissipation capacity of the system. In order to represent the behaviour of semi-rigid joints, spring models have been derived from the average of force-displacement curves of three lap joint specimens and the average of moment-rotation curves of three mortise-tenon joint specimens under monotonic bending load (Chapter 4. Experimental analysis). Thus, two nonlinear curves have been obtained from regression curve of specimens (Figure 6.3-6.4). Nonlinearity effects concentrated hinges were applied to the joints. Stiffnesses of connections which have been obtained from experiments, were progressively added at the joints. In total, three spring stiffness values were adopted to the nodals: 1. Hinge1 where lap joints between the beam and beam have been considered as semi-rigid. 2. Hinge2 where mortise-tenon connections between the beam and column have been considered as semi-rigid. 3. Hinge3 where connections between the diagonal (brace) and the main frame have been considered pinned. The translational spring is introduced at Hinge1 and rotational spring for the non-linear analysis of Hinge2. Properties of nonlinear hinges have been defined according to the force displacement and the moment-curvature, following diagrams presented at Figure 6.3 and Figure 6.4. Besides, the connection between the diagonals and the main frame as pinned. To this term, the rotational degree of freedom is released. Figure 6. 3: The translational spring of lap joint 0 2 4 6 8 10 12 14 16 18 0 0,005 0,01 0,015 0,02 0,025 0,03 0,035 0,04 0,045 Force [kN] Displacement [m] without Carbon with Carbon 130 Figure 6. 4: The rotational spring of mortise-tenon joint In Figure 6.4, the reason of the rupture of the curve of joint with CFRP (red curve) is that during the experiments the carbon fiber textile yielded and separated from the timber surface due to rotation of tenon member. Even though, the carbon fiber textile failed, CFRP has worked as a binder holding two timber members and provided continuity together until the ultimate deformation, thus the moment-rotation of the joint continued to rise after rupture. 6.2 Lateral Load Analysis Two load cases have been defined: dead loads and live loads. Dead loads consist of the permanent construction material loads comprising the roof, floor, wall, and masonry, that is 1,35 G=10.8 kN/m. Live loads come from the use and occupancy of a building, that is 1,50 Q=12 kN/m. (G=2 kN/m2, q= 2 kN/m2, l=4 m). The following load combination, depends on Eurocode 6, was considered 1,35 G + 1,50 Q + 1,00 E (in which live load, Q, snow load, S, earthquake load, E). Horizontal loads proportional to the weight of the structure have been used so to simulate seismic action on the structure. So that, different values of horizontal loads have applied at the top of frame until maximum 144 kN. In order to present the capacity curve of the model relationship between the loads and displacements, the loads of 9, 18, 27, 36, 45, 54, 63, 69, 72, 0 0,2 0,4 0,6 0,8 1 1,2 0 0,005 0,01 0,015 0,02 0,025 Moment [kNm] Rotation [rad] without Carbon with Carbon