UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Exposure based durability of FRP strengthened concrete Johnson, Martin McKay 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0377.pdf [ 25.28MB ]
Metadata
JSON: 831-1.0063492.json
JSON-LD: 831-1.0063492-ld.json
RDF/XML (Pretty): 831-1.0063492-rdf.xml
RDF/JSON: 831-1.0063492-rdf.json
Turtle: 831-1.0063492-turtle.txt
N-Triples: 831-1.0063492-rdf-ntriples.txt
Original Record: 831-1.0063492-source.json
Full Text
831-1.0063492-fulltext.txt
Citation
831-1.0063492.ris

Full Text

EXPOSURE BASED DURABILITY OF FRP STRENGTHENED CONCRETE by Martin McKay Johnson B.Sc.Eng., Queen's University, 1997 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES Department of Civil Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 2003 © Martin McKay Johnson, 2003 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of fa fod/Z ln/f^J?/rd£ • The University of British Columbia Vancouver, Canada Date /*? Q-Z> DE-6 (2/88) A B S T R A C T Deteriorated concrete structures is a current problem facing Civil and Structural Engineers. Current repair techniques include replacement of the structure and or costly and expensive retrofits. . Current research is focussing on using Fibre Reinforced Plastics (FRPs) as a possible retrofit material. This study focuses on the durability issues surround and FRP techniques pioneered at the University of British Columbia. An FRP comprised of short glass fibres (GFRP) and polyester resins is sprayed on the tensile face of the deteriorated concrete. Early tests show that this has tremendous potential to be a cost effective method of retrofitting concrete structures. In order to further develop this material as an effective repair technique additional study on the durability of the material in neutral and high pH aqueous solutions at different exposure periods, with and with out prestress on the concrete/FRP interface, was required. 72 concrete beams strengthened with GFRP and subject to different environmental exposures and concrete/FRP bond stress. The effect of exposure and bond prestress is evaluated to determine the durability issues related to the use of GFRP as a concrete strengthening material. Study results are also used to provide analysis of the concrete/FRP bond. Based on the results of this study, recommendations regarding future study and the future use of GFRP as concrete strengthening material are made. T A B L E O F C O N T E N T S A B S T R A C T ii T A B L E O F C O N T E N T S iii L I S T O F T A B L E S viii L I S T O F F I G U R E S ix N O M E N C L A T U R E xi A C K N O W L E D G E M E N T S xiii 1.0 I N T R O D U C T I O N l 1.1. General 1 1.2. Objectives 3 2.0 L I T E R A T U R E R E V I E W 4 2.1. Introduction 4 2.2. Fibre Reinforced Plastics 7 2.2.1. Introduction 7 2.2.2. Glass Fibres 7 2.2.3. Polyester Resins 9 2.2.4. Vinylester Resins 11 2.2.5. Application Methods 12 2.3. Durability of E-glass Laminates 15 2.3.1. Overview 15 2.3.2. Moisture Effects 17 2.3.3. Alkaline Effects 21 2.3.4. Temperature Effects 23 2.3.5. FRP and Concrete Bond Durability 24 iii 2.4. Previous Research 28 2.4.1. FRP Laminates Bonded to Concrete 28 2.4.1.1. Small Scale Specimen Tests 29 2.4.1.2. Large Scale Specimen Tests and Applications 35 2.4.2. FRP Reinforcing Bars 38 2.4.2.1. FRP Reinforcement Bars - Fibre Performance 40 2.4.2.2. FRP Reinforcement Bars - Matrix Performance 41 2.4.3. Research at the University of British Columbia 43 3.0 E X P E R I M E N T A L P R O C E D U R E 48 3.1. Introduction 48 3.2. Material Properties 48 3.2.1. Concrete 48 3.2.2. Glass Fibre 49 3.2.3. Vinylester Resin 50 3.2.4. Polyester Resin 51 3.2.5. Catalyst and Promoters 51 3.2.6. Post-Tensioning Steel 52 3.3. Preparation of Specimens 52 3.3.1. Casting of the Concrete Specimens 52 3.3.1.1. Preparing the Prestressing Duct 53 3.3.2. Application of the FRP Wrap 54 3.3.3. Post-Tensioning of Specimens 56 3.4. Environments 58 3.5. Summary of Specimens and Environments 59 3.6. Mechanical Testing 60 iv 4.0 E X P E R I M E N T A L R E S U L T S 63 4.1. Beam Testing Results 63 4.1.1. Observations Prior to Testing 63 4.1.1.1. Before Exposure 63 4.1.1.2. During Exposure 65 4.1.1.3. After Exposure 67 4.1.2. Observations During Testing 70 4.1.3. Results and Observations at the Conclusion of Testing 72 4.1.4. Cracking and Ultimate Strengths and Deflection 73 4.1.4.1. Calculations 74 4.1.4.2. Summary 75 4.1.5. Toughness of Specimens 76 4.1.5.1. Calculations 76 4.1.5.2. Summary 77 4.1.6. Initial Stiffness 78 4.1.6.1. Calculations 78 4.1.6.2. Summary 79 4.1.7. Overall Summary 80 5.0 DISCUSSION 82 5.1. Introduction 82 5.2. Load Versus Deflection Curves 83 5.3. Cracking Load 87 5.3.1. Control Specimens 88 5.3.2. Wrapped Specimens 88 v 5.3.3. Wrapped and Prestressed Specimens 89 5.4. Ultimate Load 90 5.4.1. Effect of Time 90 5.4.2. Effect of Laminate 91 5.4.3. Effect of Exposure 91 5.4.4. Effect of Prestress 92 5.5. Midspan Deflection 92 5.5.1. Effect of Time 92 5.5.2. Effect of Laminate 92 5.5.3. Effect of Exposure 94 5.5.4. Effect of Prestress 96 5.6. Fracture Energy 96 5.6.1. Effect of Time 96 5.6.2. Effect of Laminate 97 5.6.3. Effect of Exposure 98 5.6.4. Effect of Prestress 100 5.7. Initial Stiffness 101 5.7.1. Effect of Time 101 5.7.2. Effect of Laminate 101 5.7.3. Effect of Exposure 102 5.7.4. Effect of Prestress 103 5.8. Overall Comparisons 104 5.8.1. Chemical Exposure 106 5.8.2. Effect of Prestress 108 5.8.3. Overall 110 5.9. Impacts and Implications I l l vi 6.0 C O N C L U S I O N S A N D R E C O M M E N D A T I O N S 113 6.1. Conclusions 113 6.2. Recommendations 114 7.0 R E F E R E N C E S 116 A P P E N D I X A 123 A P P E N D I X B 128 A P P E N D I X C 138 A P P E N D I X D 146 vii L I S T O F T A B L E S Table 2.1. Summary of Results from UBC Moth Beam Tests 47 Table 3.1. Concrete Mix Design for 45 MPa Concrete - Ratios 48 Table 3.2. Typical Room Temperature Properties of Advantex Glass Fibre 50 Table 3.3. Typical Room Temperature Properties of 1/8" Clear Castings Made with Derakane XD 8084 Resin 50 Table 3.4. Typical Room Temperature Properties of Clear Castings Made with Aropol Kl 907 Resin 51 Table 3.5. Typical Room Temperature Properties of Laminate Made with Aropol K1907 Resin 51 Table 3.6. Dywidag Threadbars - Technical Data 52 Table 3.7. Summary of Specimens and Environments 59 Table 4.1. Summary of Cracking and Ultimate Strengths and Deflections 76 Table 4.2. Fracture Energy of Specimens 78 Table 4.3. Initial Stiffness of Specimens 80 Table 4.4. Overall Summary of Experimental Results 81 Table 5.1. Summary of Results 83 Table 5.2. Comparison of Findings to Banthia et al. (1996) for Wrapped Beams, Room Temperature Air Exposure 93 viii L I S T O F F I G U R E S Figure 2.1. Polyester Structure 10 Figure 2.2. Vinyl Ester Structure 12 Figure 2.3. Spray-up Schematic 14 Figure 2.4. The Spray-up Operation 15 Figure 2.5. Glass Dissolution in Water 20 Figure 2.6. Glass Dissolution in an Alkaline Environment 22 Figure 2.7. Typical Bridge in British Columbia's Lower Mainland 45 Figure 2.8. Typical Reinforcing Steel Corrosion and Concrete Spalling 46 Figure 3.1. Schematic of Post-Tensioning 57 Figure 3.2. The Post-tensioning Operation 58 Figure 3.3 Sample Identification Legend 60 Figure 3.4. Schematic of Beam Setup 61 Figure 3.5. Testing Apparatus (Japanese Yoke) and Data Acquisition Unit 62 Figure 4.1. Specimens after Exposure (Control - Top, Water Exposure - Middle, High pH Exposure - Bottom) 68 Figure 4.2. Fibre Prominence after high pH Exposure 68 Figure 4.3. Failed Specimen Showing Failure Crack and Delaminated Sheet 71 Figure 4.4. Typical Method of Slope Determination for the Initial Stiffness Calculation 79 Figure 5.1. Typical Load Versus Deflection Curve for a Concrete Beam Strengthened with a CSM FRP Laminate 84 ix Figure 5.2. Load versus Deflection Curve for a Concrete Beam Strengthened with a CSM FRP Laminate with 60 day exposure to a Room Temperature Air Environment 86 Figure 5.3. Difference in Cracking Load (High pH - Water) versus Time for pH and Water Exposures 87 Figure 5.4. Typical Curves for Load versus Midspan Deflection for a Control Specimen and Strengthened Specimen 97 Figure 5.5. Flowchart Demonstrating the Reduction and Simplification of Data 105 Figure 5.6. Combined Overall Response of the Specimens to the Variables (Exposure, Prestress, Wrap) Independent of Time 107 Figure 5.7. Load versus Deflection - Average Comparison - Prestressed Wrap - 90 Days... 108 x NOMENCLATURE A - Area ACI - American Concrete Institute AR - Alkaline Resistant AR-Glass - Alkaline Resistant Glass ASTM - American Society for Testing and Materials ASTM - American Society for Testing and Materials C - Plain Concrete Sample Caltrans - California Department of Transportation CFRP - Carbon Fibre Reinforced Plastic or Polymer CFRP - Carbon Fibre Reinforced Plastic or Polymer CoNap - Cobalt Napthenate CSM - Chopped Strand Mat DMA - Dimethylaniline e - Eccentricity FC - FRP Sprayed Concrete FHWA - Federal Highway Administration FRP - Fiber Reinforced Plastic, Fiber Reinforced Polymer FRP - Fibre Reinforced Plastic or Polymer G - Interfacial Fracture Energy G II - Interfacial Fracture Energy - sliding mode GFRC - Glass Fibre Reinforced Concrete GFRP - Glass Fibre Reinforced Plastic or Polymer G - Interfacial Fracture Energy - opening mode I - Moment of Inertia of the Section LVDT - Linear Voltage Displacement Transducers MEKP - Methyl Ethyl Ketone Peroxide MEKP - Methyl Ethyl Ketone Peroxide MoTH - Ministry of Transportation and Highways P - Applied Load PFC - Post Tensioned FRP Sprayed Concrete PH - High pH Solution Exposure RC - Reinforced Concrete RT - Room Temperature Exposure SiOH - Silicon Hydroxide tg - Glass Transition Temperature UBC - University of British Columbia UV - Ultra Violet W - Plain Water Exposure y - Distance from Centroid to outside fibre a - Stress xii A C K N O W L E D G E M E N T S This research project was made possible by ISIS Networks of Excellence. Their support is gratefully acknowledged. I would also like to acknowledge the patience, support and advice of Dr. Nemy Bathia. I am grateful for your assistance. I also extend a huge amount of gratitude to my wife Amy and our beautiful daughter, Tenley Keira McKay, who provided me with the last bit of motivation to make this thesis a reality. xiii 1.0 I N T R O D U C T I O N 1.1. General Deterioration of concrete structures has been identified as a major problem facing much of the world's civil infrastructure. Concrete, like any other building material, is subject to deterioration and damage by applied loads, environmental factors and short falls in design. In the lower mainland of British Columbia, the Ministry of Transportation and Highways estimates that there are more than 100 highway bridges in need of immediate attention because of internal reinforcing steel corrosion, concrete spalling and a possible deficiency in shear strength. In the United States alone; it estimated that there are at least half a million bridges are structurally deficient or functionally obsolete (Karbhari and Engineer, 1996b) and one bridge is taken out of service every day. Concrete bridges around the world provide the greatest evidence of reinforced concrete deterioration. Cities everywhere are growing, and with the increase in population comes an increase in the number and size of vehicles that are loading the bridges. The impact of this is that bridges that were once designed to accommodate one type of design vehicle are now being loaded by vehicles that are both heavier and longer than the original design vehicle at an increased frequency. Another contribution to the deterioration of concrete bridges is the environment. In the regions of the world where there is a significant winter climate, many bridges are subject to freeze-thaw cycling and de-icing salts. Repeated cycles of freezing and thawing lead to crack growth throughout the concrete members. The numerous cracks in the concrete provide preferential pathways for the ingress of chlorides provided by the de-icing salts used to keep the bridge safe and ice-free during the winter months. The chlorides once inside the concrete attack the reinforcing steel leading to corrosion of the reinforcing steel. The corrosion products increase 1 the original volume of steel by a factor of three, and this increase in volume results in internal tensile forces that cause the concrete cover to spall off exposing more steel to corrosion. Concrete design is a constantly evolving science which results in changes to the building codes, changes to mix designs and changes to reinforcing technology. It is not economically possible to rebuild every concrete structure with every design code change. In many cases, older concrete structures were designed with additional factors of safety. Some structures, however, need to be strengthened and retrofitted to avoid costly replacement. Fibre Reinforced Plastics (FRPs) are new materials that are beginning to gain acceptance as an appropriate material for concrete strengthening and retrofit. These materials have seen widespread use in the following industries: • chemical storage; • aviation; • sporting goods; and • performance automobiles. The light weight, high tensile strength and durability of these materials make them an ideal reinforcement for deteriorated concrete. The current factors limiting the widespread use of FRPs include: • price; • lack of performance (strength, durability, bond) data and design guidelines; and • resistance to change by engineers, designers, and construction professionals. Ongoing research at the University of British Columbia is centred on the use of short E-glass fibres mixed with a polyester resin to create an FRP laminate, which is applied using the spray-up 2 application method. The E-glass fibres and polyester resins are an economical alternative to the hand laid up continuous fibre carbon fibre/epoxy laminates that are currently in use. 1.2. Objectives The short fibre E-glass/polyester laminate has been shown to have an economical advantage over both traditional strengthening techniques and carbon fibre/epoxy laminates, and there has been demonstrated strength improvements obtained through its use (Boyd, 2000); however, there are still questions related to the durability of the laminate and the laminate/concrete bond. The research program detailed in the following report had four main objectives: 1. Investigate the durability issues surrounding the use of a glass fibre reinforcement and polyester resin based laminate as a method of concrete strengthening; 2. Investigate the effects of sustained, low level, tensile stresses on the concrete/FRP interface; 3. Identify the effects of different environments on the glass fibre; and 4. Comment on the applicability of using E-glass/polyester laminates as a strengthening method for concrete. With the information provided by this investigation it will be possible to gain a better understanding of the behaviour of E-glass/polyester laminates and their ability to strengthen concrete structures effectively in both interior and exterior environments. 3 2.0 L I T E R A T U R E R E V I E W 2.1. Introduction The current state of the world's infrastructure is far from ideal. Temperature cycling with the seasons, the use of de-icing chemicals, increases in populations and population density, and bigger, heavier vehicles are wreaking havoc with concrete infrastructure worldwide. Freezing and thawing during winter months has caused concrete to crack and deteriorate. De-icing chemicals have caused the reinforcement in concrete structures to corrode, which leads to spalling of the concrete cover. Increased numbers of people driving heavier vehicles means that the loads crossing concrete bridges are heavier and more frequent: concrete structures that were once built to code are now subject to increased loading and considered substandard by recent code revisions. Karbhari and Engineer (1996b) report that at least half a million bridges in the United States alone are structurally deficient or functionally obsolete. Mack and Holt (1999) provide a list of characteristics that make a structure deficient: excessive deflection, insufficient reinforcement, poor quality concrete, reinforcement corrosion, structural damage, fire damage, and alkali/silica reaction. There are a number of solutions to these infrastructure problems: • Rebuild and replace. This is often economically impossible with government spending priorities. • Strengthen with steel jackets or post-tension with steel members. This solution is common practice; however, the environment that causes corrosion of steel reinforcing bar will also cause corrosion of the strengthening steel. • Use fibre reinforced plastics (FRPs) to wrap the degraded concrete. This is the subject of the following review. 4 FRP laminates saw their first use in military applications, where their light weight and high strength was especially suited to the military's demands. Since then, FRPs and composite materials have moved into the public sector. The prevalence of FRP materials is at the point that we see and/or use them on a daily basis: car panels, sporting goods, boats, and furniture are just a few examples of current composite applications. The materials used in structural applications are very similar to the composites already in use today. Section 2.2. provides a detailed overview of one group of FRP materials that may provide a potential solution to our current infrastructure problems. FRP strip bonding was first introduced in 1984 by Urs Meier with the strengthening of the Ibach bridge in Switzerland. With that repair, Meier demonstrated to the world the potential use for carbon fibre and glass fibre laminates as a material for structural repair. Emmons, Thomas, and Vaysburd (1998) present the benefits of the FRP strengthening solution: • Versatility and flexibility. The laminate can be easily shaped in the field to fit the geometry of the structure. • There is no increase in the dimensions of the structure. Previous strengthening schemes would impact available overhead clearance; however, FRPs are much thinner and add at a maximum 10 mm of thickness to the structure. • There is little increase in the dead load. FRP laminates have a much higher strength to weight ratio than steel which means that not only are the laminates thinner, they are also much lighter. • Easy handling in limited spaces. The light weight of the laminates allows for their placement with a minimum of heavy equipment. The laminates can be easily handled 5 with two people eliminating the need for cranes and allowing the materials to be placed in low clearance areas. • Expediency and ease of construction. Lightweight materials with simple application procedures mean that the materials are applied easily and quickly with little impact on the use of the structure. • Good economy. Although the material costs are higher than those of steel, the savings in the installation costs and life cycle costs compensate for the higher material costs. One other attractive feature of FRP laminates provided by Debaiky, Green, and Hope (1999) is the barrier to oxygen, moisture, and chlorides that the laminate provides. Currently, there is little in the way of guidelines for using FRP laminates to strengthen existing concrete structures. Japan has adopted a set of guidelines for the use of FRP materials, and the latest revision of the Canadian Highway Bridge Design Code has a chapter dedicated to FRP use. Members of the ISIS Canada Network of Centres of Excellence have also developed the Design Manual for the Strengthening Reinforced Concrete Structures with Externally-Bonded Fibre Reinforced Polymers (FRPs). In the United States, the American Concrete Institute (AO) has a committee developing a specification and guidelines for the use of FRP as an internal reinforcement and working towards adopting set guidelines for the use of external reinforcement. Clearly there is still work to be done in developing guidelines and specifications. What is needed first are answers provided by research. The following pages review the current state of the art in FRP repair with a specific emphasis on a potentially economical solution: glass fibre and polyester/vinylester resin. 6 2.2. Fibre Reinforced Plastics 2.2.1. Introduction A composite material is defined as any material that is composed of two or more distinct materials with distinct sets of mechanical properties separated by an interface. When combined in a composite the newly created material possesses its own set of distinct properties. In the case of fibre reinforced plastics the composite created combines the beneficial aspects of both the polymer matrix and the chosen fibre reinforcement. This study considers issues surrounding the use of a composite laminate to strengthen cast concrete beams. The laminate is composed of a polyester matrix strengthened by E-glass fibre with a vinylester resin serving as the primary bond between the laminate and the concrete. The laminate takes advantage of the light weight and strength of the glass fibres, the flexibility and economics of the polyester resin, and the durability and toughness of the vinylester resin. Together these three materials act to strengthen existing concrete. It should be noted that the strength of the matrix is not a critical property in determining the strength of the composite. The resins used in composites are much weaker than the fibres used and as a result will often crack before the fibres. 2.2.2. Glass Fibres (Rubin, 1990) Glass fibres are the most abundant method of reinforcing FRPs with commercial production of glass fibre beginning in the 1930s. Other widely used fibres include carbon, graphite, and aramid; however, they are beyond the scope of this work. The popularity of glass fibres can be attributed to a number of factors: they provide a good balance of strength, Young's Modulus, dimensional stability, thermal stability, and cost. Certain glass fibres have been specifically developed for use in reinforced plastics: E-glass is one of theses fibres. E-glass is a borosilicate glass with the "E" stemming from its originally intended use in electrical applications. During production, the glass fibres are coated with a silane coupling agent. This allows for improved adhesion between the polymer matrix and the glass fibre. It also helps to retain the material properties of the glass under corrosive conditions by offering a direct chemical bond linkage between the glass and the resin. Overall, E-glass has been shown to be reasonably stable in a wide pH range (3.0-10.0) with low thermal expansion rates (Mack and Holt, 1999). The addition of glass fibre to the polymer matrix improves the overall mechanical properties of the laminate. Tensile strength, thermal stability, dimensional stability, and impact strength are all improved with the introduction of fibre to the resin. These properties all increase in proportion to the number of fibres added to the resin; however, this relationship is not linear, and there is a limit to the amount of fibre that can be successfully integrated with the polymer matrix. E-glass does not absorb water; however, the surface is highly polar allowing for the absorption of moisture during storage, transportation, and processing (Mack and Holt, 1999). Research has extensively shown the decrease in E-glass properties in moist environments, and this can be of great concern to composites that are exposed to moisture: wicking action by the exposed glass fibres can draw moisture into the composite to the detriment of the composite's mechanical properties. Section 2.3.2. provides additional information on fibre degradation in aqueous environments. Glass fibres are available in numerous two dimensional fibre arrangements. The strongest arrangement is a unidirectional weave. This aligns all of the fibres in one direction and 8 produces a laminate with anisotropic properties: extremely strong parallel to the fibre direction and weak perpendicular to the fibre direction. Woven roving is another fibre arrangement. In this fabric fibres form a weave with the fibres running perpendicular to each other producing a laminate with orthotropic properties: strong in the directions of the fibres and reduced when not in line with the fibres. Chopped strand mat (CSM) is a completely random arrangement of the chopped fibres. This mat can have two different forms. It can be manufactured in fabric form and cut to fit the required shape, or it can be produced during application by the spray up process. This subject is covered in detail in section 2.2.5. In both cases, the final laminate has isotropic properties in plane: equal properties in all two dimensional plane directions. It should be noted that there are considerable strength variations related to the directionality of the fibres in composites manufactured with multidirectional fibre fabrics. 2.2.3. Polyester Resins (Rubin, 1990) Polyester resins are among one of the four most important polymer systems: the other three are phenolic, amino, and epoxy resins. Although polyester resins were first patented in 1941, they did not begin to see extensive use until the early 1960s. The first major application using polyester resins was the construction of radomes for the United States military aircraft in the 1950s. Due to their light weight, high strength, and modest price polyester resins are one of the most widely used thermosetting resins today. In terms of numbers, polyester resin production represents 15 percent of all thermosetting resin production with 90 percent of the polyester resin produced incorporated into reinforced plastics. 