- Email: [email protected]
Use of FRP composites in civil structural applications
Construction and Building Materials 17 (2003) 389–403
Use of FRP composites in civil structural applications Lelli Van Den Eindea,*, Lei Zhaob, Frieder Seiblec a
Department of Structural Engineering, University of California, 9500 Gilman Drive, San Diego, LaJolla, CA 92093-0085, USA b Department of Civil and Environmental Engineering, University of Central Florida, Orlando, FL 32816-2450, USA c Department of Structural Engineering, University of California, 9500 Gilman Drive, San Diego, LaJolla, CA 92093-0085, USA
Abstract Fiber reinforced polymer (FRP) composites or advanced composite materials are very attractive for use in civil engineering applications due to their high strength-to-weight and stiffness-to-weight ratios, corrosion resistance, light weight and potentially high durability. Their application is of most importance in the renewal of constructed facilities infrastructure such as buildings, bridges, pipelines, etc. Recently, their use has increased in the rehabilitation of concrete structures, mainly due to their tailorable performance characteristics, ease of application and low life cycle costs. These characteristics and the success of structural rehabilitation measures have led to the development of new lightweight structural concepts utilizing all FRP systems or new FRPyconcrete composite systems. This paper presents an overview of the research and development of applications of advanced composites to civil infrastructure renewal at the University of California, San Diego (UCSD). 䊚 2003 Elsevier Ltd. All rights reserved. Keywords: Advanced composites; Carbon shell system; Modular structural systems; Retrofit; Repair; Strengthening
1. Introduction The application of advanced composite materials in civil engineering has been evolving slowly, primarily due to economic reasons. Their key advantages, such as free form and tailorable design characteristics, strengthto-weight ratios that significantly exceed those of conventional civil engineering materials and a high degree of chemical inertness in most civil environments are lost in high material and manufacturing costs. Furthermore, the current practice of one-to-one component replacement of elements in conventional structural systems by advanced composite components has shown that it is difficult to justify the use of advanced composites in civil construction, not only economically, but also structurally. Several developments have changed this situation over the past few years: (1) techniques such as pultrusion, resin transfer molding, filament winding and semiautomated manufacturing of large components have led to advances in low cost FRP manufacturing; (2) reduced material demand in the high priced defense industry, expansion of a highly competitive market for these *Corresponding author. Tel.: q1-858-822-2188, fax: q1-858-8222260. E-mail address: [email protected] (L. Van Den Einde).
materials in the sporting goods industry and prospects for large volume use in the civil sector have led to new low cost materials manufacturing; and (3) designs of these new materials in conjunction with conventional structural materials rather than individual component replacement or complete advanced composite designs, have shown that technical efficiency can be achieved within competitive economical constraints. The emerging field of renewal engineering may best describe the role of FRP composites in civil engineering. The renewal of the structural inventory, which is depicted in Fig. 1, can be divided into (1) rehabilitation, including the applications towards repair, strengthening and retrofit of structures; and (2) new construction with all FRP solutions or new composite FRPyconcrete systems. The structural effectiveness of FRPs in the rehabilitation of existing structural systems has repeatedly been demonstrated with full or large-scale structural tests at the University of California, San Diego (UCSD). Carbon fabric overlays have been used to strengthen and retrofit reinforced and unreinforced masonry walls for seismic loads, as well as to restore and more than double the displacement capacity in the repair of a fullscale five-story reinforced masonry building tested to failure under simulated seismic loads. Carbon fiber overlays and strips have also been used to strengthen
0950-0618/03/$ - see front matter 䊚 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0950-0618Ž03.00040-0
390
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 1. Use of fiber reinforced polymer composites in civil engineering.