9 The chemical makeup of polyester resins can be summarized quite succinctly. Polyester resins are low molecular weight, linear polymers with carboxylic ester linkages, and carbon - carbon double bonds. Styrene is a vinyl monomer added to the liquid resin to improve viscosity and workability. The mixture is further stabilized by adding an inhibitor. In order to form the polymer chains the resin must be cured by adding free radical catalysts. Polyester resins use peroxides as the primary catalyst. The peroxides used are obtained using methyl ethyl ketone peroxide (MEKP); however, the decomposition of peroxides occurs at a slow rate so the resin is often promoted with the accelerators cobalt napthenate (CoNap) and dimethylaniline (DMA). Curing of polyester resins is an exothermic reaction that provides adequate heat to ensure a satisfactory cure. These reactions convert approximately 90 percent of the reactive carbon double bonds into single bonds eliminating the need for post curing in an elevated heat and/or pressure chamber. Ester groups 0 0 0 H O - - C O -c —c- 0 —c--c=c- C O C — C ™ O H Reactive sites Figure 2.1. Polyester Structure 10 2.2.4. Vinylester Resins (Rubin, 1990; Dow Plastics, 1999) Vinylester resins are the reaction product of an epoxy resin and a monofunctional, ethylenically unsaturated carboxylic acid. The epoxy resin backbone is responsible for the toughness characteristics associated with vinylester resin, which exceed those associated with polyester resins. Certain vinylester resins have been further modified with a reactive elastomer that further enhances the toughness properties, adhesive strength, abrasion resistance, and resistance to severe mechanical stress. The chemistry behind these improved properties can be attributed to four key areas: the epoxy backbone, cross-linking of vinyl groups, shielding of the ester groups, and the alternating ester groups. The combination of the epoxy backbone with the vinyl groups for high reactivity and the styrene monomer for low viscosity retains the chemical resistance and strength of the base epoxy resin while improving the reactivity and basic handling characteristics. The superior toughness of the vinylester resin is attributed to the fact that cross-linking only occurs in the terminal vinyl groups. This leaves the entire polymer chain free to elongate under stress; and thus, absorbing mechanical and thermal stresses and shocks. The chemical resistance of vinylester resins lies in the pendant methyl group in the methacrylate structure. This arrangement provides excellent shielding of the ester group protecting the esters from hydrolysis, which is the initial site of attack by an acidic or basic media. A further advantage over polyester resins is that vinylester resins contain ester groups throughout the molecule, usually alternating with a glycol. Like polyesters, vinylesters require a catalyst to polymerize. The same peroxide catalyst, MEKP, also works for well with vinylester resins. The curing time is controlled through the addition of the promoters cobalt napthenate and dimethylaniline. The curing of vinylester resin, like polyester resin, is an exothermic reaction allowing complete cure at room temperatures. 11 Terminal unsaturation - reactive site o 1 I I CH 2 =C—C—O—CH 2 —CH—CH 2 —R* toughness Epoxy backbone CH3 OH A A Hydroxyl group-improved wetting and bonding to glass Methyl group shields ester linkage-improved resistance to acidic and basic media Figure 2.2. Vinylester Structure (Rubin, 1990) The toughness and durability of vinylester resins has led to their extensive use in chemically aggressive environments: absorption towers, process vessels, storage tanks, piping, hood scrubbers, ducts, and exhaust stacks. Not only are all of the previously mentioned environments chemically aggressive, thermal shock, and temperature extremes are also common. Through 30 years of use vinylesters have proven performance and reliability. It should be noted that the greatest drawback to vinylester resin is the price. At two to three times the price of polyester resins, the use of vinylester resins has some economic drawbacks; however, when compared to epoxy resins vinylesters are approximately half the price. 2.2.5. Application Methods Currently there are many different ways of manufacturing FRP; however, there are only three methods that are used in concrete strengthening applications. The most common technique is hand lay-up. In this method, a quantity of catalyzed resin is applied to the concrete bonding surface. Then pieces of pre-cut fabric are placed (laid up) on the resin 12 prior to curing. Additional resin may be applied to the fabric to ensure the complete saturation of the material (wet out). A serrated roller is used to roll out any trapped air and excess resin. Hand lay-up is the preferred method for applications that involve small surface areas, intricate geometries, restricted access, and require a consistent thickness and volume of reinforcing fibres. Another method that is used is filament winding. This technique of FRP application winds continuous fibre roving around cylindrical objects. This has shown to be quite effective in strengthening deteriorated concrete columns. Before application, the fibres pass through a resin bath where they are saturated with resin. The fibres are then wound onto the desired cylindrical object. There are three limitations to the use of this method: • Object geometry: it is difficult to wrap rectangular columns and only cylindrical columns are wound at this time. • Accessibility: in order to wind a column clearance is required around the column. This limits the use to exposed columns. It would be impossible to wind bridge beams and girders. • Economics: winding is an equipment intensive process. Vertical winders used in column winding are rare, and the set up time is equivalent, if not greater, than the winding time. 13 Figure 2.3. Spray-up Schematic (Rubin, 1990) The third method that is starting to be applied to concrete strengthening is the spray-up method of F R P application. Figure 2.3. provides a general schematic of the process. Large spools of continuous roving are fed to an air driven chopper that chops the fibre to a prescribed length as it is sprayed onto the concrete surface. Simultaneously, resin and catalyst are mixed and sprayed on to the concrete surface. The fibre and resin mix is then rolled out with a serrated roller and a chopped strand laminate is produced. This laminate has nearly identical properties to that of a laminate produced using chopped strand mat fabric in the hand lay-up process. The spray-up application is excellent for large surface areas of simple geometries. A large quantity of resin and fibre can be applied in a short period. The other main advantage is the method in which the resin is catalyzed: resin is only catalyzed immediately before application eliminating the need to catalyze numerous batches of resin during the application procedure. This method is not perfect. The equipment is maintenance intensive, and since the resin is sprayed the chemicals are vaporized leading to possible health hazards and odour complaints; however, the greatest drawback is operator skill. Considerable operator skill is required to deposit a uniform coating of the resin and glass mixture (Rubin, 1990). It is difficult to maintain consistent fibre volumes and build up thickness. High fibre rebound and fibre over-spray can also be 14 a problem if the resin/fibre/air mix is not appropriate. Test cores must be drilled from the final lay-up and tested to ensure an adequate application of material. Figure 2.4. The Spray-up Operation 2.3. Durabi l ity of E-glass Laminates 2.3.1. Overview No upgrading technique is acceptable if there is little guarantee of durability over the predicted lifespan of the strengthening solution. The use of glass fibres in FRP laminates is especially susceptible to scrutiny due to failed attempts in their use as the primary fibre in fibre reinforced concrete. It is common knowledge that glass fibres are known to degrade in the presence of water, acidic, and alkaline solutions, and even with the developments in new glass chemistry, these fears still linger. 15 Glass fibres are more susceptible to environmental effects than carbon fibres (Karbhari and Engineer, 1996b). Glass fibres have reduced durability when subjected to sustained load (Buck, Lischer, and Nemat-Nasser, 1998). Both of the previous statements are well documented in numerous sources; however, there must be some discussion to quantify these statements and evaluate their importance. Liao, Schultheisz, Hunston, and Brinson (1998) provide a comprehensive list of all the environmental effects on FRP laminates: moisture, temperature (thermal fluctuations and thermal stresses), solar radiation (UV), chemical agents, and microorganisms. With the exceptions of ultra violet effects, which are easily preventable, and micro-organisms, the following sections will examine in detail the environmental effects on FRP laminates. Liao et al. (1998) provide the following description of an FRP subject to five years of weathering. Fibre prominence (exposure of fibres) >^ Rupture of the thin resin layer covering the ridging fibres Spalling of the resin at the failure site Fibre prominence and a network of micro-cracks The initial question to be answered in an investigation of environmental effects is the onset of degradation, and here there seems to be some controversy. Karbhari and Engineer (1996b) report that most of the effects that cause the losses in strength and stiffness occur during the first few months of exposure; however, it must also be considered that military aircraft radomes manufactured with an E-glass and epoxy laminate show no appreciable structural damage after 20 years, and there is the difficult problem of dealing with 16 fibreglass boat hulls: after 50 years abandoned boat hulls are failing to break down (Mack and Holt, 1999). The findings of Mack and Holt are backed up research by Liao et al. (1998) who note that short term weathering in a moderate environment will not produce significant degradation on glass fibre composites. It must be noted that the glass fibres alone do not determine the durability of an FRP strengthening system. Karbhari and Engineer (1996b) note that matrix degradation is largely to blame for the significant decrease in load at failure, deflection at failure, and the reduction in flexural stiffness for concrete members strengthened with FRP laminates. Attention must be paid to the resin selection; as mentioned earlier, different resins have superiority in different areas when it comes to resisting environmental attack. The following sections investigate and explain the effects of different environmental conditions on FRP laminates in terms of both fibre durability and resin durability. Although far from complete, it will provide a thorough overview on the behaviour and durability of E-glass composites subject to environmental attack. 2.3.2. Moisture Effects Of the three environmental effects reviewed here, the effect of moisture on a laminate is of greatest concern. While it is true that a highly alkaline environment has a damaging effect on glass fibres, the possibility for exposure to such an environment is limited by geographic location (attack by chlorides due to de-icing of the roads and sea spray) and the state of the concrete being strengthened: aged concrete has been shown to have a pH not exceeding 9.0 (Steckel, Hawkins, and Bauer, 1999). Moisture, however, is nearly 17 impossible to avoid. Extended periods of exposure to rainfall and humidity can be extremely detrimental to FRP laminates. FRP laminates exposed to extended periods of moisture all share the same characteristics: a lowering of the glass transition temperature (tg) leading to plasticization of the matrix and stress corrosion and pitting of the glass fibres. Moisture uptake in glass fibre laminates occurs rapidly with the exposure of glass fibres to the moist environment. Continuous fibres in unidirectional laminates wick the moisture in a continuous manner along the length of the fibre. The random fibres in a chopped strand mat provide complex pathways for moisture uptake and movement in the conditioning process, and although the moisture may enter the laminate in a more convoluted route, the loss of strength associated with the moisture uptake is consistent with both types of fabric (Buck et al., 1998). The reduction of the glass transition temperature and the plasticization of the matrix are critical to the degradation of the laminate: a decrease in the tg is seen to result in significant deterioration and a loss of efficiency of the strengthening properties of the matrix (Karbhari and Engineer, 1996b). The resin matrix is the primary barrier between the structural glass fibres and the outside environment. As the first layer of defence, the resin layer degrades prior to the stress corrosion of the glass fibres. Moisture attacks the matrix layer along three routes (Mack and Holt, 1999): • capillary action along the longitudinal axis of the fibre at the fibre resin interface; • ingress through cracks and voids in the laminate structure; and • diffusion through the laminate structure. Diffusion coefficients for laminates fabricated with chopped strand mat are higher than coefficients for laminates fabricated with directional fabrics (Bank, Gentry, and Barkatt, 1995). CSM has a 18 random fibre orientation creating a higher potential for fibres creating a wicking path through a composite. Laminates made from directional fabrics also have a higher fibre content with fewer voids. Once water molecules are inside the matrix, the plasticization of the matrix begins. Bank et al. (1995) describe plasticization as the result of the interruption of the Van Der Waals bonds that occur between polymer chains, and the direct result of this is a decrease of the matrix dominated stiffness and strength properties. Increasing the fibre volume does little to delay the onset of plasticization. Studies by Buck et al. (1998) show that physical and chemical aging resulting from moisture exposure is increased by the presence of fibres with the explanation being that more water uptake is possible which further exacerbates debonding at the resin/matrix interface. Although the effects of plasticization can significantly affect the strength of the matrix, it should be noted that the effects are generally reversible on the drying of the matrix (Liao et al., 1998). Moisture damage can be detected in a number of ways. Samples of vinylester and polyester resin laminates exposed to a moist environment show the formation of a gel-like material accompanied by swelling. If degradation is allowed to continue blistering is observed followed by the eventual dissolution of the resin (Coomarasamy and Goodman, 1999). Vinylester resins consistently out performed the polyester resins in the test mentioned above. Buck et al. observed that removal of resin from the surface of laminate samples submerged in water increased with exposure time and temperature. It should be noted that moisture uptake is difficult to determine by weighing of the samples alone. Buck et al. (1998) show that the increase in weight is offset by the loss of resin so that laminate samples undergoing water absorption show at most one third of one percent increase in weight. 19 The degradation of the fibre begins once the matrix has been attacked. Unlike carbon fibres, glass fibres are especially susceptible to dissolution, stress corrosion, and pitting. Porter and Barnes (1998) provide an excellent explanation of glass dissolution. The silicon hydroxide by-product (SiOH) forms at the interface between the glass and the water. This gel layer will be less dense than the original glass structure and will transport water more readily thus accelerating the degradation process. The addition of alumina to E-glass fibres will increase the resistance of the glass to water attack; however, the presence of the alumina will decrease the resistance of the glass to alkali attack. Figure 2.5. Glass Dissolution in Water (Porter and Barnes, 1998) In order for stress corrosion to take place the fibre must be under a constant load; however, the threshold limit for stress corrosion to take place is not known precisely. Liao et al. (1998) note that chopped strand mat E-glass/polyester laminates did not fail for up to three years at 0.8 percent tensile strain (corresponding to 50 percent of the ultimate load). They also report that for E-glass/epoxy laminates no corrosion cracking threshold has been observed in distilled water. The selection of resins and the related matrix toughness has been shown to play an important role in stress corrosion by reducing stress induced micro-cracking. Vinylester resins perform on average better than polyesters. The use of silane coupling agents has also been shown to have a positive impact on delaying the stress corrosion process and minimizing the loss of mechanical properties (Liao et al. 1998). 20 Overall, moisture damage can be completely reversible, partially reversible, or permanent depending on the duration and type of exposure. In the case of plasticization, the damage is reversible upon drying of the specimen as shown previously. Specimens subject to longer exposure periods often suffer irreversible damage including damage to the fibres, matrix cracking, debonding of the fibre/matrix interphase region, and delamination due to a reduction in strength between plies caused by swelling and/or internal stress generation. 2.3.3. Alkaline Effects Although their effect on laminates is similar to the effect of moisture, alkali attack presents a great threat to the durability of FRP laminates bonded to concrete: Liao at al (1998) report that alkaline solutions have a deleterious effect on fibre morphology. Although the risk is reduced in old concrete, the mechanism of deterioration still remains. The majority of research in the area of alkali attack originates in the field of fibre reinforced concrete and glass fibre rebar (to be covered in a following section), but what has been learned in those areas can be applied to laminates bonded to concrete. Porter and Barnes (1998) provide a concise explanation of the hydroxylation dissolution of glass fibres exposed to an alkaline environment. A high pH environment is not the condition that is required for alkali attack: the combination of hydroxylation products and the presence of moisture together with the high pH acting as a catalyst provides the perfect environment for glass fibre degradation (Murphy, Zhang, and Karbhari, 1999). In the alkaline environment, the glass silicates react with the hydroxide groups of the alkalis to produce silicon hydroxide by-products along the glass interface. Hydroxylation will only occur at a pH greater than 10.0 and is rapid at first but slows as the silicon hydroxide by-product coats the surface of the glass (Porter and Barnes, 1998). Concurrent with the 21 hydroxylation reactions, calcium is leached out of the glass fibres. The calcium will combine with the water to form calcium hydroxide, which adheres to the surface of the glass further slowing the aging. Pitting and notching of the glass occurs with the growth of the calcium hydroxide crystals, and with the exposure of a fresh surface of glass the reactions are able to continue slowly reducing the cross section and leading to the embrittlement of the fibre (Porter, 1999; Porter and Barnes, 1998). By adding trace amounts of zirconium, it may be possible to improve the resistance to alkali attack; however, E-glass does not have zirconium added. — Si — O — Si — + OH" — > — Si — OH (solid) + — Si — O (solution) 1 I I I Figure 2.6. Glass Dissolution in an Alkaline Environment (Porter and Barnes, 1998) Porter and Barnes (1998) and Zhang, Ye, and Mai (1999) have shown that the degradation of FRP due to de-icing salts and sea salts follows the same mechanisms as alkali attack albeit at a slower rate. A review by Liao et al. (1998) reports of E-glass laminates that lost half of their flexural strength after 100 cycles of exposure to a sodium chloride solution. Chajes, Thomson, and Farschman (1995) show complete tensile failure in E-glass sheets bonded to concrete beams after 50 and 100 cycles of exposure to a calcium chloride environment. Their findings indicate that in the case of cracked concrete, retrofit with E-glass laminates, the reinforcement lost its effectiveness after only six months. When the same laminate was applied to intact specimens, the lifespan of effective reinforcement was extended to 15 months. It was also observed in the study that although the laminate prevented the concrete from spalling during the degradation of the concrete, there was a 22 loss of tensile capacity of the reinforced specimen due to a 36 percent drop in the tensile capacity of the fibre. It is evident that there is a fair amount of investigation into alkali attack of E-glass laminates used to strengthen concrete members. Alkali attack is one area where experience has biased the opinions of many engineers. Early failures of glass reinforced concrete specimens has led to trepidation in further implementing glass fibre laminates as concrete retrofit materials. 2.3.4. Temperature Effects It is shown that E-glass laminates are susceptible to both moisture and alkalinity. The effects of temperature serve to amplify the damage caused by both these environments, although persistent. Elevated temperatures alone have been shown to have little effect on laminate durability. Steckel et al. (1999) found that laminates exposed to 60°C dry heat showed little change in material properties after 3,000 hours of exposure. One of the few reported effects regarding negative temperature exposure involved vinylester and epoxy resins. Karbhari and Engineer (1996b) report that flexural stiffness of vinylester systems increases with exposure to a cold region type environment. It should be noted that this is a dry cold environment. Freeze-thaw testing of saturated samples can be quite detrimental to laminates bonded to concrete. Although not the focus of this thesis, freeze-thaw testing results are briefly discussed in section 2.4. Evidence that elevated temperatures act as catalysts in FRP degradation has been shown by numerous researchers. Moisture at elevated temperatures reduces the ultimate tensile strength and overall durability of the laminate far more than moisture at room temperature 23 (Buck et al., 1998). In tests comparing the effects of hot water, room temperature water, freezing water, and freeze-thaw cycles, Liao et al. (1998) report that hot water at 65°C is the worst exposure condition for laminate to concrete bond deterioration. Increased deterioration due to increased temperature can be due to increased rates of diffusion through the laminate. The effects of both moisture and alkalinity are accelerated in the presence of elevated temperatures. 2.3.5. FRP and Concrete Bond Durability It has been well documented that FRP laminates by themselves have reasonable durability, and there are many examples of aged concrete structures that are still intact. This would lead one to think that the issue of strengthening existing concrete with FRP laminate is quite trivial; however, the weakest link in the system is at this juncture: the FRP concrete bond. FRP laminates on their own are reasonably durable and concrete is reasonably durable, but the bond between these two materials requires significant investigation. Emmons, Thomas, and Vaysburd (1998) succinctly summarize the factors that control durability of the FRP-concrete bond: • tensile strength of the concrete skin; • strength of the adhesive system; • cleanliness of the concrete and FRP surfaces; • uniformity and thickness of the adhesive layer; • geometry of the exterior strengthening; • types of loads acting on the strengthened member; 24 • environmental conditions at the time of application and curing of the FRP laminate; and • workmanship and quality control. Karbhari and Engineer (1996a) list four other considerations in the adhesion of FRP laminates to concrete: mechanical interlocking, diffusion theory, electronic theory, and adsorption theory. It has been their findings that mechanical interlock is by far the dominant mechanism for composite to concrete bonding. Bizindavyi and Neale (1999) concur with findings showing that the bond strength of externally bonded plates depends mainly on surface preparation and the quality of the concrete substrate. Mechanical abrasion is the simplest way of increasing the mechanical interlock between the FRP laminate and the concrete surface. Abrasion removes the concrete latence layer and any other loose material from the bond surface. An abraded surface is obtained by sandblasting, grinding with a hand grinder, or a combination of the two methods. Prior to the application of the laminate, the surface must be cleaned using compressed dry air to remove dust generated during the abrasion procedure. It has been implied that the FRP-concrete bond is the weak link in the laminate concrete structure. Mack and Holt (1999) published a thorough list of the environmental damage mechanisms that serve to weaken and deteriorate the FRP-concrete bond: • internal influences include chemical activity, electrochemical activity, alkali content and pH level, stress, moisture infdtration, transportation of solutions and salts; 25 • interfacial influences include moisture entrapment, moisture diffusion, selective transport of chemicals, and thermal and elastic mismatch; and • external influences include humidity, moisture, temperature, temperature cycling (daily, seasonal, annually), aggressive natural and manmade agents, UV exposure, and oxygen (which may corrode internal steel in the concrete leading to expansion stresses on the bond). Electrochemical activity combined with chloride infusion occurring in the concrete may result in corrosion of the reinforcing steel. Often the corrosion of the reinforcing steel is the cause of the concrete structure's deterioration. Corrosion products require nearly three times the volume of intact steel, so when corrosion occurs in concrete, internal stresses caused by the corrosion products cause spalling of the concrete cover exposing more steel to corrosive attack. Applying FRP laminates to chloride contaminated beams is only a temporary solution. With the laminate in place, corrosion of the steel can still take place, and this leads to dilation of the concrete due to the corrosion products resulting in strain on the FRP-concrete bond and laminate, which in turn leads to accelerated aging of the laminate. Mack and Holt (1999) conclude that wrapping weak deteriorated concrete does little to improve the quality of the overall structure; however, there is some data from Soudki that indicates in fact corrosion may slow down or even stop due to an FRP wrap. Considering the available data, due diligence would dictate that a thorough repair of the corroded steel and application of corrosion inhibitors is necessary to fully rehabilitate corroded concrete structures. Moisture entrapment is considered to be one of the more critical damage mechanisms, which degrades the FRP-concrete bond. One of the benefits of bonding laminates to concrete is that the resin acts to seal the concrete surface from the ingress of water. The 26 greatest drawback to this is that the FRP is generally applied to the soffit of deteriorated structures. Concrete is a porous material with the ability to absorb moisture; however, the soffit, and often the sides, are sealed inhibiting moisture transfer so moisture collects and ponds along the FRP-concrete interface. This outward movement of moisture to the FRP-concrete bond is accelerated during freezing and thawing due to vapour pressure differences between ice and supercooled water. In their review of this problem, Mack and Holt (1999) list a number of the visual clues that indicate a moisture build up along the FRP-concrete interface: bubbles of excess moisture appear on the laminate, and there is evidence of deteriorated bonding. This problem can be avoided by proper design of the laminate-concrete interface. Steel plate bonding encountered the similar problem of moisture build-up along the concrete interface, so current practice recommends that at least 50 percent of the concrete surface be left exposed for moisture transfer. If similar guidelines are applied to FRP-concrete bonding than moisture transfer should take place and reductions in damage due to this mechanism should occur (Mack and Holt). Testing the bond strength of FRP laminates bonded to concrete has been a difficult task. Karbhari and Engineer (1996a) have performed extensive testing to determine the peel characteristics of the FRP-concrete bond using a modified peel test of their own design. One inch wide strips, 12 inches long, are peeled at a controlled angle and rate from concrete beams. Their investigations conclude that peel force is a direct measure of the work of detachment. Failure is achieved under a mixed mode of loading thereby enabling the evaluation of the interfacial fracture energy G , which can be further partitioned into the opening mode G] and sliding mode Gn. In a comparison of E-glass strips and carbon strips both bonded with epoxy resin, Karbhari and Engineer show distinct differences between the peeling behaviour of the two materials. The E-glass strip had a predominantly cohesive failure. The peeling crack propagation took place through a layer of resin with negligible 27 peeling of the concrete substrate; whereas, the carbon strips showed peeling of both materials: the resin layer and the concrete mortar substrate. 2.4. Previous Research 2.4.1. FRP Laminates Bonded To Concrete In the fifteen years since Urs Meier first began his series of tests on strengthening a pedestrian bridge with bonded CFRP strips, there has been a constantly increasing body of work reporting the investigations of various researchers and their discoveries in the field of strengthening laminates with FRP material. The following sections will briefly present some of these findings. It is meant to be simply an overview of the research that has currently taken place. To this date there are few firm conclusions that can be drawn from the body of knowledge; however, as Debaiky, Green, and Hope (1999) remind us there has been a small amount of work done and because of the considerable differences in specimen sizes and applications neither solid conclusions, nor general trends can be established for the rehabilitation of corroded concrete structures. Conclusions that have been drawn from previous research are listed below. • The application of FRP changes the failure mode of the strengthened specimen from ductile to brittle (Arduini and Nanni, 1997). • Failure mechanisms for concrete strengthened with FRP include FRP rupture, concrete crushing, shear failure, and peeling at the adhesive-concrete interface (Arduini and Nanni). 28 • Sandblasting is slightly more effective than grinding at improving the mechanical interlock of the FRP-concrete bond with sandblasted specimens showing higher loads and deflections; however, surface preparation will not change the failure mechanism (Arduini and Nanni). • The percentage increase in the ultimate strength of concrete specimens strengthened with an FRP laminate is directly proportional to the number of longitudinal layers of fabric present in the laminate; however, the increase in strength is will only happen if the design mode of failure is achieved (GangaRao and Vijay, 1998). 2.4.1.1. Small Scale Specimen Tests Steckel et al. (1999) performed a series of exposure tests on a number of different laminate systems as part of an initiative launched in 1995, by the California Department of Transportation (Caltrans), for the evaluation and qualification of advanced composite materials for seismic retrofit and rehabilitation. This initiative evaluated the performance of laminate samples subject to a number of different environmental exposures: • 100 percent humidity at 38°C for 1,000, 3,000, and 10,000 hours; • immersion in a salt-water solution at 23°C solution for 1,000, 3,000, and 10,000 hours; • immersion in an alkali solution of calcium carbonate with a pH of 9.5 at 23°C for 1,000, 3,000, and 10,000 hours; • exposure to 60°C dry heat for 1,000 and 3,000 hours; 29 • cycling between ultraviolet light exposure at 60°C and moisture at 40°C for 1000 cycles with four hours of exposure per condition; and • freeze-thaw cycling between 100% humidity at 38°C and a freezer at -18°C for 20 cycles with approximately 15 hours in the humidity chamber and 9 hours in the freezer. The laminate samples tested by Steckel et al. were E-glass/epoxy, carbon/epoxy, E-glass/polyester, and E-glass/vinylester provided by nine different manufactures. At the conclusion of the exposure periods, tensile testing was performed on the laminate coupons. Testing results showed a number of interesting trends. No significant reduction (>5%) in Young's Modulus was measured for any system after all the test durations and exposures. All of the E-glass systems showed some degree of susceptibility to tensile strength degradation from long term moisture exposure with one laminate system losing 35 percent of its strength after 10,000 hours of exposure. The findings of Steckel et al. also show a near complete recovery of strength lost to moisture after the specimens were allowed to dry out. This phenomenon has been previously discussed in §2.3.2. The most interesting result was the performance of the E-glass/polyester and E-glass/vinylester laminates compared to the E-glass/epoxy laminate. In the moisture exposure tests the E-glass/polyester and E-glass/vinylester laminates had an average tensile strength reduction of 20 percent; however, the E-glass/epoxy laminate had a tensile strength reduction of 35 percent. The explanation for the poor performance by the E-glass/epoxy laminate was attributed to higher measured moisture absorption, sourcing the E-glass from a different manufacturer, and using a different weave of glass in the laminate. The E-glass/polyester and E-glass/vinylester laminates were fabricated using a unidirectional fabric; whereas, the E-glass/epoxy laminate used a woven fabric with the fibres oriented at 0° and 90°. No significant results were obtained from exposure to the other environments, and it is significant to note that no degrading chemical effects were attributed to exposure to the salt solution or to the alkaline solution with pH9.5. Different laminate configurations have been the focus of many researchers in an attempt to fully utilise the properties of FRP laminates. Spadea, Bencardino, and Swamy, (1998) report explosive debonding of the FRP sheet bonded to the tension face of the test samples. They conclude that a CFRP plate laminated in the tension zone is inadequate to ensure composite action until failure, and beams with this laminate configuration show a significant deficiency in deflection and rotational capability. In other words, without some sort of mechanical anchorage, the full potential of the CFRP laminate is not seen. Mukhopadhyaya, Swamy, and Lynsdale (1998) show similar results in a comparison of steel plates and E-glass/isophthalic polyester resin plates bonded to the tension faces of concrete beams using epoxy resin. All of the beams tested failed by the concrete crushing prior to the failure of the GFRP plate indicating that entire strengthening potential of the GFRP was not realised. Even with its anchorage enhanced by bolting the GFRP plate still debonded at an early age. Overall, the GFRP plates increased the ultimate strength of the strengthened beam by 15 percent over its steel plate counterpart. Chajes et al. (1995) explored the effects of laminating an FRP sheet designed to have the same yield strength as a No. 3 steel bar (approximately equivalent to an Ml 0 bar) to concrete samples using a variety of reinforcement materials: E-glass, carbon fibre, aramid fibre, and epoxy resin. The laminate was applied using the vacuum bag 31 method.' After application of the laminate, the specimens were subjected to freeze-thaw and wet-dry cycling. Of the three fibres tested, E-glass and aramid lost 50 percent of the strength increase shown by the room temperature controls. The carbon fibres lost approximately 15 percent of the strength increase. In the final comparison of the test results, wet-dry cycling was shown to be more detrimental to the strengthened beams than the freeze-thaw cycling. Debonding occurred in all of the exposed beams indicating a deterioration of the FRP-concrete bond due to the exposure conditions: debonding was not a problem for the room temperature control samples. Toutanji and El-Korchi (1999) report similar findings in their study of cylinders wrapped with E-glass/epoxy and carbon/epoxy laminates. Mortar cylinders 16 mm in diameter by 120 mm in length were cast and wrapped with a single ply of laminate. Two exposure conditions were investigated: wet-dry cycling with seawater and freeze-thaw cycling. The E-glass/epoxy specimens exposed to seawater cycling showed the greatest degradation with tensile strength reductions in the order of 20 percent. The strength reductions for the freeze-thaw cycling were in the order of 10 percent. The specimens strengthened with carbon fibre showed little deterioration after exposure to the environments. Before the research of Toutanji and El-Korchi (1999), Toutanji and Balaguru (1998) published their findings based on tests involving cylinders strengthened with FRP tow sheets. Their study consisted of testing concrete cylinders 76 mm in diameter by 1 A plastic bag envelops the laminate and concrete and the air is removed using a vacuum to ensure complete saturation of the fibre by the resin. The vacuum bag is an efficient method of curing; however, its application is limited to objects with a geometry that allows full encapsulation by the vacuum bag. The use of vacuum bagging in full size structures is impractical and rarely, if ever, used. 32 305 mm in length iri compression. Strength increases based on cylinder strengthening using E-glass/epoxy wrapping systems and carbon/epoxy systems were investigated. The wrapped cylinders were exposed to salt water, wet-dry and freeze-thaw cycling. At the conclusion of the exposure period, 300 cycles, the cylinders were tested in compression. The carbon/epoxy system was shown to be superior to the E-glass/epoxy system in the harsh environments investigated. Wet-dry cycling had little effect on the stiffness of the GFRP wrapped cylinder; however, exposure caused a 10 percent reduction in strength and a 20 percent loss of ductility. Freeze-thaw exposure proved to be detrimental to both the CFRP and GFRP systems with the GFRP system showing greater deterioration. Although the freeze-thaw cycling had little effect on the stiffness of the cylinders, the strength losses were significant. The GFRP specimens lost 28 percent of their original strength, the CFRP specimens lost 19 percent. In terms of ductility, the GFRP specimens had losses greater than 65 percent and the CFRP specimens had losses in the range of 30 percent due to the freeze-thaw exposure. Toutanji and Gomez (1997) investigated the effect of saltwater wet-dry cycling on small concrete beams. The beams had the dimensions 51 mm in width by 51 mm in depth by 356 mm in length. Different glass and carbon tow sheets were applied using a selection of epoxy resins, and the strengthened beams were subject to 300 wet-dry cycles. At the completion of the loading, the beams were tested in third point loading until failure. It was shown that the epoxy selection had a significant effect on the performance of the FRP strip: one type of E-glass tow sheets outperformed the carbon tow sheets when used in conjunction with a certain epoxy. Overall there was a reduction in performance of the exposed beams compared to those beams that were left at room temperature. It should be noted that the epoxy-33 concrete interface suffered the most deterioration from the wet-dry cycling. Toutanji and Gomez base this on the fact that no fibres were observed to fail during the loading tests. Ye, Zhang, and Mai (1998) and Zhang, Ye, and Mai (1999) present the results of the effects of saline immersion on concrete cylinders, 150 mm in diameter by 300 mm in height, strengthened with an E-glass/vinylester resin system. Different E-glass fabric systems were applied to the concrete columns prior to the exposure to the saline solution. Woven roving, chopped strand mat, and unidirectional glass fibre tape were the three different fabric systems investigated. Ye et al. show that a concrete cylinder strengthened with a combination of all three of the fabric systems shows the greatest increase in performance. The addition of chopped strand mat into woven roving layers increased the load carrying capacity by 17.5 percent and increased ductility by 53 percent. Zhang et al. provide results, which follow a similar trend. Concrete cylinders, 150 mm in diameter by 300 mm in height, with an E-glass/vinylester wrap were exposed to a saline solution for eight months. Two different wrap configurations were investigated: four plies of woven roving and alternating plies of woven roving and chopped strand mat (four plies in total). After removal from the saline solution, the cylinders were tested in compression until failure. It was shown that immersion in a saline solution leads to saturation of the laminate, which ultimately leads to a reduction in the laminates strengthening efficiency; however, reduction in strengthening efficiency is dependent on the laminate used. Zhang et al. found that the woven roving/chopped strand mat showed a 21 percent reduction in load carrying capacity compared to the room temperature, wrapped, control specimens; whereas, the woven roving laminate had a 26 percent reduction in capacity compared to the control specimens. In terms of overall 34 strengthening efficiency the woven roving laminate subjected to the saline immersion showed a 41 percent increase over the plain concrete control, and the woven roving/chopped strand mat hybrid had a 57 percent increase. In both laminate scenarios, saline immersion did not affect the stiffness of the strengthened samples. Both Ye et al. and Zhang et al. conclude that the use of woven roving, chopped strand mat, and vinylester resin is a cost effective method of restoring and increasing the load carrying capacity of concrete columns. 2.4.1.2. Large Scale Specimen Tests and Applications In North America, there is a limited amount of large scale testing of FRP strengthened concrete members. There are three reasons for this: • Industry is hesitant to deviate from the status quo. The use of FRP materials is still perceived as a risk, and as a result, engineers are hesitant to use them. If a structure is in need of rehabilitation most designers feel more comfortable with tried and true methods and would rather use them than try something new that is still in the test phase; • It is expensive for universities to cast and test full scale specimens. The amount of energy required to prepare and test one full scale specimen can be redirected to the preparation and testing of a number of smaller specimens. The data from one full-scale test is statistically questionable; whereas, a number of small-scale tests produce a number of comparable results, so conclusions based on statistical truths may be made; and • For in-situ applications, it is difficult to verify the increase in strength provided by the FRP system. A theoretical strength prediction may be made, but the true value cannot be determined easily with non-destructive testing methods. Nanni and Gold, (1998) discuss a rapid load test; however, this test is not widely used. In spite of these issues there are a number of large specimen tests under way in North America. Please note that large scale testing at the University of British Columbia will be covered in section 2.4.3. Neale and Labossiere (1998) have completed the following large-scale FRP repair projects. At the University of Sherbrooke, columns damaged by the corrosion of the internal steel were repaired using an s-glass/epoxy wrap. Overpass columns that were part of Quebec's highway 10 were badly damaged by concrete spalling due to the corrosion of the reinforcing steel. Five of the columns were wrapped with a glass/epoxy laminate, four with a carbon/epoxy laminate, and three were repaired with conventional methods. Two other projects carried out by Neale and Labossiere include the strengthening of a concrete pier supporting the Champlain Bridge in Montreal, Quebec and the rehabilitation of the Webster Parkade in Sherbrooke, Quebec. Carbon and glass were used to strengthen and upgrade both of these concrete structures. The presentation of long term results are still pending, but it is believed that data provided by the previously discussed repairs has and will serve a number of purposes (Neale and Labossiere): • it has familiarized the local contractors and workers with the installation of composite wraps; • it has shown that high costs of the composite materials can be offset by reduced labour costs. In the case of the Highway 10 overpass columns, the 36 number of person hours required for the installation of the composites was reduced, and the traffic flow across and under the overpass was not interrupted; • in-situ tests of beams strengthened in the Webster Parkade have shown a 15 percent increase in bending strength, and a 20 percent increase in shear strength; • it will provide additional information regarding the durability of rehabilitation using composite materials in severe climatic conditions. The waterproof characteristics of the composite wraps are expected to arrest the corrosion of the remaining steel reinforcement; and • integrated sensors and fibre optics will monitor column deformations and infer the state of the reinforcing steel degradation. This will help to provide feedback on the performance of the wrap, as well as providing feedback on the performance of the sensors. Seible, Priestley, Hegemier, and Innamorato (1997) in conjunction with the Federal Highway Administration (FHWA) and Caltrans investigated the behaviour of reinforced concrete columns strengthened with carbon fibre, s-glass, and E-glass jackets. Full-scale bridge columns were cast and retrofit with the composite jackets. The columns were then subject to lateral loads simulating the seismic effects of an earthquake. In comparison to steel jackets, the carbon jackets met or exceeded the deformation capacities obtained by their steel jacket counterparts; however, the stiffness increases were less for the composite jackets. The results obtained by Seible et al. show that composite jackets can provide the required short-term structural effectiveness; however, they stress that reduction factors and/or increased 37 safety factors should be considered to account for uncertainties in the materials, lay up systems, jacket curing, and durability. Shehata, Rizkalla, and Stewart (1998) performed a retrofit of the roof of a Winnipeg, Manitoba water pollution treatment facility. Prior to the application of the composite to the roof, they had the opportunity to remove a number of concrete panels from the roof, retrofit them, and then test the strengthened panels to failure to obtain their strengthened characteristics. Carbon fibre strips bonded with an epoxy mortar to the concrete panels were demonstrated to provide the additional strength required for the retrofit. The strips were then applied to the panels that comprised the roof structure of the pollution treatment facility. The construction time to strengthen the roof took only four days with no interruption to the facilities operations. 2.4.2. FRP Reinforcing bars Extensive research has been conducted in the area of FRP reinforcing bars and their applications to strengthening cast concrete. The search for a non-corroding reinforcing bar has been on going since it was discovered that the corrosion of steel reinforcing bars is the major cause of much of the deterioration of the world's concrete infrastructure. FRP reinforcing bars have been the focus of a lot of this research, and an offshoot of this is insight into the behaviour of FRP in a concrete environment. The research is also helping to pin point the knowledge requirements of industry regarding composites use in civil engineering applications: in many cases the information requirements for FRP reinforcing bars are similar to those for FRP strengthening laminates. 38 The greatest obstacle preventing extensive use of FRP reinforcing bars, this applies to FRP strengthening schemes as well, is the lack of predictable material properties. Civil engineering materials are, for the most part, simple, predictable and consistent with well understood properties. Composites, however, are complex materials. A designer has many options when it comes to choosing a reinforcing bar: resin, fibre material, fibre type, fibre thickness, number of plies, and laminate thickness to name a few. There is no standard laminate. Taly and GangaRao (1999) provide a summary of the issues affecting the material behaviour of FRP reinforcing bars: • each bar depends on the fibre/resin volume, the bar size, and the method of production; • commercially available bars all have different and distinct durability and strength properties; • unlike steel, which has consistent minimum ultimate and minimum yield strengths, FRP reinforcing bars' strengths and moduli vary from producer to producer; • different reinforcement fibres have different thermal properties in the longitudinal and transverse directions. The thermal properties of GFRP are the most similar to the properties of steel. Aiello (1999) has found that FRP-concrete thermal compatibility is still very low, and thermal loading produces a state of stress at the interface between FRP reinforcing bars and concrete; and • different manufacturers provide different surface textures with no industry standard. 39 Although there are complications associated with the consistency of the material properties, Japan has included FRP reinforcing bars in their design codes since 1993 (Childs, 1999). The new revised Canadian Highway Bridge Design Code also contains a chapter on the FRP reinforcement in bridge decks (CAN/CSA, 2001). Like all composites, the two components in FRP reinforcing bars are the fibre and the resin matrix, and as has been written previously, these are subject to degradation in an alkaline environment. What follows is a brief synopsis of the research investigations into the alkaline attack on FRP reinforcement bars. The first section will focus on the fibres performance, and the second section will look at the matrix performance. 