reinforced concrete slabs (with and without openings) and to strengthen large diameter prestressed concrete pipelines to restore loss of load (water pressure) carrying capacity due to corrosion of the prestressing wires. Bridge columns have been seismically retrofitted and repaired with fiberglass, carbon and hybrid composite jackets, which were shown to be as effective as conventional steel jackets. The success of structural rehabilitation measures with advanced composite materials has led to the development of new lightweight structural concepts utilizing FRP shells and tubes to form new structural systems. FRPyConcrete composite systems have been developed for use in new lightweight bridge systems due to their simplification in construction and excellent short- and long-term structural characteristics. An overview of past and ongoing research at UCSD on application projects of advanced composites in civil engineering is provided in this paper to show the extent of current developments and to explore the feasibility of future applications in civil engineering. A prognosis for the application of advanced composites in civil engineering construction is made by outlining critical technical and implementation issues to be resolved prior to a broad acceptance by the civil engineering design and construction community. 2. Repair of constructed facilities The effectiveness of the application of advanced composite wall overlays for seismic repair and retrofitting of structural wall systems can best be demonstrated by the example of a full-scale five-story reinforced masonry building (see Fig. 2), which was tested under pseudo-dynamic simulated seismic loads at UCSD w1x. Following the initial failure loading, the crushed wall toes in the building were reconstructed with polymer concrete and the building was repaired using structural carbon composite overlays on the first two stories of the structural walls (see Fig. 2b). Only one layer of the 12 K tow uni-directional woven AS4 carbon fabricy epoxy resin composite with a nominal thickness of 1.25 mm was applied except at the toe regions, where an additional layer of the same composite was used.
The load–deflection envelopes for the original and repaired five-story building are shown in Fig. 3. The repair scheme contributed significantly to the system ductility by almost doubling the inelastic deformation capacity in the critical push direction. Measured shear deformations in the overlaid wall panels were reduced to half of those observed in the original test. Detailed information of the repair and retrofit installation and seismic response data can be found in w1x. 2.1. Strengthening of walls and slabs 2.1.1. Structural wall overlays Strengthening of structural walls can be accomplished very economically with thin advanced composite wall overlays. To date, tests have focused on (1) reducing shear deformations in structural walls; (2) retrofitting shear walls to achieve ductile in-plane behavior; (3) repairing damaged walls to increase in-plane ductility; and (4) retrofitting out-of-plane unreinforced structural walls. Seven single-story full-scale structural wall panel tests were performed at UCSD on fully grouted hollow core concrete masonry walls w2x. The overall test setup for these wall tests is shown in Fig. 4a. 2.1.1.1. Shear strengthening. Very thin overlays (only one or two layers with thickness of 0.5–1.0 mm) can achieve significant seismic improvements especially for in-plane shear wall response. Carbon fibers in the composite overlay are oriented horizontally to cross diagonal or shear cracks, while allowing horizontal or flexural cracks to open. Forces to be transferred in the composite overlays are limited by the laminar shear or principal tensile strength of the existing structural wall material, since the polymer resin typically features significantly higher tensile capacities than the concrete or masonry substrate. To improve the shear capacity of structural walls with advanced composite overlays, the following design assumptions can be made: a conservative diagonal tension crack angle of 458 can be selected and the resulting shear capacity increase can be determined based on an allowable overlay stress level derived for a maximum
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
391
Fig. 2. Five-story full-scale building test.
horizontal wall strain of 0.004, above which, aggregate interlock is assumed lost. Alternatively, stiffness criteria can be employed in the wall overlay design, limiting
shear deformations to levels that can be expected in concrete walls with conventional horizontal reinforcement (determined based on conventional design
Fig. 3. Load–displacement envelope comparison for carbon overlay repair test.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
392
Fig. 4. In-plane strengthening of structural walls.
requirements). To determine the amount of horizontal overlay fabric Aoh, the equivalent required horizontal steel reinforcement Areq sh is scaled based on the ratio between the steel stiffness and overlay stiffness as shown in Eq. (1). Aohs
Esh req Ash Eoh
(1)
where Esh and Eoh represent the horizontal steel and composite overlay laminate moduli, respectively. Three full scale 2=2 m structural wall panels were constructed with 152 mm thick fully grouted and reinforced concrete block masonry using reinforcement that was 19 mm in diameter in the vertical direction and 9.5 mm in diameter in the horizontal direction. Both vertical and horizontal reinforcement had a spacing of 400 mm and a nominal yield stress of 414 MPa (see Fig. 4b). The walls were tested in the (1) as-built condition; (2) repaired condition following diagonal shear failure; and (3) retrofitted condition. The repair of the original wall panel consisted of epoxy injection of the diagonal shear
failure crack, reconstruction of the crushed compression toe with polymer concrete and horizontal carbon composite overlays with a layer of 490 gym2 unidirectional carbon fabric on both sides of the wall. The retrofit specimen was a new as-built wall with two layers of the same horizontal carbon composite overlays only on one side of the wall. Thus, the amount of horizontal carbon fabric was the same in both cases (see Fig. 4c). For the horizontal cyclic load test, the wall panels were first loaded with a 1.8 MPa axial or vertical load and subsequently restrained against rotation at the wall top. The failure mechanism for both the repaired and retrofitted walls were similar and consisted of separation of the face shell from the core at the compression toe regions, which resulted in the bulging out and delamination of a large region of the carbon overlay and crushing of the masonry units. Envelopes for the load– deflection hysteresis curves for the three tests, which are depicted in Fig. 4d, show that both the double sided repair and single sided retrofit with horizontal carbon overlays resulted in twice the deformation capacity of the as-built wall specimen. The improved performance
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
393
Fig. 5. Out-of-plane retrofit of unreinforced walls.