2.4.2.1. FRP Reinforcement Bars - Fibre Performance Katsuki and Uomoto (1995) investigated the performance of reinforcing bars of different fibre types (E-glass, carbon, and aramid) in a vinylester matrix submerged in a sodium hydroxide solution at a constant temperature of 40°C. Their findings show that only the E-glass bars were subjected to deterioration by the alkali solutions. The performance of the E-glass reinforced bars was far below that of the aramid and carbon fibre reinforced bars, which showed little deterioration and no loss of strength even after being submerged in a solution with twice the concentration of alkalis. The examination of cross sections of the glass reinforced bars showed the radial penetration of the alkalis into the structure of the bar over time. The penetration of the alkalis attacked the outer fibres first leaving the inner core of the fibres untouched by the alkalis. This was confirmed by the glass reinforced bars have a lower ultimate strength (due to the loss of cross section), but having the same Young's Modulus (the inner bars were intact and had the same 40 stress-strain relationship). After 120 days of exposure, Katsuki and Uomoto found that the reinforcing bars reinforced with E-glass fibre had strength decreases of more than 50 percent. Porter and Barnes (1998) report inconclusive findings in a study of reinforcing bars constructed from E-glass fibres and a vinylester resin. The overall trend showed the fibres as the critical component of the reinforcing bar; however, due to considerable scatter in the results there were few conclusions. There is evidence that E-glass is not the ideal material for FRP reinforcement for severely alkaline environments. One possible solution to this might be alkaline resistant (AR) glass (Clarke, 1999). This type of glass was developed especially for use in glass fibre reinforced concrete (GFRC) when researchers noticed the breakdown of ordinary glass fibres in the concrete environment. Currently, AR-glass is not widely used as a reinforcement material in FRP; however, there exists the possibility for its use. 2.4.2.2. FRP Reinforcement Bars - Matrix Performance The matrix is one of the most easily manipulated variables in an FRP system. Once the fibre is selected, it is possible to experiment with an almost infinite number of resin variations. What follows is a very brief overview of some of the previous work conducted using E-glass fibre and polyester and vinylester resins as the composite materials for FRP reinforcing bars. 41 Coomarasamy and Goodman (1999) investigated the effects of reinforcing bars in a simulated concrete pore water solution. The solution consisted of potassium hydroxide, sodium hydroxide, and saturated calcium hydroxide maintained at a temperature of 60°C. With a solution pH of 13.5, it was their intent to mimic the interior environment of freshly cast concrete. FRP reinforcing bars produced by different manufacturers with different resin materials, polyester and vinylester, were submerged in the solution for a period of six months to investigate their performance. Within 11 weeks of exposure, the E-glass/polyester samples had lost 80 percent of their initial strength compared to the 45 percent loss of strength experienced by the E-glass/vinylester specimens. There were also noticeable decreases in the stiffness of both the E-glass/polyester and E-glass/vinylester samples. The overall trends indicated that the E-glass/vinylester samples had better durability; however, there was also an indication that the manufacturing of the bars played a role in their overall performance. The results of others partially substantiate the findings of Coomarasamy and Goodman (1999). In studies investigating the performance of reinforcing bars with AR-glass in either a polyester or a vinylester resin, Tannous and Saadatmanesh (1999) found that the vinylester matrix provided better protection in a marine environment. This was a follow up to their research published in 1998 that details a study of reinforcing bars with E-glass/polyester resin and E-glass/vinylester resin submerged in alkaline solutions, de-icing salt solutions, and sea water solutions (Tannous and Saadatmanesh, 1998). Their findings show that exposure to de-icing salts and alkaline solutions were responsible for the greatest losses in strength. They also note that exposure to de-icing salts induces greater losses in strength than the equivalent exposure to seawater. In all cases, the reinforcing bars with the vinylester 42 resin retained more of their original strength than their polyester companions. To further validate the findings of Katsuki and Uomoto (1995) no significant changes in the Young's Modulus were recorded in any of their tests (Tannous and Saadatmanesh, 1998). Porter (1999) presented the results of tests at Iowa State University comparing the performance of E-glass/isophthalic polyester reinforcing bars with E-glass/vinylester reinforcing bars submerged in a 60°C solution with a pH of 12.8. The specimens were provided by two different bar manufacturers. After the accelerated equivalent of 50 years of exposure, the results showed that all specimens were subject to significant strength losses after accelerated aging. One set of specimens with an E-glass/isophthalic polyester resin has strength losses of 48 percent, another set of E-glass/isophthalic polyester resin specimens had strength losses of 66 percent, and the E-glass/vinylester resin specimens had losses of 65 percent. Not only do the results indicate deterioration due to exposure, but they also show that the bar manufacturer can have as great an influence on bar durability as the materials themselves. 2.4.3. Research at the University of British Columbia A report by Banthia, Yan, and Nandakumar (ND) outlines the first use of the FRP spray-up application method of concrete strengthening. Small concrete beams (100 mm in width by 100 mm in height, by 350 mm in length) were cast. After curing, specimens were lightly sandblasted and coated with a 3 mm coating of 9.5 mm long E-glass fibre and a blend of polyurethane and polyester resin. The fibre content of the coating was 8 percent by volume. After the laminate had cured, the specimens were tested in third point loading to obtain the load deflection curve. ASTM Cl018 (1989) was used as the testing protocol. 43 The beams strengthened by the laminate showed a 36 percent increase in the ultimate load carrying capacity. The fracture energy of the strengthened beams was calculated, and this showed that the beam strengthened with the FRP absorbed 62 times the fracture energy of the control beam. It was observed in all of the tests that the laminate remained bonded to the concrete throughout the duration of the test. The use of the spray-up method to apply FRP was shown to be a viable method for increasing the absorbed fracture energy and the load carrying capacity of concrete specimens. Following up on the research of Banthia et al., Boyd and Banthia (1999) further investigated the use of the spray-up process of GFRP to strengthen shear deficient concrete beams. Concrete beams (96 mm in width by 125 mm in height) were designed to be deficient in shear. E-glass fibre and polyester resin were used to create a 3.5 mm thick laminate with a fibre content of 20 percent by volume. In order to enhance the concrete laminate bond a vinylester resin was used to prime the surface of the concrete before the application of the laminate. After the application of the GFRP laminate, the strengthened beams were tested in third point loading. Three laminate configurations were investigated: laminate on the sides, laminate on the bottom and the sides, and laminate on all four sides that completely encapsulated the beam. During testing, the specimens with the laminate bonded to their sides failed due to the laminate debonding, specimens with the laminate on the sides and bottom failed with the crushing of the concrete, and specimens encapsulated with the laminate failed due to fibre failure. It should be noted that complete encapsulation is an unrealistic method of strengthening a concrete specimen. Analysis of the data showed good correlation with predicted values calculated using the Canadian Standards Association guidelines; however, the failure mode was difficult to predict. Boyd and Banthia showed the dramatic increase in fracture energy ranging from 307 percent, for the 44 laminates on the sides of the beam, to 3788 percent, for the encapsulated beam: further proof of the effectiveness of the FRP spray. During the summer of 1999, two full-scale concrete bridge girders were strengthened and tested until failure in the UBC labs. These beams were supplied to UBC by the Ministry of Transportation and Highways (MoTH) for British Columbia. The beams, having been removed from the Neale Bridge in British Columbia earmarked for demolition, were representative of a number of the bridges in the lower mainland and showed signs of concrete spalling and reinforcing steel deterioration. It was also believed that the bridges were deficient in shear. Two methods of laminate repair, GFRP wrap and GFRP spray were used on the beams in a direct comparison. Figure 2.7. Typical Bridge in British Columbia's Lower Mainland 4 5 Figure 2.8. Typical Reinforcing Steel Corrosion and Concrete Spalling Before the laminates were applied to the beams, the missing concrete was replaced with a nonshrink grout. The grout was allowed to cure for four days before the surfaces of the beams were sandblasted and the laminate was applied. The wrap system used was supplied by Master Builders Technologies, and consisted of an E-glass unidirectional fabric and an epoxy primer and resin. The spray-up laminate used an E-glass fibre 48 mm in length and a polyester resin and vinylester primer. The final wrap laminate had a thickness of approximately 3 mm with a fibre content of approximately 30-35 percent by weight. The spray-up laminate had a thickness of approximately 10 mm with a fibre content of approximately 20 percent by weight. Testing of the samples took place using four hydraulic jacks which loaded the beam in third point loading. Table 2.1. shows a summary of the results from these tests. It is obvious that the spray-up laminate is comparable in performance, if not superior, to the wrap. The questions concerning the strengthening potential for sprayed-up FRP had been addressed; however, there were still questions remaining surrounding the durability of the sprayed-up laminate. 46 These questions will be partially answered in the experimental report contained in the following pages. Table 2.1. Summary of Results from UBC MoTH Beam Tests (Ross, 2000) Beam Configuration Ultimate Load, kN Midspan Deflection, mm Initial Stiffness, Nmm2 Control 237 114 3.52E+13 FRP - GFRP Wrap 323 143 4.16E+13 FRP - GFRP Spray 470 128 4.97E+13 47 3.0 E X P E R I M E N T A L P R O C E D U R E 3.1. Introduction The experimental procedure followed for the investigation of the durability characteristics of concrete strengthened with FRP mat is outlined below. One of the advantages of the hand lay-up is the simplicity of the application operation. A minimal amount of equipment and skill is required to produce a laminate of acceptable quality. What follows is a brief description of the materials used and their properties, the procedures followed in preparation of the samples, and an explanation of the testing protocol. 3.2. Material Properties The material properties of the FRP laminate and its components presented in this section are specific to the experimental procedure followed in this investigation. For a general, more detailed investigation, please refer to Section 2.2. Fibre Reinforced Plastics. 3.2.1. Concrete Table 3.1. Concrete Mix Design for 45 MPa Concrete - Ratios Contents Ratio Description Water 0.50 Tap water Cement 1.00 Type 10 Portland Cement Fine 2.00 Fineness Modulus 2.5 Coarse 3.50 Maximum size 3/8 in. The concrete mix design selected for the concrete beams was designed to provide 45 MPa concrete in 28 days. The components of the mix are shown in Table 3.1. The specimens were cast using nine different concrete batches. The total volume of each batch was 48 limited by the capacity of the pan mixer. The mixed volume of each batch produced enough concrete to cast 16 beams (four beams from each mix were retained as replacements if any specimens proved faulty) and three cylinders. After casting the specimens were stored in a water bath saturated with lime (calcium hydroxide) for 28 days. Three concrete cylinders were cast for each batch. The dimensions of these cylinders were 100 mm diameter by 200 mm in height. After 28 days of curing the top and bottom of the cylinders were ground smooth, in lieu of capping, to provide a defect free loading surface. The cylinders were tested in accordance with ASTM C39-86 (Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens). Testing took place using an Amsler hydraulic compression machine with a capacity of 200,000 lbf. The 28-day strength for the nine batches had an average of 49.0 MPa with a standard deviation of 4.95%. 3.2.2. Glass Fibre Owens Coming's Advantex Glass Fibre Chopped Strand Mat was the glass reinforcement selected for the laminate. This fabric is produced using the Advantex® glass fibre: an E-glass fibre (the same fibre supplied in bobbins for use in the spray-up process). The fabric is composed of fibres ranging in length from one to two inches (25.4 to 50.8 mm) with a highly soluble silane binder used to bond the fibres together prior to application of the resin. The Advantex fibre is the same fibre that was used previously at the University of British Columbia in the investigation of the spray-up method of strengthening concrete. By using the same fibre in a chopped strand mat fabric form it was possible to guarantee that laminates with consistent fibre content were used to strengthen the concrete specimens: this eliminated one of the variables that is associated with the spray-up process. 49 Table 3.2. Typical Room Temperature Properties of Advantex Glass Fibre (Supplied by Owens Corning - Refer to Appendix A for a copy of the faxed data sheet) Tensile Strength 3100-3800 (MPa) 450-550 (ksi) Tensile Modulus 80-81 (GPa) 11.6-11.7 (psix 10") Elongation, % 4.6 4.6 3.2.3. Vinylester Resin Dow Chemical's Derakane XD 8084, an elastomer modified vinylester resin, was selected as the primer resin used to prime the surface of the concrete before the application of the structural laminate. The advantages of vinylester resin have been discussed previously in Section 2.2.4. The specific benefits that can be associated with Derakane XD 8084 include an increased toughness (due to elastomer modification), increased adhesive strength, and superior resistance to mechanical strength. Vinylester resin would make an ideal resin for the entire structure; however, the aim of this study was to find a cost effective alternative to expensive epoxies. Table 3.3. Typical Room Temperature Properties of 1/8" Clear Castings made with Derakane XD 8084 Resin (Dow Plastics, 1999) Tensile Strength 69-76 (MPa) 10-11,000 (psi) Tensile Modulus 3.5 (GPa) 5.1 (psix 105) Elongation, % 10.0 10.0 Flexural Strength 110-124 (MPa) 16-18,000 (psi) Flexural Modulus 3.0 (GPa) 4.4 (psix 105) 1.0% vol./wt. Lupersol DDM-9 MEK P Catalyst. 50 3.2.4. Polyester Resin Ashland Chemical's Aropol K1907 was selected as the resin matrix for the laminate used in strengthening the concrete. This resin is an isophthalic polyester designed for room temperature curing. Its beneficial characteristics include excellent corrosion resistance, excellent wet out and roll out (ease of application), and high stiffness. Table 3.4. Typical Room Temperature Properties of Clear Castings made with Aropol K1907 Resin (Ashland Canada Inc., 1998) Tensile Strength 75.8 (MPa) 11,000 (psi) Tensile Modulus 3.78 (GPa) 5.48 (psi x 10') Elongation, % 2.43 2.43 Flexural Strength 112 (MPa) 16,300 (psi) Flexural Modulus 4.00 (GPa) 5.81 (psi x 10") 1.0% vol./wt. Lupersol DDM-9 MEKP Catalyst. Table 3.5. Typical Room Temperature Properties of Laminatef made with Aropol K1907 Resin (Ashland Canada Inc., 1998) Tensile Strength 59.5 (MPa) 8,640 (psi) Tensile Modulus 9.17 (GPa) 1.33 (psi x 10") Flexural Strength 117 (MPa) 17,000 (psi) Flexural Modulus 8.55 (GPa) 1.24 (psi x 10") 1.0% vol./wt. Lupersol DDM-9 MEKP Catalyst. fLaminate construction: 3 ply 450 g/m2 (l'/2 oz.) chop mat. Glass content 28% by weight. 3.2.5. Catalyst and Promoters The catalyst used to cure all resins was the industry standard Lupersol DDM9 Methyl Ethyl Ketone Peroxide (MEKP) recommended by both resin manufacturers. 12%-cobalt napthenate (CoNap) and dimethylaniline (DMA) were used to promote the resin. These materials were supplied by Ashland Canada. 51 3.2.6. Post-Tensioning Steel Based on the equipment available and the small size of each sample, it was decided that the specimens would be post-tensioned. The post-tensioning bars selected were 5/8 inch (15 mm) diameter Dywidag bars cut to 700mm in length. These bars are used extensively in post-tensioning operations where threaded bar is required. Dywidag Systems manufactured the nuts required for locking off the loaded bar, and standard  lA inch (6.5 mm) steel plate was used as the backing plate between the nut and the concrete surface. Table 3.6 provides the engineering properties of the Dywidag bars used in the post-tensioning operation. Table 3.6. Dywidag Threadbars - Technical Data (Dywidag Systems International, 1998) Nominal Diameter 15 (mm) 5/8 (in.) Steel Area 177 (mm") 0.27 (in.2) Yield Strength 900 (MPa) 130 (ksi) Ultimate Strength 1100 (MPa) 160 (ksi) Yield Load 159 (kN) 35.1 (kip) Ultimate Load 195(kN) 43.2 (kip) Modulus of Elasticity 204,500 (MPa) 29,700 (ksi) 3.3. Preparation of Specimens 3.3.1. Casting of the Concrete Specimens Each specimen had the following dimension: 100 mm high x 100 mm wide x 350 mm long. Eight Plexiglas moulds with two beams per mould were used to cast the beams. Based on the number of moulds it was possible to cast 16 specimens at a time. A thin layer of form oil was applied to each mould before each batch was cast. The oil would act as a 52 mould release and facilitate the removal of the beams from the moulds after the initial 24-hour curing period. Concrete was added to the moulds in three layers with vibration between each concrete layer. The final surface of the specimens was given a trowel finish before being set aside to cure. The freshly cast beams were covered with a layer of impermeable polyethylene film for 24 hours. After 24 hours had passed the beams were of sufficient strength to be demoulded and placed in the curing baths. The curing bath was a solution of water saturated with powdered lime (calcium hydroxide). At the conclusion of the 28-day curing period 36 beams were separated and placed in three categories for each curing period. 12 beams would be left as control samples for their respective environments, and 24 beams were selected for FRP application, 12 of which would be post-tensioned following the FRP cure. Please refer to Table 3.7. for the breakdown of the test program. 3.3.1.1. Preparing the Prestressing Duct In order to develop a low-level bond stress it was necessary to prestress some of the concrete specimens. Based on the resources available it was decided that the prestressing would be achieved by post-tensioning the sample using threaded Dywidag bar. This decision required that all specimens were cast with a small diameter duct in the appropriate location. Schedule 10 PVC pipe, % inches (19 mm) in diameter was selected as the duct material. Each duct was cut to the appropriate length. Wooden supports were placed in each of the moulds to ensure that the duct would be cast at the appropriate depth. Early casting attempts resulted in the PVC ducts floating to the surface during vibration. In order to compensate for their lack of density the ducts were weighed down with a 10 mm steel reinforcing bar cut to 53 the correct length. The remaining void in the duct was filled with sand. Each end of the duct was sealed with duct tape, and the concrete was poured in the mould. After demoulding, it was possible to chip away the thin, hardened, paste of concrete covering each end of the tube, cut away the duct tape, and remove the steel bar and sand creating a % inch (19 mm) duct running the length of the pipe. 3.3.2. Application of the FRP Wrap For each exposure period, 24 samples required the application of an FRP wrap. The wrapping procedure consisted of the preparation of the samples, application of the primer coat, and the application of two plies of 450 g/m2 (P/2 oz.) chopped strand mat. The specimens were allowed to cure for a minimum of 24 hours after which the rough edges of the FRP laminate were ground smooth and catalyzed resin was applied to the exposed fibres on those specimens that were to be immersed in the aqueous environments. Another 24 hour curing period followed the final application of resin. The sample face to which the FRP was to be applied was not a mould surface; and therefore required both grinding, to remove large surface protrusions (> 2mm), and sandblasting, to remove any concrete laitance that had not been removed by grinding. After sandblasting and grinding was complete the dust from the specimens was removed using compressed air. The primer coat consisted of a thin layer of catalyzed resin. The resin selected for this was Derakane XD 8084. The properties can be found in Section 3.2.3. The resin was promoted with 0.2 percent (by weight) 12%-cobalt napthenate (CoNap) and 0.2 percent (by weight) dimethylaniline (DMA). Prior to use, the resin was catalyzed with 2 percent (by weight) 54 Methyl Ethyl Ketone Peroxide (MEKP), trade name Lupersol DDM9, to give a gel time of 30 minutes. The primer was applied with a brush with coverage of 4.9 m2/L. The primer was allowed to cure for one hour before the application of the polyester resin and CSM structural laminate. The laminate was applied in four stages. The Dow Chemical Company publishes a brochure entitled Fabricating Tips that provides thorough instructions on the construction of an FRP laminate (Dow Plastics, 1994). First, a catalyzed coat of the polyester resin, Aropol Kl 907, was applied using a brush or roller with a coverage of 0.65m2/L. The properties of Aropol K1907 are outlined in §3.2.4. The resin was promoted by the supplier with CoNap and DMA, and catalyzed with 1.5% MEKP (by weight) to give a gel time between 20 and 30 minutes. A pre-cut piece of CSM was applied to the liquid resin and rolled out using a serrated roller. The remaining resin was applied and rolled out onto the fabric using the serrated roller. It was ensured that the glass had become saturated with resin before the second ply of fabric was applied. Saturation had occurred when the glass mat had become translucent in appearance and there were no air voids visible. The application of the second ply followed the same steps as the first ply. Catalyzed polyester resin was applied with coverage of 0.65 m2/L. The second ply of CSM was applied and rolled out using a serrated roller. A final topcoat of catalyzed resin was applied with coverage of 0.65 m2/L. The laminate was allowed to cure for a minimum period of 24 hours before proceeding. After curing of the laminate had occurred, the rough edges of the laminate were ground smooth using a hand grinder. This procedure was strictly aesthetic. A thin coat of catalyzed Aropol K1907 polyester resin was applied to the freshly ground surface of the laminate for those samples that would be immersed in an aqueous environment. The 55 application of the resin would prevent the premature exposure of the glass fibre possibly deleterious conditions. Glass fibres when exposed to aqueous conditions have been shown to provide a wicking effect drawing liquids into the laminate causing premature failure (Mack and Holt, 1999). After this final application of resin, there was another 24-hour curing period before post-tensioning and exposure. 3.3.3. Post-Tensioning of Specimens Of the 24 specimens, for each exposure period, that had been strengthened with FRP 12 were further post-tensioned in order to induce a stress field along the FRP-concrete interface. These tensile stresses would stress both the concrete-FRP bond, as well as, the glass fibre, polyester laminate. The presence of these low level stresses would also help to show the possibility of stress corrosion cracking of the glass fibres in the laminate due to tensile stress. 56 Dywidag Bar Steel Bearing Plate P I _ Lock Nut 25 mm 100 mm o %" <fr P V C Tube 3mm F R P Laminate t P Figure 3.1. Schematic of Post-tensioning Post-tensioning of the samples was achieved using 15 mm (5/8 in.) diameter Dywidag Rods (§ 3.2.6). The rods were inserted into the post-tensioning duct that was cast into each of the samples. The rods were held in place using Dywidag supplied nuts bearing against 6.5 mm ('A in.) thick steel plate, which was bearing against the concrete. Prior to loading, the nuts were finger tightened to keep the plates from shifting during the prestressing operation. The rods were loaded to 133 kN (30,000 lbf) using a hydraulic Baldwin Universal Testing machine. The ultimate capacity of the testing machine is 266 kN (60,000 lbf)- While the load was held at 133 kN, the nuts were tightened using a crescent wrench. After the nuts had been tightened, the load was released and the samples were removed. P Pey CT = — ± -A I ( 0 57 Based on equation 1 it was possible to determine the level of stress along the FRP-concrete interface. In the CPCI Design Manual for Precast and Prestressed Concrete (1996), losses for prestressed elements are generally in the range of 18 to 25 percent. Assuming 25% losses due to relaxation of the prestressing steel, the final level of stress along the FRP/concrete interface was 5 MPa. This was deemed an acceptable low level of sustained bond stress for the FRP laminate. It was noted during visual observation of the prestressed specimens that microcracks were present in the concrete. Figure 3.2. The Post-tensioning Operation At the completion of the environmental exposure, the prestressing was released using the Baldwin Universal Testing machine. The prestressing bars were removed from the samples, and the samples were prepared for testing. 3.4. Environments In order to evaluate the performance of the FRP laminate, three different environments were selected: room temperature, plain water, and high pH. The room temperature environment was 58 the ambient conditions of the laboratory with an average temperature of approximately 20°C. All of the room temperature specimens were stored on a laboratory bench away from the windows and out of direct sunlight. The plain water exposure consisted of submerging the sample beams in a large tank full of tap water. The pH of the water was monitored during the exposure period to ensure that the environment never exceeded a pH of 8. In order to control the pH, there were times when it was necessary to drain the tank and refill it with tap water of the correct pH. The tendency of the plain water solution to increase in pH may be attributed to hydroxides leaching out of the concrete specimens into the plain water. In order to achieve constant exposure to a high pH environment, the beams selected for exposure were submerged in a solution of water saturated with calcium hydroxide (lime). The pH of this solution was carefully monitored, and never dropped below a pH of 12. Although not exact, the saturated solution would simulate an extreme alkaline environment that an FRP strengthened concrete member would experience during its service life. 3.5. Summary of Specimens and Environments Table 3.7. Summary of Specimens and Environments Sample Environment & Exposure Control (room temp.) [RT] Plain Water (room temp.) [W] High pH Solution (room temp.) [PHI Days Exposed 30 60 90 30 60 90 30 60 90 Plain Concrete [C] 4 4 4 4 4 4 4 4 4 FRP Sprayed Concrete TFC1 4 4 4 4 4 4 4 4 4 Post Tensioned FRP Sprayed Concrete [PFC] 4 4 4 4 4 4 4 4 4 59 Prior to testing each specimen was labelled in the following fashion. This allowed for easy identification and data manipulation required for analysis. Sample Exposure PFC60PH1 Days Exposed Sample No. Figure 3.3. Sample Identification Legend 3.6. Mechanical Testing All flexural tests for the concrete beams were conducted following ASTM C 1018-89 (Standard Test Method for Flexural Toughness and First-Crack Strength of Fibre-Reinforced Concrete [Using Beam With Third-Point Loading]), performed on an Instron mechanical testing machine, which had a maximum axial compressive loading capacity of 150 kN. The cross-arm displacement was maintained at a constant rate of 0.1 mm/min. 60 c E %" (j) P V C Tube CN r 1 J) 1 3mm F R P Laminate P/2 P/2 I 1 100 mm typ. h s r 300 mm 350 mm Figure 3.4. Schematic of Beam Setup Prior to testing, each specimen had the centre line on the side faces marked as well as the third points. This would facilitate the application of the measuring devices and placement of the specimen in the loading frame. A Japanese yoke was mounted on the specimen to eliminate spurious deflections caused by the concrete crushing at the supports. This provided the data to generate the load displacement curves required for the detailed analysis. Two Linear Voltage Displacement Transducers (LVDTs) were mounted at midspan to measure the deflection of the beam. Readings from the load cell and the LVDTs were displayed and recorded using a microcomputer and standard data acquisition software. The time period for each beam test ranged from 20 minutes to 30 minutes. The specimen was considered to have failed when the displacement exceeded 2 mm, and/or the load reading had dropped below 5 kN. 61 Figure 3.5. Testing Apparatus (Japanese Yoke) and Data Acquisition Unit At the completion of the testing, each beam was examined for trends in failure mechanism, delamination patterns, and other deficiencies. These trends were duly noted for further examination. The specimens were photographed in based on their reinforcement and exposure groups. For reference these photographs can be found in Appendix B for further examination. 62 4.0 E X P E R I M E N T A L R E S U L T S 4.1. Beam Testing Results 4.1.1. Observations Prior to Testing During the course of the investigation, careful observations were made of the laminate, the concrete specimens, and the laminate concrete bond. There were three distinct stages for each specimen prior to testing. The following observations are divided into those three areas: before exposure, during exposure, and after exposure. There were few notable observations with the most notable being the transition from a clear to opaque appearance of the resin in both the plain water and high pH environments, the resin flaking and peeling after exposure to the high pH environment, and fibre prominence after exposure to the high pH environment. All of these observations are explained in detail below. 4.1.1.1. Before Exposure Immediately after casting and during the 28 day period of curing there was little to observe of the concrete samples. Demoulding of the specimens and clearing of the post-tensioning duct had no detrimental effects on the samples. At the completion of the 28-day curing period, the samples were removed from the curing bins and allowed to air dry before specimen preparation continued. Observations at this time recorded no significant change in the appearance of the specimens. The curing process had affected the colour of the specimens changing the concrete colour to a lighter shade of grey, but this was the only noticeable change. 63 The samples selected for FRP laminate application required surface preparation on the face to which the laminate was to be bonded. The surface preparation consisted of both sand blasting and grinding to remove any large deformities and the concrete laitance layer, it was noted that grinding was a much more efficient method of surface preparation removing the laitance layer and smoothing any protrusions. The surface abrasion created by the sand blasting; however, would act to provide better mechanical interlock between the laminate and the concrete specimen. It should also be noted that size played an effect in surface preparation. Although grinding seemed to be more efficient the surface area was only 0.035 m2. Sandblasting would be far more efficient for large concrete members where the amount of surface requiring attention is in the range of multiple square meters. During the application of the FRP laminate, there were a few observations regarding the interaction between the concrete and the laminate materials. The concrete surface noticeably absorbed the primer coat during the hour-long curing period. Absorption was indicated by a change in colour of the concrete surface from light grey to dark grey, any void spaces on the surface of the specimen would quickly wick away the primer as soon as it was applied using the brush, and the meniscus of the primer coat was observed to diminish during the curing period. Polymerization of the resin was apparent towards the end of the pot life. After approximately 25 minutes had passed, the resin would begin to gel. The onset of the gel period was preceded by a change in viscosity and the appearance of thread-like strands of polymerised material. When these changes were observed the resin was deemed unfit to use and a new batch of resin was catalyzed. It was possible to laminate one ply on eight beams in a 20-minute period, so the appropriate amount of resin was catalyzed to minimize waste. 64 Air bubbles in the resin were frequently observed during the application of the laminate. During the reaction of the resin and the catalyst foaming occurs. The foaming will often manifest itself in the laminate with the appearance of tiny air bubbles in the hardened resin. Foaming and the resulting air bubbles can be minimised in the final laminate with the use of an anti-foaming catalyst and/or careful attention during the application and rollout of the laminate. Based on the scale of the beams and the accessibility of the laminated surface, the rollout of laminate was simplified and air voids were kept to a minimum. Once the concrete specimens had been laminated and the laminate had cured, the specimens were divided into their exposure groups. Eight specimens from each exposure group were post-tensioned. Examination of the post-tensioned specimens before and after showed that small hairline (<0.1 mm) cracks originating from the tensioned face of the beam (due to their small size photographs did not have the resolution to capture the cracks). The tension face was also the same face to which the FRP was applied. No audible cracking was observed during the post-tensioning operation, and no delamination of the FRP was observed at the completion of the post-tensioning. 4.1.1.2. During Exposure Three types of exposure were considered in this investigation: room temperature, plain water, and high pH. Observations of the specimens were consistent based on the exposure environment. In other words, all of the laminated specimens at room temperature underwent the same changes, all of the specimens in the water bath 65 underwent the same changes, and all of the specimens in the high pH solution underwent the same changes. The laminated specimens that were kept at room temperature showed no observable change during the 30, 60, and 90 day exposure periods. The small cracks that were observed in the post-tensioned samples did not propagate, or change in length. In fact, the cracks seemed to close-up over the course of the exposure. This is most likely attributed to the relaxation of the post-tensioning bars that would have occurred with time or self healing in concrete. The appearance of the laminate remained the same throughout the duration of the exposure. The specimens in the water environment underwent drastic changes in appearance. The initial colour of the FRP laminate is a translucent brown/amber colour; however, within 24 hours of exposure to the water solution the colour of the laminate had changed to a cloudy and opaque milky grey colour. Investigation of the laminate showed that the colour change was not due to a residue from the water deposited on the specimen. In fact the colour change had occurred throughout the laminate and was not only restricted to the outer layer. This colour change stayed with the laminate for the duration of the exposure without further change to the colour of the laminate or the concrete specimen. Deposits of corrosion were observed on the post-tensioning steel. This was expected in the neutral pH aqueous environment. The corrosion was restricted to the post-tensioning steel and did not affect the laminate or the concrete specimen. Observations of the specimens were difficult in the high pH solution. The solution was a combination of tap water saturated with calcium hydroxide. Measurements of 66 the pH frequently confirmed the pH of the solution met or exceeded a pH of 12, and this was confirmed by the absence of corrosion on the post-tensioning steel. The high pH solution was quite cloudy due to the presence of suspended calcium hydroxide particles. These particles settled out of the solution with time making it possible to observe the specimens; however, any movement of the specimens resulted in the generation of murky clouds of lime reducing visibility in the exposure bath to zero. The few observations that were possible showed a degree of clouding in the once translucent resin; however, it should be noted that this clouding was not to the same degree as the clouding experienced by the laminate in the plain water bath. It was also possible to observe that the calcium hydroxide particles that had settled out of the solution were resting directly on the laminate. This direct contact ensured the laminate's exposure to highly alkaline solutions. The high pH bath was very similar to the curing bath for the concrete specimens. Based on this it is not surprising that there were no observable changes to the concrete specimens themselves. 4.1.1.3. After Exposure Prior to testing, each specimen was labelled and had the third points and centreline drawn on to facilitate the test set-up. During this time, each specimen was carefully examined and any new observations were noted. With the exception of the high pH solution exposure, there were no significant observable changes in any of the specimens that would indicate damage to the laminate. The specimens exposed to the high pH solution showed significant fibre prominence: the only visible sign of laminate deterioration. 67 Figure 4.1. Specimens after Exposure (Control - Top, Water Exposure - Middle, High pH Exposure - Bottom) Figure 4.2. Fibre Prominence after high pH Exposure The room temperature environment had little effect on the samples exposed to that environment. No changes were noticed in the specimens at the completion of the exposure period. The small cracks in the post-tensioned specimens were still present; however, their prominence had been reduced. The reduction in crack size can be attributed to the relaxation of the post-tensioning steel that had occurred 68 during the exposure period, and the removal of the steel that had occurred prior to testing. The FRP laminate was intact with no signs of ultra violet deterioration, delaminating, fibre prominence, and/or cracking. In comparison with the initial, before exposure, appearance there was no appreciable difference. The samples exposed to the plain water environment showed noticeable differences after exposure. After the samples had been removed from the water, the FRP laminate was covered by a white film. Attempts were made to rinse of the white film; however, it was difficult to remove and scrubbing it would have only damaged the fibre. It was also observed that the laminate below the surface was lighter in colour when compared to the room temperature specimens. This lightening in colour can be attributed to the absorption of moisture by the resin matrix. The specimens were allowed to dry for 24 hours. At the completion of the drying period, the white film had become a chalky residue that could be wiped off, but never completely removed. The resin had a darker appearance indicating that some of the absorbed moisture had evaporated. Examination of the laminate showed that although some of the resin had been dissolved by the water there was still a layer of resin covering the fibres; therefore, fibre prominence was not a problem. No peeling, delaminating, and/or cracking of the FRP laminate was observed on any of the plain water exposure specimens. The specimens exposed to the high pH solution showed the greatest change at the end of the 30, 60 and 90 day exposure periods. Upon removal, the specimens were covered with a white film. Unlike the water exposure samples, the white film was rinsed of easily as the specimens were removed from the exposure bath. The white film present in the high pH bath can be attributed to the settling out of the calcium hydroxide particles suspended in the saturated solution. The most notable change was observed at the same time the white paste was rinsed off the specimens. Immediately after rinsing the specimens fibre prominence was observed on every specimen. The fibre prominence was not isolated to spots on each specimen, but was present along the length and width of the laminate. By running a gloved hand along the surface of the laminate, it was evident that a thin layer of resin had been removed exposing the fibres directly to the deleterious environment. The specimens were removed from the environmental bath and allowed to dry for 24 hours. At the completion of the 24-hour drying period, other indications of resin damage to the FRP laminate were observed. During application of the resin and laminate, resin was applied to the sides of the beam to seal the glass fibres exposed after the rough edges of the laminate had been ground down. Resin had also dripped down the sides of the beams during the application of the laminate. After the 24-hour drying period, those beams that had been exposed to the high pH exhibited peeling of the resin painted on to the sides. The dried specimens has extensive peeling of the plain resin on the sides of the beam, and the peeling resin would flake off easily when the surface of the beam was rubbed with a dry cloth. A putty knife was used to remove the peeling resin before the centre line and third points were applied with a marking pen. The specimens exposed to the high pH environment showed the greatest observable changes of all of the specimens in the experimental program. 4.1.2. Observations During Testing All of the specimens, regardless of their age or exposure, exhibited similar behaviour during loading. During the first few minutes while the specimen was undergoing the initial loading, little change in the load carried by or the displacement of the specimen was 70 recorded. After the specimen and load cross head had seated correctly the load began to increase with little change to the midspan deflection of the specimen. At the first crack of the specimen the midspan deflection began to increase at a constant rate of 0.1 mm/min., which was the rate specified by the ASTM standard. In the majority of cases, the first crack quickly grew into one large crack that propagated up the beam face throughout the duration of the test. In the case where more than one crack was observed, only one crack became the dominant failure crack. In some cases the formation of the dominant crack occurred when the multiple cracks all propagated together to form one single crack. If the crack spacing was too great to allow this to happen one crack would eventually dominate leaving the other crack to remain at its final size, or even to close up slightly as the dominate crack became the main source for energy to be released. Figure 4.3. Failed Specimen Showing Failure Crack and Delaminated Sheet With the exception of 3 specimens, failure of the laminated specimens occurred when the FRP sheet delaminated and popped of off the specimen. Warning of the failure was given in advance in a number of ways. With the formation of the dominant concrete crack, debonding of the FRP laminate began adjacent to the crack mouth. The crack created 7 1 between the laminate and the concrete was observed to propagate in both directions away from the crack mouth. As the loading progressed, the laminate stopped delaminating in both directions with the crack propagating in one direction only. The crack between the laminate and the concrete progressed to one end of the specimen. Failure occurred when the crack reached the support and "popped" the laminate off the beam. The delamination of the FRP sheet was preceded by a series of audible cracking noises with the final failure accompanied by a loud snap. 4.1.3. Results and Observations at the Conclusion of Testing The following sections present the results obtained during the testing program. Chapter 5.0 will present a more detailed analysis of the results. Any observations that may correlate with the test results and a brief interpretation of the results will be mentioned in the following sections for information. The tabulated results in the following sections contain the averages of all the specimens for the appropriate variable, exposure, and tested property. For a complete set of results, please refer to AppendixC. Appendix B provides a complete set of the photographs of the failed specimens. Although delamination occurred in 69 of the 72 specimens strengthened with FRP and despite the cracking in the specimens, the PVC prestressing duct, located in the compression zone of the specimen, prevented the specimens from breaking into two separate pieces. Examination of the surface of the peeled FRP laminate and the concrete face from which the sheet peeled showed an adhesive failure and a minimal amount of concrete bonded to the laminate. This indicates that the bond between the vinyl ester primer resin and the concrete is not stronger than the cement aggregate bond of the concrete specimen. The small amount of concrete that was bonded to the E-glass/polyester and vinyl ester consisted 72 mainly of small grains of sand and paste that had not been completely removed during the sand blasting and grinding of the specimens. Previous research by Johnson (1997) using carbon fibre/epoxy and glass fibre/epoxy laminates bonded to concrete showed that the failed samples had significant concrete bonding to the laminate; evidence that indicates that the epoxy concrete bond was stronger than the cement aggregate bond. Examination of the specimens that had achieved 2 mm of deflection exhibited a degree of mechanical interlock. During the sand blasting and grinding procedure, air pockets and aggregate protrusion appeared. These raised and depressed surfaces increased the overall roughened texture of the bonding surface allowing for a greater bonding surface area and more resin bonding and interaction with the concrete surface. The indentations and protrusions also acted as miniature shear keys restraining the shearing forces that were acting on the laminate sheets during delamination. Surface treatment played a role in maximizing deflection prior to complete delamination. Although sandblasting and grinding effectively removed the laitance layer the degree to which they effectively abrade the surface remains a discussion point. Future investigations could help to determine to what extent the surface treatment of a reinforced specimen affects the both the bond strength and the bond durability. 4.1.4. Cracking and Ultimate Strengths and Deflection For the complete table of results please refer to Appendix C. 73 4.1.4.1. Calculations The recording equipment had been calibrated before testing and the recording equipment directly converted the electrical signals to their kN or mm values as the values were recorded. This meant that no conversions of the data were required during the analysis of the results. For the cracking loads and the ultimate loads, the value presented is the average value for all four of the specimens for the exposure condition, exposure duration, and strengthening condition. The only data manipulation required was zeroing the load cell data and removing the numerical offsets that were introduced by the recording equipment and the load cell. The data registered a sight load prior to application of the load. The load was zeroed by subtracting this load from all of the recorded values. During the test, a Japanese yoke was mounted on the specimen to eliminate spurious deflections caused by the concrete crushing at the supports. Two LVDTs were mounted to the yoke on either side of the beam at midspan. A separate channel was used for each LVDT to record two independent values for each beam. The recorded values were corrected for any initial offset, and the average of the two readings was taken to produce the load versus midspan data for a tested beam. The values presented in Table 4.1 are the averaged values for the four specimens subject to the applicable exposure condition, exposure duration, and strength condition. 74 4.1.4.2. Summary Table 4.1 provides a summary of the calculated values for the first crack load, ultimate load, and maximum midspan deflection for the samples tested. The results are varied making it difficult to provide the exact tendencies of the data; however, the general trends of the data are • the presence of the E-glass/polyester laminate increases the first cracking strength, ultimate strength and maximum midspan deflection of the specimen; • first cracking and ultimate strengths tends to increase with time; • the maximum midspan deflection tends to increase with time; • those specimens exposed to a room temperature environment tends to have higher measured values for first cracking strength, ultimate cracking strength and maximum midspan deflection; and • low level (5 MPa or less) prestressing of the laminate/concrete bond does not appear to have a significant effect on the performance of the strengthened concrete specimens for the exposure periods investigated. Please note that these are general comments. Please refer to the following chapter, 5.0 Discussion, for a more detailed breakdown of the performance of the samples complete with explanations for the variations in performance. 75 Table 4.1. Summary of Cracking and Ultimate Strengths and Deflection Specimen Cracking Load, kN Ultimate Load, kN Deflection, mm Days Exposed Days Exposed Days Exposed 30 60 90 30 60 90 30 60 90 C-RT 17.7 21.0 22.9 17.7 21.0 22.9 0.05 0.07 0.06 C-W 14.2 15.5 16.9 14.2 15.5 16.9 0.06 0.06 0.06 C-PH 20.0 18.3 17.5 20.0 18.3 17.5 0.05 0.05 0.05 FC-RT 23.3 22.6 31.1 23.9 24.3 31.3 1.23 1.39 1.04 FC-W 21.1 20.5 23.9 22.0 24.9 31.6 0.79 1.43 1.03 FC-PH 21.4 21.8 26.4 22.3 25.2 30.4 1.09 1.32 0.96 PFC-RT 23.8 23.0 27.7 24.3 23.7 33.6 1.26 1.37 1.36 PFC-W 23.1 19.8 24.1 23.5 27.6 30.0 0.71 1.63 1.18 PFC-PH 22.8 20.9 26.4 23.0 24.4 31.0 0.84 1.20 1.02 C - Contro , FC - FRP Wrap, PFC - Stressed FRP Wrap RT - Room Temperature Air Exposure, W - Room Temperature Water Exposure, PH - Room Temperature High pH Solution Exposure 4.1.5. Toughness of Specimens For the complete table of results please refer to Appendix C. 4.1.5.1. Calculations The fracture energy for each specimen was obtained by calculating the area under the Load-Deflection curve. The data were stored in a spreadsheet form, which made the manipulation, and calculation of the area based on thousands of data point a simple task. For the calculation, the following formula was employed. area = J-(^„+ 1-y„X*„+, ~xn)+yn{xn+,-xn) i 2 where (2) yn - nth. load value xn = nth. deflection value 76 By programming this formula into a spreadsheet, calculation of the area under the Load-Deflection curve was calculated quickly and efficiently. 4.1.5.2. Summary Table 4.2 provides a summary of the calculated values for the fracture toughness of the samples tested. The results are varied making it difficult to provide the exact tendencies of the data; however, the general trends of the data are • the presence of the E-glass/polyester laminate increases the fracture toughness of a concrete specimen when compared to an unstrengthened specimen; • fracture toughness tends to increase with time; • those specimens with a room temperature exposure tends to have increased values of fracture toughness; and • low level (5 MPa or less) prestressing of the laminate/concrete bond does not appear to have a significant effect on the fracture toughness of the strengthened concrete specimens for the exposure periods investigated. Please note that these are general comments. Please refer to the following chapter, 5.0 Discussion, for a more detailed breakdown of the performance of the samples complete with explanations for the variations in performance. 77 Table 4.2. Fracture Energy of Specimens Fracture Energy, Nm Specimen Days Exposed 30 60 90 C-RT 0.5 0.9 0.9 C-W 0.6 0.7 0.7 C-PH 0.6 0.6 0.6 FC-RT 21.2 27.7 28.0 FC-W 14.2 32.0 28.0 FC-PH 17.1 30.1 26.7 PFC-RT 23.1 25.6 38.6 PFC-W 13.1 34.9 31.6 PFC-PH 15.7 25.2 27.5 C - Contro , FC - FRP Wrap, PFC - Stressed FRP Wrap RT - Room Temperature Air Exposure, W - Room Temperature Water Exposure, PH - Room Temperature High pH Solution Exposure 4.1.6. Initial Stiffness For the complete table of results please refer to Appendix C. 4.1.6.1. Calculations Initial Stiffness was calculated calculate by taking the slope of the trend line of the Load-Deflection curve for each specimen for the portion of the curve in the 2-10 kN range. This range was selected because each of the specimens tested showed a general linearity in this region. The trend line method was used to eliminate any noise due to sensitivity in the electronically recorded data. Microsoft Excel software was used to calculate the equation for the trend line and from this equation if was possible to extract the slope of the line which represented the stiffness of the specimen. Figure 4.4., below, illustrates this method of obtaining the stiffness for a specimen. 