of the repaired wall specimen (over the other two tests) can be attributed to the enhancement of the wall toe that was rebuilt with polymer concrete, which has a higher compression strength, following the initial failure test. 2.1.1.2. Flexural strengthening. For out-of-plane wall strengthening, the key design aspects are (1) to achieve force transfer from the substrate to the overlay material in regions of high moment; and (2) to avoid the potential risk of buckling and delamination of the thin and stiff overlays on the compression side of the flexural member. Two out-of-plane flexural wall tests were also performed at UCSD on 2=2 m unreinforced fully grouted masonry wall panels, see Fig. 5a, with the same geometric and axial load conditions as the shear walls described in Fig. 4. To provide flexural capacity to the wall, a single vertical layer of carbon fabric overlay was developed at the base of the cantilever wall. End anchorage or fabric development was achieved with four different anchorage types. The two concrete anchorage solutions both incor-
porated bolted base beams and used different wraparound curvature for the carbon fabric (see Fig. 5b), while the two steel anchorage solutions consisted of using either a bolted steel angle or individual steel straps (‘Simpson ties’), as shown in Fig. 5c. The steel angles were placed on top of a carbon fiber overlay while the steel straps were sandwiched between a carbon base strip and a carbon fiber overlay. The cantilever load–deflection envelopes for the wall retrofit schemes with the four different connection details are depicted in Fig. 5d. In the first quadrant, the dashed line represents the ‘Concrete Curve’ case, while the dotted line represents the ‘Steel Angle’ case. In the third quadrant, the dashed line represents the ‘Concrete Fillet’ case and the dotted line represents the ‘Simpson Tie’ case. By far, the largest deformation capacity was obtained using individual steel straps on the tension side as anchorage, since the right-angle-bent steel straps (‘Simpson Ties’) straighten out with increasing cantilever loads. Lateral load capacities do not depend on the anchorage type as long as the full carbon fabric capacity
394
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 6. CFRP strengthening of large-scale slabs with openings.
can be developed. Test results show that the design of flexural carbon overlays is controlled by strain limits on the compression side, which is less than half their coupon tensile capacity. As a general rule of thumb, a limit of one-third the ultimate coupon tensile strain, ´ou, should be applied. 2.1.2. Slab strengthening The use of Carbon FRP (CFRP) composites in the form of strips or overlays, offers a cost-effective technique for the flexural strengthening of substandard reinforced concrete slabs. The efficiency of this application has been demonstrated through numerous experimental investigations and field applications, e.g. w3,4x. A recently completed experimental test program at UCSD investigated the effectiveness of the CFRP strip strengthening method on large scale slabs weakened by the introduction of a rectangular hole at the center of the slab w5x. A series of four tests on large-scale slabs with a rectangular cutout in the center retrofitted with CFRP strips was conducted to evaluate the use of this strengthening method for typical slab upgrades. Of the four test units that were considered for this project, two slabs were used as base-line references, one loaded in three
point bending and the other in four point bending. The remaining two specimens were reinforced with carbon strips in the two principal directions and were tested in the two respective loading configurations. The dimensions of the slabs considered were 6 m in length, 3.2 m in width and 0.18 m in thickness. A rectangular opening measuring 1 m in length and 1.6 m in width and centered with respect to the slab was later introduced. The dimensions of the cutout were chosen as half the width of the slab such that, the main reinforcement ratio was reduced by half and 1 m in the longitudinal direction to maintain a reasonable strength capacity. The test setup for the two flexural slabs tested in three-point bending is shown in Fig. 6a. A steel distribution beam over the cutout allowed introduction of load at mid-span. An overview of the test results for the three- and four-point bending tests is provided by the load–displacement curves in Fig. 6c and d, respectively. The figures demonstrate that the ultimate strength of the slab was increased to more than twice that of the weakened slab with cutout. Under both loading conditions, the strengthening measure was successful in recovering the initial strength of the previously tested slab without the cutout w5x, while also increasing the stiffness. In addition to a higher load carrying capacity, the CFRP strength-
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
395
Fig. 7. Carbon fiber jacket shear column retrofit test.