78 12 10 •o 6 flj o y = 607.66X + 0.6606 y = 478.53x - 0.5746 y = 543.42x +0.5169 y = 535.39x - 0.7371 —PFC90PH1 PFC90PH2 PFC90PH3 — PFC90PH4 — Linear (PFC90PH2) — Linear (PFC90PH4) — Linear (PFC90PH1) — Linear (PFC90PH3) 0.005 0.01 0.015 Midspan Deflection, mm 0.02 0.025 Figure 4.4. Typical Method of Slope Determination for the Initial Stiffness Calculation 4.1.6.2. Summary Table 4.3 provides a summary of the calculated values for the initial stiffness of the samples tested. The results are varied making it difficult to provide the exact tendencies of the data; however, the general trends of the data are • the initial stiffness of the plain concrete specimens and the strengthened concrete specimens tends to increase with time; • those specimens exposed to the room temperature exposure tend to have lower values for initial stiffness when compared to those sample exposed to both a high pH environment and a plain water environment; 79 • those specimens strengthened with an E-glass/polyester laminate tend to have higher values for initial stiffness when compared to the unstrengthened concrete specimens; • low level (5 MPa or less) prestressing of the laminate/concrete bond appears to result in lowered values of initial stiffness compared to both the plain concrete specimens and those specimens strengthened with an E-glass/polyester laminate. Please note that these are general comments. Please refer to the following chapter, 5.0 Discussion, for a more detailed breakdown of the performance of the samples complete with explanations for the impact on performance. Table 4.3. Initial Stiffness of Specimens Specimen C-RT C-W C-PH FC-RT FC-W FC-PH PFC-RT PFC-W PFC-PH Initial Stiffness, kN/mm Days Exposed 30 434.6 414.6 486.2 507.1 449.9 450.1 400.7 481.8 500.9 60 469.8 484.9 497.0 466.4 453.4 492.8 454.3 474.7 469.0 90 549.3 464.5 523.2 595.2 586.6 565.2 527.2 525.8 541.3 C - Control, FC - FRP Wrap, PFC - Stressed FRP Wrap RT - Room Temperature Air Exposure, W - Room Temperature Water Exposure, PH - Room Temperature High pH Solution Exposure 4.1.7. Overall Summary Table 4.4 provides an overall summary of the tested and calculated results. Appendix C contains a colour coded summary sheet, which helped to identify the trends in the data. Chapter 5.0 provides a detailed discussion of the results. 80 Table 4.4. Overall Summary of Experimental Results Specimen Cracking Load, kN Ultimate Load, kN Fracture Energy, Nm Deflection, mm Initial Stiffness, kN/mm Days Exposed Days Exposed Days Exposed Days Exposed Days Exposed 30 60 90 30 60 90 30 60 90 30 60 90 30 60 90 C-RT 17.7 21.0 22.9 17.7 21.0 22.9 0.5 0.9 0.9 0.05 0.07 0.06 434.6 469.8 549.3 C-W 14.2 15.5 16.9 14.2 15.5 16.9 0.6 0.7 0.7 0.06 0.06 0.06 414.6 484.9 464.5 C-PH 20.0 18.3 17.5 20.0 18.3 17.5 0.6 0.6 0.6 0.05 0.05 0.05 486.2 497.0 523.2 FC-RT 23.3 22.6 31.1 23.9 24.3 31.3 21.2 27.7 28.0 1.23 1.39 1.04 507.1 466.4 595.2 FC-W 21.1 20.5 23.9 22.0 24.9 31.6 14.2 32.0 28.0 0.79 1.43 1.03 449.9 453.4 586.6 FC-PH 21.4 21.8 26.4 22.3 25.2 30.4 17.1 30.1 26.7 1.09 1.32 0.96 450.1 492.8 565.2 PFC-RT 23.8 23.0 27.7 24.3 23.7 33.6 23.1 25.6 38.6 1.26 1.37 1.36 400.7 454.3 527.2 PFC-W 23.1 19.8 24.1 23.5 27.6 30.0 13.1 34.9 31.6 0.71 1.63 1.18 481.8 474.7 525.8 PFC-PH 22.8 20.9 26.4 23.0 24.4 31.0 15.7 25.2 27.5 0.84 1.20 1.02 500.9 469.0 541.3 C - Control, FC - FRP Wrap, PFC - Stressed FRP Wrap RT - Room Temperature Air Exposure, W - Room Temperature Water Exposure, PH - Room Temperature High pH Solution Exposure 81 5.0 D I S C U S S I O N 5.1. Introduction In order to evaluate the effectiveness of the laminate strengthening, it was necessary to quantify the performance based on numerical comparisons. For this evaluation, five properties that could provide insight into the material and specimen behaviour were measured and/or calculated. The five properties that were considered are listed below. • Cracking Load. The measured load corresponding to the first crack in the specimen. • Ultimate Load. The maximum load resisted by the specimen. • Fracture Energy. This value was calculated by measuring the area under the load deflection curves for each of the beams. • Midspan Deflection. This was the maximum average midspan deflection experienced by each beam prior to failing. It was calculated by taking the average of measurements made by the two LVDTs mounted at the midspan of the beam. • Initial Stiffness. This value was calculated by measuring the initial slope of each load deflection curve between 2 and 10 kN of load. A summary of the average measured values is presented again for convenience. 82 Table 5.1. Summary of Results Specimen Cracking Load, kN Ultimate Load, kN Deflection, mm Fracture Energy, Nm Initial Stiffness, kN/mm Days Exposed Days Exposed Days Exposed Days Exposed Days Exposed 30 60 90 30 60 90 30 60 90 30 60 90 30 60 90 C-RT 17.7 21.0 22.9 17.7 21.0 22.9 0.05 0.07 0.06 0.5 0.9 0.9 434.6 469.8 549.3 C-W 14.2 15.5 16.9 14.2 15.5 16.9 0.06 0.06 0.06 0.6 0.7 0.7 414.6 484.9 464.5 C-PH 20.0 18.3 17.5 20.0 18.3 17.5 0.05 0.05 0.05 0.6 0.6 0.6 486.2 497.0 523.2 FC-RT 23.3 22.6 31.1 23.9 24.3 31.3 1.23 1.39 1.04 21.2 27.7 28.0 507.1 466.4 595.2 FC-W 21.1 20.5 23.9 22.0 24.9 31.6 0.79 1.43 1.03 14.2 32.0 28.0 449.9 453.4 586.6 FC-PH 21.4 21.8 26.4 22.3 25.2 30.4 1.09 1.32 0.96 17.1 30.1 26.7 450.1 492.8 565.2 PFC-RT 23.8 23.0 27.7 24.3 23.7 33.6 1.26 1.37 1.36 23.1 25.6 38.6 400.7 454.3 527.2 PFC-W 23.1 19.8 24.1 23.5 27.6 30.0 0.71 1.63 1.18 13.1 34.9 31.6 481.8 474.7 525.8 PFC-PH 22.8 20.9 26.4 23.0 24.4 31.0 0.84 1.20 1.02 15.7 25.2 27.5 500.9 469.0 541.3 C - Control, FC - FRP Wrap, PFC - Stressed FRP Wrap RT - Room Temperature Air Exposure, W - Room Temperature Water Exposure, PH - Room Temperature High pH Solution Exposure The following discussion will focus on these five areas considering the effect of time, FRP strengthening, exposure, and prestress. After examining each of these individual variables and the related effects an overall comparison of the beams was used to determine the overall trends applicable to concrete reinforced with FRP laminates. The chapter concludes with a brief discussion on the unexpected results obtained in the course of this investigation including the effectiveness of prestressing the concrete/laminate bond. 5.2. Load versus Deflection Curves The most important tool used for the analysis of the results was the load versus deflection curve generated for each of the tested specimens. Figure 5.1 shows a typical load versus deflection curve for a concrete beam strengthened with a chopped strand FRP laminate. The slope of the initial linear portion was used to calculate the initial stiffness (load versus deflection) for each of the samples. Typically, the curve remained linear until the concrete cracked. The load at which 83 the first crack occurred was affected by the presence of the laminate with the strengthened specimens generally cracking at a higher load. After first crack, the load typically dropped of as the deflection increased. The explanation for this is quite simple. With the first crack, stress concentrations at the crack mouth and the FRP laminate/concrete interface caused the FRP sheet to debond from the concrete. As the debonding of the FRP laminate progressed, the load continued to drop and deflection continued to increase. The debonding of the laminate sheet stopped when the stresses were redistributed and an equilibrium condition halted the progress of the FRP sheet separation from the concrete. Typical load increase with the First c rack \^ redistribution of stresses I Typical partial bond \ failure at peak of \ ~~load increase followed by a reduction in load Typical failure \ occurs with the debonding of the FRP laminate Uncracked, slope of initial stiffness 0 0.5 1 1.5 2 Midspan Deflection, mm Figure 5.1. Typical Load versus Deflection Curve for a Concrete Beam Strengthened with a CSM FRP Laminate Why the FRP sheet debonding stopped is not exactly known. It can be hypothesized that along the bonded surface there is not uniform bond strength. In this hypothesis, it can be assumed that the chemical bond between the laminate and the concrete surface was uniform. The other factor affecting bond strength is the mechanical interlock between the concrete and the laminate. This 84 mechanical interlock would not be uniform in nature. Prior to the application of the laminate, the concrete surface was abraded using a hand grinder and by sandblasting allowing the concrete specimens to develop a rough surface that would enhance the mechanical interlock between the FRP sheet and the concrete specimen. The points of mechanical interlock would act as barriers to the progress of the concrete/laminate debonding crack. After the specimen first cracked, the bond between the FRP sheet and the concrete was not strong enough to withstand the concentrated stress at the crack mouth and as a result, the sheet began to peel. The peeling continued until a point of increased bond strength (due mechanical interlock). At this point, the concrete beam was able to halt the onset of the debonding crack and carry more load. The load was able to increase while the stresses between the sheet and the concrete surface were redistributed around the site of increase bond. With the load increase, the stress between the concrete and the laminate increased until failure of the bond caused the peel crack to grow and the load to drop. Figure 5.1 illustrates the load increase and decrease, and peeling relationship. This pattern of load increase followed by decrease was common to the test results of all of the strengthened samples. The propagation of the peeling crack was accompanied by audible cracking noises. The volume of the snapping sounds was directly related to the accompanying drop in loud. The loudest cracks preceded the largest drops in load. Failure of the majority of the specimens was announced with a substantial and often startling snapping sound. The multi-peaked load versus deflection curve is typical for most concrete specimens strengthened with bonded FRP sheets. Although sudden failure takes place eventually, the cyclical load increase followed by growth of the debonding crack and load drop leading to a halting of the debonding crack and subsequent load increase means that there are definite warning signs before the sudden failure of the bond. The accompanying cracking noises that were observed also provide early warning of failure. The existence of these audible warning signs 85 suggests that there is the potential to incorporate acoustic emissions testing and monitoring into FRP laminate strengthening of concrete. 40 30 0 ! - FC60RT1 FC60RT3 —FC60RT2 - FC60RT4 0.5 1 1.5 Midspan Deflection, mm Figure 5.2. Load versus Deflection Curve for a Concrete Beam Strengthened with a CSM FRP Laminate with 60 days exposure to a Room Temperature Air Environment Comparing one batch of four beams where one of the four beams reached 2 mm of midspan deflection (Figure 5.2.) shows that the bond had little effect on the first cracking strength; however, bond did have significant impacts on ultimate strength, midspan deflection and fracture energy. It is noted that three of the four beams in this batch were showing the same trend in the load/deflection relationship. Beams 3 and 4 had were showing similar deflection increases when the sheet peeled causing failure. Examination of all four of the laminate sheets after failure showed that there was more concrete material bonded to sample 2 than to 1, 3 or 4. The bond that had developed on the concrete/FRP interface was stronger than the concrete itself. It is evident that the quality of bond plays a significant role in increased midspan deflections and ultimate load. When a good chemical and mechanical bond is present, there are significant increases in the midspan deflection and ultimate load. 86 5.3. Cracking Load Test results showed that the cracking load tended to increase with time. With the exception of the control specimens exposed to a high pH environment, all of the beams tested required a greater load to initiate cracking after 90 days of aging compared to the load required after 30 days of aging. The increase in cracking load with time is consistent with the strength gain behaviour of all concrete specimens. Figure 5.3. Difference in Cracking Load (high pH - water) versus Time for pH and Water Exposures 87 5.3.1. Control Specimens It is interesting to note that the samples exposed to the high pH environment, the saturated calcium hydroxide solution, showed a decrease in first crack performance with time. This is in direct contradiction of typical concrete behaviour: strength increases with time. All of the samples were cured for 28 days in a saturated calcium hydroxide bath. This method of curing is the method of preference for curing concrete cylinders. A saturated solution of calcium hydroxide used as the curing bath inhibits additional calcium hydroxide, which is also a product of the hydration of the cement, from leaching out of the concrete which in turn leads to stronger concrete. This is contradicted by the control specimens that had a reduced first cracking load with time even though they were exposed to what should have been an ideal curing environment. 5.3.2. Wrapped Specimens As expected, the specimens strengthened with an FRP laminate had higher first cracking loads than the plain concrete control samples. The presence of the FRP laminate acted as tensile reinforcement for the concrete thus increasing the load required to initiate cracking. The general trend for the samples showed that the specimens that had been exposed to the dry room temperature environment had the highest first cracking loads followed by the specimens exposed to the high pH solution. The specimens exposed to the plain water solution had the lowest first cracking strengths. This is similar to the behaviour of the control samples; however, based on the differences in cracking loads it is evident that the exposure environment played an effect. 88 In all cases, the laminated samples exposed to the moist environments did not perform as well as the samples exposed to the room temperature condition. This is in agreement with previous research showing the susceptibility of FRP laminates to aqueous solutions. Examination of the samples after exposure found that the samples exposed to the high pH solution had exposed fibres indicating that the resin matrix had deteriorated due to the alkaline environment. The specimens in the plain water solution did not have any exposed fibres; however, the resin had grown cloudy in appearance. After examining the differences between the cracking loads, it was shown that the aqueous environment had the greatest effect in decreasing strength for the 30 and 60 day exposures. After 90 days of exposure, the differences between the first cracking strengths for the alkaline environment are significant when compared to the tap water environment. Figure 5.3. is a representation of this. After 90 days of exposure, the tap water was shown to have the most significant effect on reducing the load required to initiate cracking. 5.3.3. Wrapped and Prestressed Specimens The concrete specimens that were wrapped with FRP laminate which was further stressed with a prestress of 5 MPa exhibited similar first cracking behaviour to the specimens that had the unstressed laminate. The specimens exposed to room temperature had the highest first cracking loads. Although the water exposed specimens had a higher first cracking load after 30 days exposure compared to the high pH exposure, by 90 days of exposure the specimens in the high pH solution had a higher cracking strength. The difference in strength followed the same linear pattern as the laminated samples without the prestress (Figure 5.3.). 89 The presence of the prestress had little effect on the cracking strength of the specimens. As mentioned earlier in the procedure, the prestressing load was only in place during the exposure period. Examination of the averaged results shows that there is little difference in the first cracking trends shown by the prestressed samples compared to trends shown by unstressed samples. Based on the results it can be inferred that a prestress of only 5MPa on the FRP laminate/concrete, approximately 10 percent of the concrete's compressive strength, is not enough to affect the load required to initiate cracking in the strengthened specimens. Overall, the specimens subjected to the room temperature exposure had higher first cracking loads. This is similar to the results obtained when wet cylinders are tested and the results are compared to their air dry counterparts. Wet concrete tests lower than dry concrete primarily due to stress corrosion of the cementitious components of the concrete. Although all of the specimens were air dried for a minimum of 48 hours before testing it is still possible that the samples taken from the wet environments still had enough moisture present to significantly affect the cracking strength of the samples. 5.4. Ultimate Load 5.4.1. Effect of Time Following a similar trend to the cracking load, the ultimate load of the specimens increased with time. For the control specimens the ultimate load was equivalent to the cracking load, so the samples exposed to the high pH solution had the same uncharacteristic reduction in load with time. 90 5.4.2. Effect of Laminate The FRP laminate has obvious effects on the strength of the concrete specimens. The concrete beams that had an FRP laminate bonded to their bottom faces were able to withstand a greater flexural load than the control specimens. Increased load capacity was common to all of the strengthened beams regardless of the exposure conditions. The FRP laminate acts much the same way as steel reinforcing bar in traditional reinforced concrete. Without the presence of reinforcement in the tensile portion of a concrete beam subjected to bending, ultimate failure occurs when the concrete cracks. By having a material that is stronger than concrete incorporated into the tensile portion of the beam it was possible to exceed the tensile capacity of the region which ultimately leads to an increase in the ultimate load capacity of the beam when loading in bending occurs. 5.4.3. Effect of Exposure Analysis of the data shows inconclusive results for the effect of exposure on the ultimate load capacity of the concrete beams. Five of the nine sets of beams tested produced results showing the room temperature exposed beams as having the greatest ultimate load capacity; however, two of the control sets are in this group of five making the results inconclusive for the beams with the FRP laminate bonded to the tension face. Of the six sets of beams strengthened with FRP laminates analysis showed that three sets subject to the room temperature exposure showed the highest average ultimate load for the exposure duration, two sets exposed to the tap water solution had the highest average ultimate load, and one set exposed to the high pH solution had the highest ultimate load. For the beams 91 strengthened with FRP laminates the room temperature exposure was shown to generate marginally better results for ultimate load capacity when compared to the tap water and the high pH exposures. 5.4.4. Effect of Prestress Examination of the results for ultimate load reveals little in the way of discernable trends. The immediate conclusion that can be drawn is that 5 MPa of prestress on the concrete/laminate interface during exposure has little effect on the ultimate strength of the sample when compared to the equivalent unstressed sample. 5.5. M i d s p a n Deflection 5.5.1. Effect of Time The trend for the samples tested showed that a peak in recorded deflection occurred after 60 days of exposure. This evidence suggests that the concrete samples became more brittle with time which is consistent with the theories of concrete science. 5.5.2. Effect of Laminate The presence of the laminate dramatically increases the ductility of the concrete specimen. On average, the midspan deflection at failure increased approximately 20 percent. The increase in ductility is easily explained. The control samples were unreinforced concrete. When the first crack occurred there was nothing to resist the tensile forces in the concrete and the crack rapidly propagated through the depth of the beam. For the unreinforced 92 control samples, the first crack initiated immediate failure of the sample; hence, the midspan deflection was directly proportional to the ductility of the concrete. The laminate acted in the same way as traditional steel reinforcement. When the concrete began to crack, the tensile stresses were transferred to the FRP sheet that was bonded to the tensile (bottom) face of the beam. The presence of the FRP sheet meant that not only could the concrete beam sustain more load, but the ductility of the beam increased as well. Had the laminate bond remained intact and the failure occurred by concrete crushing or FRP tensile failure increased ductility would have been recorded. This is shown by the three specimens that reached 2 mm of midspan deflection before failure: 60 days, prestressed wrap subject to tap water exposure; 60 days, wrap subject to tap water exposure; and 60 day; wrap room, control. Table 5.2. Comparison to Findings of Banthia et al. (1996) for Wrapped Beams, Room Temperature Air Exposure Specimen Johnson (2003) Banthia et al. (1996) Days Exposed 30 60 90 Cracking Load, kN 23.3 22.6 31.1 17.59 Ultimate Load, kN 23.9 24.3 31.3 22.70 Fracture Energy, Nm 21.2 27.7 28.0 20.47 Midspan Defection, mm 1.23 1.39 1.04 0.92 The ductility of the specimen was limited by the flexibility of the sheet and the strength of concrete/laminate bond. Due to the relatively high fibre content of the FRP laminate (~30 percent by weight, ~15 percent by volume) the laminate itself was quite stiff. Banthia et al. (1996) found that it was possible to fail a bonded laminate in a similar testing program; however, it should be noted that in that series of tests the laminate had a fibre volume of 8 93 percent by volume. The stiffness of a laminate is determined by the fibre content, so it would be expected that a laminate with fewer fibres, though not as strong, would have a greater overall flexibility leading to increased values for deflection. The failure of the sheet in the Banthia et al. test program is also evidence that there may have been a better concrete-FRP bond. In these tests the stiffness of the laminate created increased shear stresses along the concrete/FRP interface. The rigidity of the sheet compared to the flexibility of the cracked concrete meant that as the deflection increased the net shear force along the bond interface increased resulting in a peeling type failure rather than a rupture of the FRP sheet. In the samples that reached 2 mm of defection the bond was strong enough to resist the peeling stresses. Additional comparison with the data from Banthia et al. shows that the increased fibre content in the laminate resulted in higher values for cracking load, ultimate load, fracture energy and midspan deflection. Based on all of the test results it can be concluded that strengthening of the concrete using chopped strand mat FRP laminates is an effective method of increasing the ductility of concrete specimens loaded in bending. 5 . 5 . 3 . Effect of Exposure For the most part, the specimens subjected to a room temperature environment had the greatest ductility. After 60 days of exposure, the specimens that had been exposed to the tap water environment recorded the greatest midspan deflections, followed by the room temperature exposure. All of the samples after 60 and 90 days of exposure to the high pH environment had the lowest average values for midspan deflection. It is obvious from this 94 that long term exposure to an alkaline environment is detrimental to the ductility of concrete specimens strengthened with FRP laminates. In Chapter 2, previous research into the susceptibility of glass fibre laminates subject to deterioration by alkaline environments was discussed. Porter and Barnes (1998) and Liao et al. (1998) have both found the alkaline solutions have a deleterious effect on fibre morphology. Hydroxylation occurs in a high pH environment and ultimately results in the embrittlement of the glass fibre. The embrittlement is apparent in this investigation with the reduction in midspan deflection and specimen ductility. The samples exposed to the high pH solution also showed significant fibre prominence due to dissolution of the resin matrix. Buck et al. (1998) and Coomarasamy and Goodman (1999) report significant deterioration of the resin after long-term exposure to a plain water environment. In this case the visual observations of the samples exposed to the plain water environment showed deterioration of the resin; however, this did not significantly effect the ductility of the strengthened specimens. The specimens in the highly alkaline solution, however, showed significant deterioration after only 30 days, and the flaking and peeling of the resin that was prominent along the sides of the beam indicating that the alkaline exposure contributed significantly to the deterioration and embrittlement of the resin matrix. The findings of this investigation confirm that long-term exposure to an alkaline environment is detrimental to the ductility of concrete specimens strengthened with FRP laminates. 95 5.5.4. Effect of Prestress Analysis of the data shows that there are no conclusive trends in the data comparing the prestressed samples to those samples that were not prestressed. Analysed data shows that prestress on the bond had a distinct effect only after 90 days of exposure. After 90 days of exposure, the samples strengthened with an FRP laminate and then prestressed with 5 MPa along the concrete/laminate interface on average showed an increased ductility compared to the samples that had not been prestressed. It should be noted that with the exception of the room temperature exposure this increase was at most 14 percent. For the other exposures, the samples that had not been prestressed generally had greater deflections. Based on the inconsistency of the results, no conclusions are made regarding the 5 MPa of prestress that was applied to the concrete/laminate interface during exposure and its effect on deflection. 5.6. Fracture Energy The value for fracture energy provides an indication of the toughness of the specimen. Higher values indicate that the specimen has a higher toughness. Fracture energy for a specimen is determined by calculating the area under the load deflection curve. 5.6.1. Effect of Time Fracture energy is dependent on both load and deflection; two properties that have been previously shown to increase with time. Based on this and by comparing the calculated values it can be shown that fracture energy also tends to increase with time. The exception to this observation occurs in the series of specimens that had been strengthened with a 96 laminate and exposed to the tap water solution. In the case of both the unstressed and the prestressed samples, the fracture toughness peaked at 60 days exposure and then started to decrease. This phenomenon will be dealt with in detail in Section 5.6.3. 25 20 z -6 00 o 15 10 Region of increasing deflection with little decrease in load - Control Strengthened with FRP Laminate Failure occurs after a period of sustained loading with a dramatic increase in deflection 0.2 0.4 0.6 Midspan Deflection, mm 0.8 Figure 5.4. Typical Curves for Load versus Midspan Deflection for a Control Specimen and Strengthened Specimen 5.6.2. Effect of Laminate All of the specimens strengthened with the FRP laminate showed greater fracture toughness when compared to the plain concrete control specimens. Figure 5.3 shows the typical curves for a control specimen and a strengthened specimen. The area under the curve for the specimen strengthened with the FRP laminate is far greater than the area under the curve for the control specimen. Plain concrete is quite brittle which means that it also has low fracture toughness. Cracking in concrete occurs first with the formation of microcracks, and as the load increases, the microcracks open into larger cracks with the 9 7 propagation of the crack tip. As the load increases, the crack tip propagates through the material. Failure occurs suddenly when the cracking load for the control sample is reached. The ultimate load for plain concrete is equivalent to the cracking load because with the first crack the fracture energy available causes rapid propagation of the crack and immediate failure of the specimen. The specimens that are strengthened with the FRP laminates are able to control the crack tip propagation by absorbing the fracture energy that would normally propel the crack tip. Once the specimen has cracked, the load drops off as the crack begins to propagate; however, the laminate bonded to the tensile face of the beam bridges the crack mouth preventing the rapid opening of the crack and rapid failure of the specimen. For this investigation, the loading crosshead was lowered at a constant rate of O.lmm/min., and failure of the strengthened specimen occurred when the sheet peeled off the specimen. The peeling mechanism is tied directly to the cracking of the concrete. The strengthened specimen still must transfer the cracking energy into the concrete and FRP components. Once the concrete has cracked, the FRP is preventing the concrete from further cracking; however, at the same time there is a horizontal crack propagating along the FRP/concrete interface. The load for this portion may remain relatively constant, or may increase with the increased deflection. The energy that builds up due to the increase in deflection and load is released with the peeling of the sheet and when failure inevitably occurs is accompanied by loud snapping noises as the remaining energy is released as sound. 5.6.3. Effect of Exposure No conclusions regarding the effect of exposure on fracture energy can be made without the consideration of period of exposure. With this consideration a number of conclusions 98 are drawn from the experimental data. Room temperature conditions are shown to be the best conditions for maximizing the fracture energy for a specimen after a lengthened period of exposure. After 90 days of exposure, the room temperature exposure resulted in specimens with the highest values for fracture energy regardless of strengthening and prestressing. Samples subject to a tap water environment had the second highest values for fracture energy followed by the samples conditioned in the high pH solution. The explanation for this is quite straightforward, and again shows that long-term exposure to a high pH solution is detrimental to the physical properties of glass fibre FRP laminates. It has been previously been discussed that exposure to tap water and alkaline solutions are detrimental to glass fibre. Both deflection and ultimate loads tend to be reduced with time and exposure, so it comes as no surprise that fracture energy, which is the area under the load deflection curve, is reduced as well. The results for the samples tested after 30 days and 60 days exposure are largely inconclusive. After 30 days of exposure, the unstressed samples strengthened with FRP laminates exposed to the room temperature environment displayed the highest average values for fracture energy. After 30 days of exposure, fracture energy for the control specimens is maximized after exposure to the high pH solution. After 60 days of exposure, the samples that had been strengthened with a glass fibre laminate had the highest fracture energies after exposure to the plain water environment, and the control samples for the same exposure period had the highest values after exposure to the room temperature environments. Clearly, for shorter periods of exposure, there are no discernible trends. 99 5.6.4. Effect of Prestress As has been the case with the other measured and/or calculated properties there appears to be no real correlation with the presence of prestress and performance of the specimen. Section 5.9. specifically addresses the overall role and performance of prestress in this investigation. For the specimens that were subject to prestress the common trend that can be identified is that after 90 days of exposure, regardless of environment, the specimens that had the bond between the laminate and the concrete prestressed had higher calculated values for fracture energy than the samples that were strengthened with the laminate without the prestress. For the 30 day and 60 day exposure periods, the dominant trend was for those samples with an unstressed bond between the laminate and the concrete to have higher calculated values for fracture energy. It is interesting to note that the rankings for fracture energy are identical to the rankings for deflection, but differ from the rankings for ultimate load capacity. This indicates that deflection has a greater impact on the calculated value for fracture energy than ultimate load capacity. Certain caveats should apply to this statement. In the case of this experimental program, the differences in the ultimate load capacity for the strengthened specimens did not vary to the same degree (max. 7%) as the variation in deflection (max. 31 %) based on comparisons between the prestressed and non-prestressed results. Based on this, the variations in deflection were more critical to the fracture energy of the samples than the ultimate load capacity of the specimen, and it can be further inferred that prestressing had a greater effect on the ductility of the samples than it did on the ultimate strength. Further discussion of this will take place in Section 5.9. 100 5.7. Initial Stiffness 5.7.1. Effect of Time The general trend for the samples tested showed a peak in initial stiffness after 90 days with the stiffness tending to increase with duration of exposure. This is expected considering the first crack strength also tended to increase with duration of exposure. 5.7.2. Effect of Laminate For the 30 day and 60 day exposures, the FRP strengthened specimens had increased values for initial stiffness compared to the unreinforced control specimens. After 60 days of exposure, however, the unreinforced concrete samples had higher values for initial stiffness compared to their strengthened counterparts. After 90 days of exposure the, those sample that had been strengthened with an FRP laminate showed higher average values for initial stiffness. The interpretation of this trend indicates that the stiffness benefits of FRP strengthening are greatest after an increased period of exposure. The general trend discussed in Section 5.7.1 showed that for all of the specimens initial stiffness tended to increase with time. With that in mind, it can be surmised that the combined stiffening effects of the FRP laminate and concrete are best utilized in the early stages, or after periods of longer duration. This premise supports the use of FRP laminates for strengthening existing concrete beams where the concrete has aged significantly and any initial increase in stiffness may have decreased with time. 101 5.7.3. Effect of Exposure For the 30 and 60 day exposures, those specimens exposed to the highly alkaline solution showed higher initial stiffness than both the specimens exposed to the tap water solution and the specimens exposed to room temperature conditions. After 90 days, however, it was the specimens subject to room temperature exposure that tended towards having the highest initial stiffness. The explanation for these results is quite straightforward. Exposure to highly alkaline solutions affects concrete quite differently than it affects E-giass/polyester laminates. Concrete reacts positively to an alkaline environment. The curing solution for the concrete specimens was a saturated lime solution, and the same solution was used for the alkaline exposure for the strengthened specimens. The same highly alkaline solution, however, has a negative effect on stiffness and strengthening efficiency of the E-glass/polyester laminate. Karbhari and Engineer (1996b) note that matrix degradation is largely to blame for the reduction in the flexural stiffness of concrete specimens strengthened with FRP laminates. The effect of alkaline environments is also shown in the findings of Coomarasamy and Goodman (1999) which report that noticeable decreases in the stiffness of both the E-glass/polyester and E-glass/vinylester are shown after only 11 weeks of exposure to a calcium hydroxide solution. It has been shown that FRP laminate strengthened concrete samples exposed to an alkaline solution show increase in the strength and stiffness of the concrete and a decrease in the stiffness of the FRP laminate. Knowing this, the initial stiffness results with time for the test specimens are easily explained. In the early days of exposure, the FRP laminate had not been significantly affected by the alkaline environment and any degradation that may have occurred had little overall effect compared to the stiffness effects of the bonded laminate. Any degradation that may have occurred was also compensated for by the 102 increased stiffness of the concrete due to the extended curing period in the calcium hydroxide solution. This is why, for the 30 and 60 day exposures, the specimens exposed to the alkaline environment generally showed higher initial stiffness values. At 90 days of exposure, the samples that had been strengthened with the E-glass/polyester laminate began deviate from this trend. The reinforced but not prestressed samples exposed to the room temperature conditions had higher values for initial stiffness than the specimens exposed to the alkaline environment. The explanation for this is simple. The detrimental effect of the alkaline environment on the FRP laminate was shown to have a more significant effect on the initial stiffness than the beneficial effects of the extended curing period. 5.7.4. Effect of Prestress Examination of the tabulated results shows that on average prestressing the FRP/concrete bond had a detrimental effect on the initial stiffness of the concrete specimens. Six of the nine exposure groups tested had results that showed samples without prestressed interfaces having higher values of initial stiffness compared to those specimens that had their FRP/concrete interface prestressed, and of these six groups all three of the 90 day exposure groups showed a decrease in the initial stiffness of the samples that were prestressed. There is a simple explanation for this. One of the purposes of this research was to investigate the effects of low-level sustained tensile stresses on the interface between concrete and the E-glass/vinyl and polyester laminate. The decrease in the initial stiffness of the samples subject to prestressing indicates that the prestress had some effect on the integrity of the FRP/concrete bond. Deterioration of the bond led to a decrease in bond integrity between the laminate and the concrete. Microcracking of the matrix and/or stress corrosion of the E-glass fibres most likely took place. This led to a decrease in stiffness of 103 the laminate, which ultimately led to an overall decrease in the initial stiffness of the concrete/FRP system. During the period of prestressing the area of the interface was subject to tensile stresses of 5 MPa. Over a short period, this level of stress would have little effect on the bond and the initial stiffness of the laminate and the laminate/concrete composite. After an extended period of prestress, however, the extent of matrix cracking and stress corrosion of the fibres leads to the deterioration of the bond, the laminate, and ultimately the initial stiffness of the specimen. This is evident in the fact that over time those specimens that had not been subject to prestressing had increased values for initial stiffness. 5.8. Overall Comparisons This investigation considered three variables affecting the performance of the FRP laminate: exposure environment, period exposed, and prestressing. In total 108 specimens were tested. In order to best compare the samples, it was necessary to simplify the representation of the data. Four samples were used to measure the impacts of each variable, and from the four sets of data the average value for each of the five properties (cracking and ultimate loads, fracture energy, deflection, and initial stiffness) was obtained. This reduced the total number of values from 108 to 27, which was a more manageable set of numbers to work with. The complete set of reduced and simplified data can be found in Appendix C . 104 Cracking Load, kN Specimen Days Exposed 30 60 90 PFC-W 0 -1 1 24.12, 19.84 23.08, 26.6 21.7 23.3 24.8 21.9 20.3 20.3 16.9 21.83 24.19 21.78 24.5 F igure 5.5. Flowchart Demonstrating the Reduction and Simplification of Data After ranking the average values in a -1,0,1 format it was possible to directly evaluate the performance of the laminate after exposure and prestress. The evaluation was made by adding the values for each beam group (i.e. C-RT30, C-RT60, C-RT30) at each measured/calculated characteristic (Cracking Load, Ultimate Load, Fracture Energy, Deflection, Initial Stiffness). These values were then combined to calculate the ranking for each beam treatment and exposure (i.e. C-RT, FC-RT, PFC-W, etc). The spreadsheets showing this data reduction and normalization are contained in Appendix C. The direct result of this comparison is that the specimens with a laminate at room temperature exposure showed the best overall performance regardless of exposure and days exposed. Figure 5.6 shows the overall performance of the nine specimen types. 105 5.8.1. Chemical Exposure Of the two variables investigated, exposure and bond stress, exposure had the greatest effect on the durability of the specimens. Looking at Figure 5.6 it is not difficult to see that the room temperature exposure has the least effect on decreasing the properties of the five properties of the samples. Both the tap water solutions and the high pH solutions were expected to be deleterious to the integrity and strength of E-glass/polyester laminate. The test results confirm that when compared to room temperature exposure concrete specimens strengthened with E-glass/polyester laminates exposed to either a tap water solution or a highly alkaline solution show an overall reduction in performance. This is based on overall comparisons obtained by taking the averaged values for each unique exposure type, duration, reinforcement, and prestress. The data was further refined by ranking each set of three data points as outlined in the previous section. The results showing that the specimens exposed to a room temperature environment were as predicted. Although concrete benefits from extended periods of exposure to highly alkaline solutions and to a lesser degree aqueous solutions (the presence of moisture allows for an extended period of hydration), the strength and durability of the FRP laminate is seriously compromised. The laminate is subject to chemical deterioration to the extent that any increases in strength due to the strengthened concrete are negated due to the deteriorated FRP laminate. 106 20 15 10 (0 V 5 JB .5 CO 0 > c a> E -5 "o a> a -10 10 -15 -20 -25 C - Plain Concrete Control FC - Wrapped Concrete PFC - Prestressed Wrapped Concrete PH - High pH Exposure at Room Temperature RT - Air Exposure at Room Temperature W - Tap Water Exposure at Room Temperature i l l F C - W P F C - F C - P H P F C - P F C - F C - R T P H W R T Normalized Values Figure 5.6. Combined Overall Response of the Specimens to the Variables (Exposure, Prestress, Wrap) Independent of Time These results are further validated with examination of the average load displacement curves. The average load displacement curves (Appendix D) show that for all exposure periods the room temperature air exposure had the highest cracking load, and for four of the six exposure periods had the highest midspan deflection and highest ultimate load. A sample average load displacement curve is shown in Figure 5 . 7 . 1 0 7 40 30 T3 20 (8 o 10 — Control —Water Exposure —High pH Exposure 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Midspan Deflection, mm Figure 5.7. Load versus Deflection - Average Comparison - Prestressed Wrap - 90 Days 5.8.2. Effect of Prestress Figure 5.6 shows little conclusive evidence required to determine the effects of prestressing the concrete/laminate interface. Examination of Figure 5.6 shows that prestressing the concrete/laminate interface has a detrimental effect for specimens subject to a room temperature exposure, prestressing the concrete/laminate interface has a beneficial effect for specimens subject to tap water environment, and prestressing the concrete/laminate interface has a detrimental effect for specimens subject to a high pH environment. The explanation for the difference in performance between the unstressed and the prestressed samples in the high pH environment can be attributed to the increased stress corrosion of the E-glass fibres due to the application of low-level sustained tensile stress during the exposure period in the high pH solution. Increased matrix cracking due to the applied stress and chemical degradation due to the high pH environment also contributed to the overall decrease in performance. 108 The explanation for the performance of the samples exposed to the tap water solution is not a clear. It was expected that the specimens subject to prestress during the course of exposure would have a decreased overall performance compared to the samples that were not prestressed. Figure 5.6 shows that this was not the case and instead the samples with a prestressed interface showed an overall increase in performance compared to their unstressed counterparts. Possible explanations for this include • low-level sustained tensile stresses do not induce stress corrosion to the same extent in aqueous solutions when compared to solutions with a high pH; • prestressing tendons applied a compressive stress to the entire beam; however, since this stress was applied eccentrically a tensile force was generated along the laminated face of the beam to create the bond stress. The compressive force that was applied to the beam may have served to compress the concrete to create a denser, more impermeable concrete. The compression of the concrete would hinder the leaching out of the calcium hydroxide compounds resulting in a stronger concrete, and this may be one possible explanation for the prestressed, laminated samples displaying overall increased performance compared to their unstressed, laminated counterparts exposed to the aqueous (tap water) solution; • four samples of each type of exposure type and duration, prestress, and laminate configuration may not have been enough to obtain a truly representative statistical distribution. Despite the inconsistencies with that samples exposed to a tap water solution, it can be concluded based on the results presented in Figure 5.5 that prestressing the concrete/laminate interface with a low level sustained stress negatively affects the 109 durability and integrity of a concrete specimen strengthened with E-glass/polyester laminates. 5.8.3. Overall Of the variables investigated in this experiment, chemical exposure had the greatest effect on the durability of concrete specimens strengthened with an E-glass/polyester laminate. Second to the exposure is the level of prestress. Time of exposure may play a role; however, a maximum exposure period of 90 days is not long enough to effectively demonstrate deterioration resulting from extended exposure periods. Based on the results presented throughout this discussion there are the following conclusive results. • Given the exposure conditions and durations, prestress of the bond, and laminate configuration it is shown that concrete samples strengthened with an E-glass/polyester laminate demonstrate the best durability characteristics when the bond is unstressed and the strengthened samples are exposed to a dry room temperature environment. • Tensile stresses are more detrimental to strengthened specimens when that specimen is exposed to a high pH solution compared to a similarly prestressed specimen that is exposed to tap water. • Overall, high pH solutions have a greater detrimental effect on the performance of concrete specimens strengthened with E-glass/polyester laminates when compared to an equivalent sample that is exposed to a tap water environment. 110 Chapter 6.0 provides a more comprehensive look at the both general and specific conclusions. 5.9. Impacts and Implications The guiding premise of this investigation was to explore the durability aspects of concrete strengthened with E-glass/polyester laminates. Based on the results of this investigation a number of areas for concern and future investigation have been identified. One of the most promising end uses for the laminate strengthening system investigated in this report is for the strengthening of concrete building members and bridge beams. The performance of the concrete specimens strengthened with the E-glass/polyester laminates was shown to provide acceptable strengthening and durability in strengthening applications where the strengthened specimen was exposed to a dry, room temperature environment. This would make the system ideal for internal building members where the environment is easily regulated and controlled. Although this strengthening system is ideal for internal building members, there is concern surrounding its use in outdoor settings. One of the greatest potential applications for this concrete/laminate strengthening system is for the strengthening of deteriorated bridge beams. The exposure conditions for these beams include salt water attack from sea spray and de-icing salts, rain and melting snow leading to attack from high pH solutions leaching from the concrete at the concrete/FRP interface, and variable loading due to vehicles travelling across the bridge. This experiment attempted to duplicate a number of these conditions (high pH, rain water, sustained loading), and based on the results it can be shown that there is a decrease in performance by the laminate strengthened concrete when exposed to the aforesaid conditions. I l l The strengthening ability of the E-glass/polyester system has been demonstrated both in this investigation and in the work of Boyd (1999). It must be noted, however, that both Boyd and this investigation agree that the greatest strengthening occurred when the strengthened specimens were exposed to a controlled, dry, room temperature environment. This investigation clearly shows that performance of the strengthened specimens is compromised when the specimens are prestressed and/or exposed to either high pH solution or a tap water solution. Although this experiment investigated the behaviour up to 90 days of exposure, there are still concerns as to the long-term durability of the E-glass/polyester strengthening system that should be addressed before this method should be recommended for widespread use. 112 6.0 C O N C L U S I O N S A N D R E C O M M E N D A T I O N S The purpose of this investigation was to examine concrete strengthened with E-giass/polyester and the effects of environment, time and bond stress on the concrete specimens. Based on the analysis of the results, the following conclusions and recommendations may be made. 6.1. Conclusions 1. FRP short fibre laminates bonded to the tensile face of unreinforced concrete beams effectively increase the cracking capacity, ultimate capacity, ductility, and fracture energy of the beams when the beams are subjected to third point loading. 2. Long-term exposure to an alkaline environment is detrimental to the ductility and fracture energy of concrete samples strengthened with chopped strand FRP laminates. 3. Low-level sustained tensile stress on the FRP/laminate bond negatively impacts the initial stiffness of the specimen. 4. Low-level sustained tensile stress on the concrete/laminate bond is detrimental to the durability of a concrete beam strengthened with E-glass/polyester laminates. 5. The durability of concrete strengthened with an E-glass/polyester laminate appears to be acceptable for specimens exposed to dry, room temperature conditions. The results of this investigation show that samples strengthened with E-glass/polyester laminates are susceptible to environmental degradation and deterioration. 113 6. Surface treatment plays a role in maximizing deflections. A roughened surface was shown to increase bond strength resulting in increased values for midspan deflection. Testing observations and analysis of the load versus deflection curves indicate that the a roughened surface appears to increase the mechanical bond at the concrete/laminate interface. 7. Tests periods of 30, 60, and 90 days at room temperature conditions are not long enough to provide adequate information for long term exposure trends. 8. The chopped strand mat and polyester laminate used in this experiment is nearly identical to the laminate produced using the spray up method. Based on the positive results provided by this investigation, the use of spray-up fibre has the potential to be used as a concrete strengthening material. 