ening reduced the strain levels in the steel reinforcement and enhanced the development of a more evenly distributed crack pattern. Failure of the strengthened test units was due to delamination or peeling-off of the strips (see Fig. 6b). 3. Seismic column retrofit Recent earthquakes in California such as the Wittier 1987, Loma Prieta 1989 and Northridge 1994 have repeatedly shown the vulnerability of existing bridge columns built before the 1971 San Fernando Earthquake. Tests on 0.4 scale bridge columns at UCSD have shown that carbonyepoxy jacket retrofits can be just as effective as comparable steel jacket retrofits w6x. As an example, the dimensions and reinforcement layout of one shear column test unit are shown in Fig. 7a together with the
carbonyepoxy jacket design of variable thickness. A jacket thickness of only 0.4 mm was required over the shear critical center region of the column to prevent brittle shear failure. Experimental load–deflection hysteresis loops for the carbon fiber retrofitted shear column are shown in Fig. 7b. Completely stable hysteresis loops up to a displacement ductility level of mDs10.5 were achieved at which point the test was terminated due to test setup limitations rather than retrofit failure. Comparative load–deflection envelopes for (a) the unretrofitted or ‘as-built’ shear column; (b) a companion steel jacket retrofitted column with 5 mm jacket thickness; and (c) the carbonyepoxy retrofitted column are depicted in Fig. 7c. Clear improvement of deformation capacity can be seen for both the steel and carbon retrofitted cases over the ‘as-built’ case, which failed in brittle shear at a displacement ductility level of mDs2.0.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
396
Fig. 8. Modular FRP deck test.
The test results in Fig. 7b and c show that the column retrofitted by a steel jacket exhibited a slightly higher initial stiffness and a small increase in lateral load carrying capacity with increasing displacement levels due to the isotropic nature of the steel jacket, resulting in a more concentrated plastic hinge and more strain hardening at the column ends. Both stiffness and capacity increases are not necessarily desirable in bridge column retrofits, since typically higher seismic force levels are transmitted to adjacent structural elements. Thus, the carbon jacket with only horizontal or hoop directional strength and stiffness can better accommodate the requirements for minimal stiffness or strength increase than steel jackets. FRP jackets have also been shown to be effective for rectangular columns. For most column retrofit designs, the stiffness of the overlaying jacket material controls the design. In general terms, the required jacket thickness follows the relationship found in Eq. (2).
tjf
D C Ej
(2)
where D is the column dimension in the loading direction, Ej is the jacket or overlay modulus and C is a general coefficient w6x depending on the retrofit demand. While FRP jacket retrofitting clearly provides sufficient structural effectiveness based on short-term load tests, appropriate reduction factors for differences and uncertainties in the materials and lay-up systems, jacket curing and durability considerations need to be applied for actual retrofit application.
4. Structural replacement—new bridge systems 4.1. All FRP systems The advantages of FRP composites make them attractive for use in replacement decks or in new bridge systems as well. In addition to the potential lower lifecycle costs, FRP decks would be significantly lighter resulting in large savings in column and foundation costs and enabling higher live load levels through the replacement of heavier decks. FRP deck systems also have a high application potential in areas where longer unsupported deck spans are necessary or where lower weight would translate into lower seismic demands. Research at UCSD on full-scale FRP deck elements has indicated that appropriate performance levels can be achieved at significantly lower weights w7x. The challenge is to optimize the configuration and use of advanced composite material to match both the performance and cost of reinforced concrete decks. Initial tests on a number of different configurations show that FRP decks are capable of much higher failure loads with comparable stiffness at service demand. Fig. 8 shows the testing configuration and gives representative loaddeformation results of the testing program conducted at UCSD on modular E-glassyvinylester deck panels composed of pultruded cores with hand lay-up face sheets. Additional tests have been performed at UCSD on FRP decks with cores made of (1) endgrain balsa; (2) foam cubes made of triangular and trapezoidal prisms; (3) corrugated premanufactured glass sheets; (4) pultruded I and tubular shapes; and (5) aluminum honeycomb. Manufacturing methods include, but are not limited to pultrusion with hand lay-up top and bottom
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
397
Fig. 9. Concrete filled Carbon Shell System (CSS).
sheets, vacuum assisted resin transfer molding (VARTM) and complete hand lay-up w7x. The high material costs of the advanced composites virtually limit the material choice for replacement bridge decks to EGlass and even then cost competitiveness with conventional concrete or orthotropic steel bridge decks is difficult to achieve. 4.2. Innovative FRP-concrete composite systems The development of new structural concepts and systems, that combine the directional and tailorable characteristics of FRP composites in tension with the dominant characteristics of conventional structural materials such as concrete in compression, show great potential for advances in the design and construction of new civil engineering structures. 4.2.1. Carbon shell system (CSS) The first modular FRP-concrete composite structural system consisting of concrete filled carbon shells has been developed at UCSD w8x and is schematically shown in Fig. 9. The concrete filled Carbon Shell System
(CSS) concept creates new structural elements by using prefabricated filament-wound carbonyepoxy thin shells filled on-site with concrete. The shell serves the dual function of longitudinal and circumferential confinement reinforcement and stay-in-place formwork for the concrete core. The concrete provides compression force transfer, stabilizes the thin shell against buckling and provides anchorage of connection elements. The design of the carbon tubes consists of carbon fibers in the longitudinal tube direction ("108 from the member longitudinal axis due to manufacturing limitations) and hoop fibers (908) in the transverse direction. Circumferential ribs are provided on the inside of the carbon shell for full force transfer between the concrete infill and the shell. The concept is suitable for both columns and girders. For the development of bridge superstructure components, the concrete filled carbon shells are combined with a structural deck system, which may consist of either a conventional cast-in-place reinforced concrete (RC) slab or a modular FRP deck system as shown in Fig. 10. The connection between the deck and the carbon shell girders is accomplished by embedding shear
Fig. 10. Modular CSS beam and slab assemblages.
398
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 11. Beam-and-slab assembly test (‘3-girder’ system test).
connector dowels into the carbon shell system during grouting. In the slab, the dowels are either cast directly into the RC deck or anchored in polymer concrete filled sections of the cellular E-glass deck systems. The modular FRP E-glass deck consists of pultruded trapezoidal elements bonded together and overlaid with continuous face sheets under factory conditions. Typical composite decks weigh 1–1.5 kNym2 or 1y4 of a conventional reinforced concrete deck, resulting in significant weight savings for construction and mass reductions for seismic zones. 4.2.1.1. ‘3-girder’ system test. To validate the overall performance of the CSS system, a full-scale superstructure system consisting of three carbon shell girders and a FRP deck system was constructed and tested in the laboratory at UCSD w9x. The test specimen simulated one half of the width and the distance between the support and the inflection point for one span of the prototype Kings Stormwater Channel Bridge in California. Characterization of the system durability in terms of structural integrity and damage tolerance was investigated by subjecting the bridge superstructure first to stiffness characterization tests up to service load levels
and then to 2 million cycles of fatigue service loading. A photograph of the test setup is shown in Fig. 11a. The loading pattern consisted of four servo-controlled actuators applying simultaneously a load of 56 kN each at a frequency of 1 Hz. This load level duplicates the shear force demand on the prototype bridge at the girder–deck interface under full service loads. Test results show that the strength of the structure did not degrade during fatigue loading. This conclusion was further validated when the stiffness characterization tests were repeated after completion of the 2 million fatigue cycles without any noticeable change in structural response. The load–displacement response of the system up to failure at mid-span is shown in Fig. 11b. The system remained primarily linear up to failure, which occurred in the top face sheet of the deck panel. At the peak load (490 kN per actuator), which is equivalent to approximately 8.8 times the service demand level, the face sheet of the deck delaminated from the core and the core material buckled near its connection with the inclined web (see Fig. 11c,d). After failure, the specimen was unloaded and then cycled three times with a load of 445 kN force per actuator. The final permanent settlement at the comple-
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
399
Fig. 12. Hybrid tube bridge system.