6.2. Recommendations The findings of this investigation provide positive answers concerning the use of random short fibre E-glass/polyester laminates as a material to strengthen and reinforce existing concrete structures; however, further investigation needs to take place. A better understanding of the behaviour of concrete strengthened with random, short fibre E-glass/polyester laminates will be developed if the following ideas are considered: 1. Larger scale experiments, similar to the channel beam study conducted at UBC in 1999 (Boyd, 1999 and Ross, 2000), should be carried out with exposure conditions and time of exposure as the considered variables. It is clear that in a room 114 temperature, laboratory environment short fibre E-glass/polyester laminates are an effective method of strengthening concrete; however, there needs to be more data on the real world behaviour of these FRPs and their ability to strengthen concrete. 2. Future test programs should investigate the effect of periods of exposure longer greater than 90 days. This may be achieved through extended time periods and/or through a process of accelerated testing. With longer exposure periods, it will be possible to obtain more definitive answers regarding the behaviour of the concrete/laminate bond and its performance in both neutral and high pH aqueous environments. 3. Future test programs should investigate the effect of different methods of surface preparation of the concrete face to which the laminate is to be bonded. 4. In 97.6 percent of the tested specimens the FRP sheet delaminated before the deflection target of 2mm was reached. Future investigations should investigate the effect of fibre volume and laminate thickness and the role of these variables on premature debonding of the laminate sheet. 5. The application of E-glass/polyester laminates to concrete surface has, to this point, taken place in the controlled environment of the laboratory. Future investigation considering the practicality of applying this material in the field will be beneficial to industry acceptance of this material as a concrete strengthening system. 115 7.0 References Aiello, Maria Antonietta. (1999). Concrete cover failure in FRP reinforced beams under thermal loading. Journal of Composites for Construction. 3(1), 46-52. Arduini, Marco, and Nanni, Antonio. (1997). Behaviour of precracked RC beams strengthened with carbon FRP sheets. Journal of Composites for Construction. 1(2), 63-70. Ashland Chemical Incorporated. (1998, March). K1907/K1908 Series Technical Data Sheet. Kelowna, BC: Ashland Chemical Incorporated. ASTM C1018. (1989). Standard test method for flexural toughness and first-crack strength of fiber-reinforced concrete [Using beam width third-point loading] (ASTM Designation C1018-89). 1990 Annual Book of ASTM Standards: Vol. 04.02 Concrete and Aggregates (pp. 500-506). Philadelphia, PA: American Society for Testing and Materials. ASTM C39. (1986). Standard test method of compressive strength of cylindrical concrete specimens (ASTM Designation C39-86). 1990 Annual Book of ASTM Standards: Vol. 04.02 Concrete and Aggregates (pp. 20-24). Philadelphia, PA: American Society for Testing and Materials. Bank, Lawrence C , Gentry, T. Russell, and Barkatt, Aaron. (1995) Accelerated test methods to determine the long-term behavior of FRP composite structures: environmental effects. Journal of Reinforced Plastics and Composites. 14, 559-587. Banthia, N.; Yan C ; and Nandakumar, N. (1996). Sprayed Fiber Reinforced Plastics (FRPs) for Repair of Concrete Structures. Proceedings, 2nd International Conference on Advanced Composite Materials. ACMBS II. 537-546. 116 Bizindavyi, L., and Neale, K.W. (1999). Transfer lengths and bond strengths for composites bonded to concrete. Journal of Composites for Construction. 3(4). 153-160. Boyd, A.J., and Banthia, N. (1999). Shear strengthening of reinforced concrete beams with sprayed GFRP. Structural Faults & Repair, 8th International Conference (Proceedings). Boyd, Andrew James. (2000). Rehabilitation of reinforced concrete beams with sprayed glass fiber reinforced polymers. The University of British Columbia, Vancouver, Canada. Buck, Stephanie E., Lischer David W., and Nemat-Nasser, Sia. (1998). The Durability of E-Glass/Vinyl Ester Composite Materials Subjected to Environmental Conditioning and Sustained Loading. Journal of Composite Materials, 32(9), 877-892. CAN/CSA-S6-00. Canadian Highway Bridge Design Code. Canadian Standards Association. Chajes, Michael J., Thomson Jr., Theodore A., and Farschman, Cory A. (1995). Durability of concrete beams externally reinforced with composite fabrics. Construction and Building Materials. 9(3). 141-148. Childs, Philip J. (1999). Glass Rebar's Growing Pains. Civil Engineering. 69(2). 65-67. Clarke, John L. (1999). Fibre-reinforced plastic reinforcement for concrete. Concrete (London), 33(1). 15-16. Coomarasamy, Arunasalam, and Goodman, Stephen. (1999) Investigation of the characteristics of fiber reinforced plastic (FRP) materials in concrete environment. Journal of Thermoplastic Composite Materials, 12, 214-226. CPCI. (1996). Design Manual Precast and Prestressed Concrete. Ottawa, ON: Canadian Prestressed Concrete Institute. 117 Debaiky, Ahmed, Green, Mark F., and Hope, Brian B. (1999). FRP rehabilitation of corrosion-damaged concrete structures. International SAMPE Symposium and Exhibition (Proceedings). 44(11). 2170-2181. Dow Plastics. (1994, October). Fabricating Tips - Derakane epoxy vinyl ester resins. (Form No. 125-00082-196 SMG). Midland, MI: The Dow Chemical Company. Dow Plastics. (1997, February). Derakane epoxy vinyl ester resins - Reference & Formulating Guide - A Field Book for North American Fabricators. (Form No. 125-00275-297 SMG). Midland, MI: The Dow Chemical Company. Dow Plastics. (1999, March). Derakane Epoxy Vinyl Ester Resins Chemical Resistance and Engineering Guide. (Form No. 125-00043-399X SMG). Midland, MI: The Dow Chemical Company. Dywidag Systems International. (1998, October). Dywidag Threadbars - Technical Data [Imperial and Metric Units]. Surrey, BC: DSI. Emmons, Peter H., Thomas, Jay, and Vaysburd, Alexander. (1998). Muscle made with carbon fiber. Civil Engineering, 68(1), 60-61. GangaRao, Hota V.S., and Vijay, P.V. (1998). Bending behavior of concrete beams wrapped with carbon fabric. Journal of Structural Engineering, 124(1), 3-10. Johnson, Martin. (1997). Freeze-Thaw Durability of RC Beams Strengthened with CFRP Sheets. Kingston, ON: Queen's University. 118 Karbhari, V.M., and Engineer, M. (1996a). Investigation of bond between concrete and composites: use of a peel test. Journal of Reinforced Plastics and Composites, 15, 208-227. Karbhari, V.M., and Engineer, M. (1996b). Effect of environmental exposure on the external strengthening of concrete with composites - sort term bond durability. Journal of Reinforced Plastics and Composites, 15, 1194-1216. Katsuki, F., and Uomoto, T. (1995). Prediction of deterioration of FRP rods due to alkali attack. International RILEM Symposium FRPRS-2 (Proceedings). 82-89. Liao, Kin, Schultheisz, Carl R., Hunston, Donald L., and Brinson, L. Catherine. (1998). Long-term durability of fiber-reinforced polymer-matrix composite materials for infrastructure applications: a review. Journal of Advanced Materials, 30(4), 3-40. Mack, Jaime K., and Holt, Erika E. (1999). The effect of vapor barrier encapsulating of concrete by FRP strengthening systems. International SAMPE Symposium and Exhibition (Proceedings), 44(H), 2182-2193. Mukhopadhyaya, P., Swamy, N., and Lynsdale, C. (1998). Optimizing structural response of beams strengthened with GFRP plates. Journal of Composites for Construction, 2(2), 87-95. Murphy, K., Zhang, S., and Karbhari, V.M. (1999). Effect of concrete based alkaline solutions on short term response of composites. International SAMPE Symposium and Exhibition (Proceedings). 44(H). 2222-2230. 119 Nanni, Antonio, and Gold, William J. (1998). Strength assessment of external FRP reinforcement. Concrete International. 20(6), 39-42. Neale, Kenneth W., and Labossiere, Pierre. (1998). Fiber composite sheets in cold climate rehab. Concrete International, 20(6), 22-24. Porter, M.L.; and Barnes, B.A. (1998). Accelerated aging degradation of glass fiber composites. In H. Saadatmanesh and M. R. Ehsani (Eds.), Second International Conference on Composites in Infrastructure ICCI-2 (Proceedings), Vol. 2, 446-459. Porter, Max L. (1999). Durability issues for FRP reinforcement to concrete structures. International SAMPE Symposium and Exhibition (Proceedings). 44(11), 2246-2254. Rubin, Irvin I. (Ed.). (1990). Handbook of Plastic Materials and Technology. New York: John Wiley and Sons, Inc. Seible, Frieder, Priestly, M.J. Nigel, Hegemier, Gilbert A., and Innamorato, Donato. (1997). Seismic retrofit of RC columns with continuous carbon fiber jackets. Journal of Composites for Construction, 1 (2), 52-62. Shehata, E., Rizkalla, S., and Stewart, D. (1998). Roof retrofit. Concrete International. 20(6). 44-46. Spadea, G., Bencardino, F., and Swamy, R.N. (1998). Structural behavior of composite RC beams with externally bonded CFRP. Journal of Composites for Construction, 2(3), 132-137. 1 2 0 Steckel, Gary L., Hawkins, Gary F., and Bauer Jr., Jerome L. (1999). Durability issues for composites in infrastructure. International SAMPE Symposium and Exhibition (Proceedings). 44(11). 2194-2208. Taly, Narendra, and GangaRao, Hota V.S. (1999). Guidelines for design of concrete structures reinforced with FRP materials. International SAMPE Symposium and Exhibition (Proceedings). 44(11), 1689-1696. Tannous, Fares E., and Saadatmanesh, Hamid. (1998). Environmental effects on the mechanical properties of E-Glass FRP rebars. ACI Materials Journal, 95(2), 87-100. Tannous, Fares E., and Saadatmanesh, Hamid. (1999). Durability of AR glass fiber reinforced plastic bars. Journal of Composites for Construction, 3(1), 12-19. Toutanji, Houssam A., and El-Korchi, Tahar. (1999). Tensile durability of cement-based FRP composite wrapped specimens. Journal of Composites for Construction. 3(1), 38-45. Toutanji, H., and Balaguru, P. (1998). Durability characteristics of concrete columns wrapped with FRP tow sheets. Journal of Materials in Civil Engineering. 10(1). 52-57. Toutanji, Houssam A., and Gomez, William. (1997). Durability of concrete beams externally bonded with FRP composite sheets. Cement and Concrete Composites, 19, 351-358. Ye, L., Zhang, S., and Mai, Y-W. (1998). Strengthening efficiency of E-Glass fibre composite jackets of different architectures for concrete columns. Applied Composite Materials, 5(2). 109-122. 121 Zhang, S., Ye, L., and Mai, Y-W. (1999). Effects of saline water immersion on glass fiber/vinyl-ester wrapped concrete columns. Journal of Reinforced Plastics and Composites, 18( 17), 1592-1604. 122 A P P E N D I X A Literature from Owens Corning Composites 123 A P P E N D I X B Beam Photographs 128 = E . — u u c. EC a M ca "S P-O o "fi u Cu u -CO O C-X u (U Q o 1 3 1 o 1 " 1 ° C30W2 C30W3 -5 e *o o 130 s II 5 0 S p s-B i « St «9 « O s £ • — o P i . X *0 h-1 g W ) " U ffi rn 131 1 8 i o LL. CL j PFC60W4 at E 3 u a m a c a "3 0-I -3 O CL X W I H c/5 C CL X p o 133 135 § § -a ••« cx-2 a* r ^ <u a W CL, / I S u 13 1 w co X 1 ^ ^ U ffi o 137 A P P E N D I X C Experimental Results Normalization of Data 138 c o S E I E Q CD CO I LU CO > CD Q CN oo co CO CN CO o ro ro r -as CD CD CD CO T3 CO o -J cu «-» CO E 3 T 3 CD CO » LU CO a CN CO CM I <N CD CO in cd CN CN cn CD 7" CD CD CD T3 T3 CO CD O co _ l O o>z C L X c ^ LU 2 CO o >i as CO Cr Q 00 T CN •<-CN CM o r-CD CD in ai in co CN co CN -IC CN 5 c cu E _^ cu 3 CL tn X UJ tc c a> E 5 cu CL CO 1) I CD > < o CN o CO o -~ I CO s_ o > < pi CD 1— 0 > < CD r i cu CD > < o CN o oo O o CJ — > < 139 CD CN •ST CD oo 00 •sr I s- 7: •sf 00 CD CD 00 -- co CO LO OT CM CO CD CO CD O 00 •sf CO CM I s-I s -CO co CO CM o T — o O o o T — " d " T— d d LO CO CO •s* CM co CM 00 rs. - •st LO CM CM CM LO CM I s-co 00 co CO CD LO oo o o CO CO o •sf LO O o co CO I s-co o CM 00 lO 00 LO CM CO 01 o o •sf CD CD LO CD CM OT 00 CD CD I s -I s -o LO Is-CO o •sf O •sf LO LO I s -•sf 00 T — o T — T "~~ T — d T — * d d d d d d d d CD CO o •sr LO CM co OT I D co CD oo LO o LO CD •sf ro o ro ro CO I s-CM CD CM o CO d CO CO CO LO CO CO M •3 r-^  CM •sf CO i d CM CO d CO CO d CO CM CO d CM CO I s - LO LO CM LO co CO oo I s- CD •sf I s - I s - co co LO 00 CD •sf oi CN CM CM CO CM LO CM LO CM •sf CM CO CM CO CM CO CM CO CM CO I s -CM CM OT CM CM •sf CM d CM CM CM d CM CM CM oo I s - CO co CO CM CD CO OT CM •sf LO LO O CM oo o •sT CM T— CM o CM CM CM CM CM CM LO CM CD CM *» CM %r CM CM CM •sf CM c\i CM •sf CM CO CM CM CM d CM d CM d CM d CM CD O T— co •sl- LO CD x— co I s- CD I s- CO 00 CO I s - CM CO •sf CO CM CM CO CM CM CM ed CM d CO CM CD CM CD CM CM d CM CM CO CM •sf CM •sf CM d CM d CM d CM d CM d CM m r - CO LO 00 LO CO CO O CD CO CO OT 00 •sf I s - OT CM CM CM CM CM o CM CM CM CM CO CM CO CM CO CM CM d CM d CM d OT CM CM d CM CM CM d d CM CD CO CM •sf CM co CM CD 00 ro 00 OT T — CO IN. LO O CM OT oo CM CM d CM CM CM cb LO CM CO CM •sf CM CO CM CM ^r CM CM •sf CM d CN CM CM d CM d CM CM CM CM CM FC-PH1 CM I D_ I: O L L CO X C L 1 o L L X C L ! o L L 0 CD co > < r -DL 1 o L L C L CM I— CH 6 L L C L co 1— Cd 1 O L L C L •sf \— r r i o L L CL 0 CD co _ 0 > < i i o L L C L CM 6 L L C L CO 1 O L L C L •sf i o L L C L CU CD CO — CD > < X C L 1 O L L C L CM X C L 1 O L L C L ro X C L 1 o L L C L •sr X 0. 1 o C L 0 CD CO I— 0 > < rs £ CO E CO © 3 a u> x cu H I ce CO •sr in CO CO in in 00 in •sr co o 00 CM CO in CN o •sr in CD 0) CO posed o CO CD m r-J CO m CO in CM CM in co •Nt in o in CO CO ^r CO CO •sr •sr CO •sr o m CO in CM O m 1^  m m CO CM in co CO in 00 m 00 co CD LO iri CT) in in i n m iri CN in oo CM r-CD CO in ffne E posed in o CD CO 00 oo r- co to CO CO o CO CO o co co CN •sr in m q oo •sr tial Sti E 2 X LU CO >* CO o CD o in CD in CO CD CM OO CO co co ai CD in in oo •sT CM •sr •sr co •sr 00 CO CO iri m •sr 00 •sr CM •sr r~ CO •sr •sr m •sr CO o in in r -•sr CM CO •sr CO CD •sr CD •sr co m •sr T— in •sr 00 CO ^r CO in •sr c Q CNI T - CN o co co CO O •sr co h- in •sr •sr CM o in co 1- i - •sr r - CO CO o CO i n in cb in •sr CO CD co CO CO •sr CN T— •sr CO CD •sr CO o •sr ai CO •sf •sr CM CD •sr iri h~ •sr CO o in CO •sr co 00 CD m •sr CO in iri CM •sr iri CO o in CD O in oo •sr o CO o ^r •sr ai •sr •sr o CT> h-CD q T— oo CD CM CT) CD oo CD •sT O co CD oo CD CD CD CD oo CD in CD CD 3 19.7 36.2 20.3 35.9 28.0 28.9 30.9 27.4 24.6 28.0 >» CD T3 jre Enei £ z ays Expose o CD CD CO CD 00 d CO o •sr CD o 00 CD 00 CD CD •sr CD in CD 00 CD o CD CD 13.3 47.2 29.9 20.5 27.7 22.1 27.7 35.3 43.0 32.0 Fracti ays Expose Fracti Q o CO m o in CD CD CD CO CD in CD CD CD •sr CD m CD CO CD CO CD CD in CD CO CD co CD 26.1 23.2 18.9 16.6 21.2 22.2 12.5 14.1 o CO 14.2 Specimen C-RT1 I C-RT2 C-RT3 C-RT4 Average C-W1 C-W2 C-W3 C-W4 Average C-PH1 C-PH2 C-PH3 C-PH4 Average FC-RT1 FC-RT2 FC-RT3 FC-RT4 Average FC-W1 FC-W2 FC-W3 FC-W4 Average 141 to c cu E co a> 3 Q. (0 X © UJ CC IN •sr in CN T— •sr •sf o CN CO LO CN 00 •sr Is- in CO CO Is-LO CO CD d CN m d in m d CD in •sf CD in d CD •sr •sf CD -sf csi CN LO Is-: CN LO r^ m c o •sr o> CO m CO LO d CN in d CO in Is-o CD d Is-•sr d LO LO o co c o •sr c o o Is- •sr CO co CD Is- CD CN LO CD o oi in •sr •sr CN LO csi 00 •sf csi CD •sf 1^  CO •sf d LO d o •sf d CD •sr •sr LO •sr •sr CN in Is-•sf •sr co CO •sr •st Is-•sr d cn d d 00 •sr d Is-•t d CD •sr o LO c o o> oo CN Is- Is- in o CN CO CO CD •sr CO o CD csi CO •sr csi •sr ^r d o •sf •sf Is-•sf d m •sf d CO CO d CD CO LO •sf csi CD CO d o •sr d CO •sr co •sr d CO m r-CN •sr 00 •sr d CO •sr r-: CD •sr d CN in d LO d a in CN >sr CO Is- Is- 00 CD co •sf CD o CD •sr CO LO CN LO CN d CN d T— d CN d CN d •sr d LO d CO d CO •sr CN d CO csi r o co CO Is-: CO •»— CN d CO d CN o c o CD Is- co •sr Is- CO Is- CO CD CD CD o Is- in Is- CN d CN LO CN Is-CN 1^  •sf CN CO d d CN CN d CN d CO •sr d CN •sr •sf CO CO csi CN d CN d CN d CN in CD in CD m CN o CD Is- m oo _ CD c o o CD Is-o CO d CN •sf CN T— d d CN d d 00 d CN Is-T— •sr T— d d d d d d d FC-PH1 FC-PH2 FC-PH3 FC-PH4 Average PFC-RT1 PFC-RT2 PFC-RT3 PFC-RT4 Average PFC-W1 PFC-W2 PFC-W3 PFC-W4 Average PFC-PH1 PFC-PH2 PFC-PH3 PFC-PH4 Average 142 ? S §? I 143 144 T J CD Q . E CO •o •a o o 3 TJ CD re • CO E E 3 C O 8 CS Q Total Ranking in o CM CM o CO CO *-E P w CO - 7 LO o CM Initial Stiffness, kN/i Days Exposed o o> CN CM o CO CM o o CN Initial Stiffness, kN/i Days Exposed o CO O CM CM 7 7 CM CM Initial Stiffness, kN/i Days Exposed o co V CO 7 CM CM 7 CO CM Deflection, mm Days Exposed w CNJ CM in CO o CM 7 Deflection, mm Days Exposed o o O 7 O O CN CM o Deflection, mm Days Exposed o CO 7 CO CM CM CO o Deflection, mm Days Exposed o CO CO CN 7 Fracture Energy, Nm Days Exposed w CNJ CO CM CN 7 Fracture Energy, Nm Days Exposed o o> o 7 CM o o CO Fracture Energy, Nm Days Exposed o CO o 7 CO O CM o CO 7 Fracture Energy, Nm Days Exposed o CO CO CM o O 7 o CN 7 Ultimate Load, kN Days Exposed <9 CM CO CN -Ultimate Load, kN Days Exposed o OI 7 CM CO o CO O CM Ultimate Load, kN Days Exposed o CO o CM 7 O o CN CM CN O Ultimate Load, kN Days Exposed o CO Csi CO o CM 7 CN O 7 Cracking Load, kN Days Exposed w 7 CO CM •* CM CN LO O o Cracking Load, kN Days Exposed o cn - T CM CO O CM CM - -Cracking Load, kN Days Exposed o CO o CM 7 o 7 - CM 7 Cracking Load, kN Days Exposed o CO CM CO - - 7 7 CM o Specimen C-RT C-W 1 C-PH FC-RT 1 FC-W FC-PH PFC-RT PFC-W [ PFC-PH c CS a. TJ C (0 c o CO "C CS Q . E o o CO > O • TJ CD a E <o T J C cs TJ « U 3 TJ CU • CS E E 3 C O 3 CS a Overall o CM CN in o in CO Overall C-W C-PH C-RT FC-W PFC-PH FC-PH PFC-W PFC-RT FC-RT Stiffness in CO 7 o - - CM ••»-Stiffness PFC-RT C-W FC-W PFC-W C-RT C-PH FC-PH PFC-PH FC-RT Deflection LO CM CM 7 o CM CO Deflection C-PH C-RT C-W PFC-PH FC-PH FC-W PFC-W FC-RT PFC-RT Fracture CO CM 7 CM CM •t Fracture C-PH C-W C-RT PFC-PH FC-W FC-PH FC-RT PFC-W PFC-RT Ulitmate CO CM CM CO Ulitmate C-W C-PH C-RT FC-RT FC-W FC-PH PFC-PH PFC-W PFC-RT Cracking CO CN CM 7 o o CN in Cracking C-W C-PH FC-W C-RT PFC-W PFC-PH FC-PH FC-RT PFC-RT Overall Days Exposed w in o CM CM o in CO Overall Days Exposed o o> in CO o> LO o CO CO Overall Days Exposed o CO CO CN CD CN in "3-Overall Days Exposed o CO 7 CM o r^ - CM CO CM 7 Specimen C-RT 1 C-W 1 C-PH 1 FC-RT 1 FC-W 1 FC-PH PFC-RT 1 PFC-W | PFC-PH 145 A P P E N D I X D Plots Used for Analysis 146 to >. ro Q i o ro 0> =5 (0 o q. x lu q. £ H £ o o a: o *-* sz o o in CO c CO N>| 'peon 147 CO ro a Qi i _ 3 CO O a. x UJ ro a> a. E a> I-E o o cc o o I <0 (0 CD c cn ro i n o o o m 'peon 1 4 8 L O CN O CL CL CL o o o •st i— cr o CD CO CO CO O O O O c o C M -sr I— I— I— CL CL CL o o o co co co O O O CD CO CO CU CD 0 c c tz in ro Q o CO I 22 3 in O a. x LU 0 k. 3 0 Q. E 0 I-E o o CL c o o 0 CO CN h-o o CN O L O o d E E c o u T— 0 o c r-; 0 ro Q. in T3 L O o o L O o o NM ' P B O l 149 00 o oo CD CSl oo co >. ro G o co co O CL X UJ CP a. E a> I-£ o o I a co (0 CO cu £ fc CO "(0 oo CM "** CM d + X O) o c r i T— CD II CM I-QL DL QL L t o o o o CO CO CO CO O O O O L L L L L L L L CD CO CD CU CD CU c c c LO CM O CM O LO E o E O £ o o a> * -CD a CO a. (0 o •© d "5 LO o o o LO o o m 'peon 150 Q o CD <D L-3 U) O a. x LU CD i_ 3 4-* CO k. 0) a E o I -E o o cn I CL CO (0 to e fc CO NM 'Peon 151 in CM O d IS-CD CO X cn o co -sr CM o CO CM CO I— cr cr cn cn o o o o oo co co co O O O O U L L L . L L . L L . cc o CO o CD CD c cr o oo O cr o CO O CD CD cr o CO O CD 0 C CM O to o CO 3 co o CL X LU 3 ro i _ 0 CL E 0 tr E o o rr • CL CO to 0 c CO "co N M ' P B O T 152 cn ro cx> •sr d •sf CO CSI CO o d 1- csi co •sr I — r — I — I  Cd Cd Cd Cd o o o o CO CD CD CO O O O O <0 CO Q CD i _ 3 SO o CL X UJ CD CL E CD E o o I a CO TJ CD CO CO CD i -0. I to CO CD £Z CO "(0 QL QL Cd Cd o o o o CO CD CD CD O O O O L L L L L L L L CD CD CO 0 OJ 0) LO II LO CM o Csi o o d E E c o o CD a c CO a. co T3 o o LO o o CM 00 CD 'st NM 'PBOT CM 154 It) CO It) o CL X LU a co i— cu CL £ CD r-£ o o ac i CL CO TJ C U It) II) cu 0_ I It) It) 0) c fc CO "to £ £ £ o **-o _cu cu o c CO CL It) T3 155 to o o d m 'peon 156 157 co -sr T-5 £ 5 CN cn CN o o i- CN co -sr o o o o OO OO CD CD O O O O o o CD CD o o CD CO CD CD c c o CO o o CO O o CO O — CD CD C CN O cn o d E E c o o CD Q c ro D-CO TJ O O LO o o CN 00 co CN m 'peon 158 LO CO O o o o o CO CO CO co O O O O — T — ^ O to •sr CM o Wi 'pecn CM CO 159 CO T - CM 5 5 5 5 o o o o CD CO CO CO O O O O LL LL LL LL 5555rrrr O O O O C 0 C 0 C D C 0 C D C D C D C O C D C D C D C D O O O O = .E .£ £ to >. CO Q o CD s 3 CO o CL X LU CL CO CL CO m CO LO CM O CM O L O o o E E o a> Q c CO a CO "O o d L O o o CM CO CD NM 'peon CM 160 CO >< CB Q to o Q. X UJ CO Q. CO Q . CO E £ e o o O cu Q c ro a to • D m 'p«n 161 L O CM O CM CO -fl-5 £ £ o o o CO CO CO O O O T - CM CO ^ U_ LL. LL § 5 § I t t t o o o o . . . r? J? f? • 9 * O O O O CD CD cu LL LL c c g D . 0. Q . 0. _ l _J _| ID >. CO Q o CO CO o Q. X UJ l_ CD +J CO Q. ro Q. CO T J CD CO CD 10 CD CO CO CD c +^ CO o CO u LL CL C M O LO o d E E o 0) CD Q c CO CL CO T J o d LO o o CM 00 CO CM m 'peon 162 163 co -sr o o CJ) Oi Oi O) O O O O co ro Q O i_ 3 (0 o CL X L U i _ ro CL ro CL ro § TJ CU (0 CO cu s -CO £ Q . I to to cu c fc CO "ro E E c o HS u _cu 4 -0> Q C CO C L co TJ m 'peon 164 CN CO X X X L 1 1 o o o CO CO CO X 0_ o CO co -sr T-X X X CL CL CL o o o CO CO CO o o o CO CO CO d ) CU CD c c c 0 0 0 0 _ I _ I _ I _ I co >. CO Q o CO ( 0 o C L X 111 o o co co CD C |fc co CM X CL O CO O i _ CO CD CZ t o CM o d CO LO CD co d + X LO d I S-II CM O LO o d E E E O u CD a c ro CL ( 0 T J o o L O o o CM CO C D C M NM 'Peon 165 m 'peon 166 m CN o d I D_ O CN CO X X Q. o CN x x o o cn cn O O CD CD m CO CD oo cn oo oo c tz O O O O J J X D_ o cn O t_ CD CD £= CO X CL O cn O i ro CD co >» CO Q o cn to o C L X UJ o o (0 to CD c |C CO CM O L O o d E E c o o 0) **-CD Q c CO a o d m o o CM CO CD m 'peon CN 167 i o CM o CJ) o o C O X o C M C O CM CO X X D _ o o o o o LL LL LL LL CO >< rs Q o ro • OJ i_ 3 CO O EL X LU CL rs co (0 cu CO TO • s f CO X X 0- L L . O O CO CO CO O O O L L L L L L CM X CL O CO CO CD CD CD CD c c c X CL O C O O CM O LO o d £ £ c o Cw o cu Q c CO CL CO T J O d i n o o o CM CO c o m 'peon CM 168 CO CO O in o o_ x LU CL fO in in CO ]E 'E 169 CM CO CO -<3-m 'peon CM 170 in C M O 5* co Q o C O CO o CL X LU I C L i: CL co T3 0 CL CO It) 0 c CO re CD CM C O O C M C O X X 0- L L . O O T - C O - sr C M X X X X D_ L L . Q. CL O O O O C O C O C O C O O O O O O O O O L L DL co 0 L L L L 2_ CL co co 0 0 cr. c m 'peon 171 ro O d T - CN CO •<*• X X X X 1 CL CL D. o o o o CD CD CD CD O O O O CO >. CO Q o CO co o CL X LU CL CO (0 O a -+-» CO CD k . 0) c !t t- CN co -sr X X X X CL CL D. CL o o o o CD CD CD CD o o o o LL LL LL LL CL CL a, 1 1 1 s s co co co CD co cu a> c c c CD LO LO + X CO CD CD LO CN O CN O LO O d £ £ c o u _co co Q c CO CL o d CO o o CM CO CD CM m 'peon 172 LO CN O T - CN CO ' S f I X X X 0_ CL CL CL o o o o 0) OO OJ CD O O O O CNJ •sf CO X X X X C L C L C L C L O o O O CD CD CD CD O O O O LL LL LL LL C L C L C L C L • — * — * — " — i _ CD CD CD CD CU CD CD CD c C c C _ l 1 _ l 1 _1 1 _ l 1 CO >» CO Q 3 CO O Q . X LU I a. • Q. co co co CD u . +-> CO CO co CD c fc CO "re CNJ o d LO o d E E c o ' o _CD <*-CD Q c ro a (0 TJ O d LO o o CNI 00 CO ' s f m 'peon CNI 173 C N r - CM CO s f oc a: cn o o o CO CO CO CO O O O O CH o oo C D CO CO CO o C L X U J CJ a -3 *J CO 1-CD C E CD t— E o o CL 2 c o o E E u CD o co C L CO •o 00 d CD Q CD O E CD > d CNI O o o CO o CM Wi 'peon 1 7 4 CM CM co •sr H t- t~ *-cn cn or cn o o o o CO CD CD CD o o o o CQ Q o CO as *_ 3 CO o a x LU CD CD CL E CD t— E o o o o o CJ CD D CO > TJ co O E E £ o Cw CJ a> «+-CD Q c co CL CO TJ NM 'Peon 175 co > . co Q 3 in o Q. X LU o 0) Q. E o r— £ o o 0-o o o o o o Q !2 cu > X3 CO o CM T - CM co -<r I- I- I- r -01 DL DL DC o o o o cj) oo CTJ cn o o o o oo CD E £ o 0) D ro D_ CO T J C O CD O d CM o NM 'Peon 176 177 178 CM T - CM CO -sr I— I— r— I— Ct Lt Ct Ct o o o o o o o o o •sf CO CM m 'peon 179 CO Q 3 CO o Q. X LU OS ro L. OS Q. £ GS r— £ o o DL a co as +-* to GS o CO 3 as > T3 CO o £ £ c o +s o as a c CO a . CO TJ m 'peon 181 182 CM i - CM n ^ § § 5 § o o o o CO CO CO CO O O O O CO CD CO >< ro Q o ro CD h-3 CO O O-X LU CO CL ro r-O o CM E E c o o o 4 -d) Q c CO CL CO TJ CO o CJ 0> "«-0) O CO 3 cu > CO o CD o d CM d o o CO o CM m 'peon 183 CM T - CN CO ^ 5 § 5 § o o o o co co co co O O O O CO C D CO >» CS a o CD (0 o Q. X LU co t a CO O C o o CM E E c o *3 O 0) *-a> Q c co Q. (0 TJ CO o o o •4— a> a CJ > TJ CO O CD d N" d CM d o o CO o CM m 'peon 184 T - <N CO ^ 5 g 5 5 o o o o CD CD CD CO o o o o oo C D CO > < CO Q CO o CL X LU CM £ £ CL CO o 0) Q O O :i c o CO CL (0 T J CO d co 52 > CO o CO o o C M d o O CO o CM m 'peon 185 CM 1 - IN CO Sf 5 5 5 5 o o o o CO 00 CO CO O O O O LL LL 11 LL 00 C D 186 187 188 NM 'Peon 190 CM o o o o CO CO CO CD O O O O oo U _ L L . L L . U _ r • CL CL a. a. • CD o o o o o -t ro CM - -NM 'peon 191 10 (T3 X 5 O CO to o a. X LU I a. o (J o i t -CD Q a> > to o CM CM CO X X CL CL O O CO CO CO CO O O O O X CL O •st X CL O oo CD CM E E E O Z-o a> t-a> a c CO Q . CO d CD CD O CM O m 'peon 192 CM CO >» re X 5 O CO _5 co o a x UJ I a i o o a u **-a> Q e CD > -o C8 O X a. o CM CO X X D_ o X a. o CD CD CD CO o o a o CO CD CM E E s£ o o 0 CD Q c co CL CO co d CD CD o' CM o o - 4 " o co o CM NM 'Peon 193 1- CN CO X X CL Q_ o o CL o X CL o 0 ) 0 3 05 0) o o o o 00 CD co >. CO T3 o Oi CN 3 CO o o_ X LU £ E E O o o o _0) I t -CD Q c co a CO •a o a> n -0) a CO 3 CO o > T3 CO O 00 o CD O O CN O o o CO o CM m 'Peon 194 >» CO Q o CO CO o a x UJ D. CL CO o _0> »»-0) o CO a> > T J CO o E E c o u a> Q c co CL-IO T5 NM 'Peon 195 CM i- CM tO t X X X I a . c_ a . n . o o o o CD CO CD CO O O O O LL. LL. LL. L L o o o o o - J - CO CM T -m 'peon 196 CN T - CN CO ^ X X X I Q_ Q- D_ 0_ o o o o 00 03 0) 0) o o o o o o o o o o CO CN T -m 'peo-| 197 CO >> CO Q o CO CA O a. x LU Q . ( 0 CO CD c o o _Q) i t -CD a CD > Ti-ro O CN OO CD E E o a> M -<D Q ro Q . CO T5 00 d CD d o CN d 198 (0 >> rs Q 0) k . 3 to o CL X LU CL ro TJ CO CO CJ o 0) Q (0 3 CO v > TJ ro O E E o o « Q c ro CL u> TJ NM 'Peon 199 i- CM r o s f D_ a. D_ D_ o o o o C> CD CO CO o o o o L L L L L L L L CL GL CL CL CD 200 201 CM 0) NM 'Peon 202 CM 00 0) o o o o o •<J- CO CM T -m 'peon 203 204 CM NM 'peon 205 CM CO CD CD -3 NM 'Peon 206 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0063492/manifest

Comment

Related Items