tion of the test was less than 5 mm. No visual damage to other components, such as girders, end blocks and connections was observed w9x. 4.2.2. Hybrid tube system (HTS) Another modular FRP bridge system has been developed based on pultrusion and hand lay-up of a carbony glass hybrid material system w8x. The hybrid tube system (HTS) beam-and-slab bridge proposes the use of hollow E-glassycarbon hybrid beams that are connected along their tops with a polypropylene fiber reinforced concrete deck system as shown in Fig. 12. The girders consist of
Fig. 13. Hybrid tube bridge system characterization studies.
pultruded or hand laid-up E-glassyvinylester rectangular sections, which are reinforced along the bottom flange with unidirectional carbon fibers. An FRP panel is snaplocked to the pultruded girders providing flexural reinforcement to the concrete slab between girders as well as stay-in-place formwork for the slab. The end hooks of the panel are anchored by filling polymer concrete in the ‘dovetail’ pockets of the top portion of the transverse girder. Prefabricated carbonyvinylester snap-in stirrups are also snapped into the grooves to provide horizontal shear transfer between the concrete deck and the girders. A parametric study of a hybrid tube bridge system was performed by evaluating the maximum simply supported span length of the system under a set of allowable design criteria for service and factored loads. The bridge system under consideration consisted of seven transversely spaced HTS girders supporting a 13m wide road surface. The results of the parameter study as shown in Fig. 13, indicate that the deflection under live load governs the design, which is common for FRP structures. 4.2.2.1. Two-girder test. A two-girder assembly test was also conducted at UCSD to verify the potential application of the HTS concept. The test specimen was simply supported and the hollow rectangular girders were locally filled with concrete near the pin supports, FRP deck panels were installed between the two girders and a polypropylene fiber reinforced concrete slab was poured on top, as discussed in Section 4.2.2 w10x. The load cases for the test were designed to evaluate the response of the system under loads equivalent to the demands of a prototype bridge and to assess potential
400
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 14. Two-girder hybrid tube system (HTS) test.
failure mechanisms and sequences in the system assembly. Fig. 14a shows a photograph of the test under a simulated wheel load condition. Results from one of the flexural loading cases are provided in Fig. 14b. The two-girder system showed significant strength reserves over the demand of the proposed prototype bridge. The system and its concrete slab behaved linearly under wheel and global flexural loads up to factored load levels. Even after the concrete slab was severely damaged at multiple locations (from punching shear tests) and showed signs of slippage (bilinear point in Fig. 14b at 60 mm), the system was still able to carry loads greater than the demand with minimal stiffness loss. The girder-to-deck connection displayed good integrity up to 150% of the factored shear demand. The deflection and strains in the deck between girders were compared with those obtained from simply supported
deck tests w10x. It was concluded that under an AASHTO HS 20 truck wheel load w11x, a significant strength reserve existed in the deck system. 4.3. Advanced technology bridge To demonstrate the application of the advanced composite bridge systems presented above, a 137 m long cable-stayed bridge supported by a 46 m high A-frame pylon was designed to be constructed at UCSD utilizing advanced composites. The bridge structure is designed for two 3.7 m vehicular lanes, two bicycle lanes, two pedestrian walkways and a utility service tunnel. Based on a preliminary study of different structural systems, the design presented in Fig. 15 was selected. The structure proposes the use of the carbon shell system (CSS) for the pylon legs and edge longitudinal girders,
Fig. 15. The I-5yGilman Advanced Technology Bridge.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
401
Fig. 16. Pylon splice connection test.
supported by a dual cable plane system. In the transverse direction, partially grouted carbon tube cross-beams are employed, which in turn support longitudinally spanning prefabricated E-glass or reinforced concrete deck panels w12x Prior to construction of the I-5yGilman Advanced Technology Bridge, prototype component and system
validation testing of all FRP members is required. The following section describes results from one of the fullscale tests that have already been performed. 4.3.1. Pylon connection test The A-frame pylon in the proposed I-5yGilman Advanced Technology Bridge is composed of 1.5 m
402
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
diameter, 9.75 m long filament-wound cylindrical carbon composite tubes filled with concrete that are spliced together up the pylon height. As part of a Validation Test Program for the bridge, a Pylon Splice Connection Test was conducted w13x to validate the design and analysis methods used to develop the connection splice reinforcement details, to examine the ductile failure of the mild steel reinforced splice section, to assess the confinement effects of the carbon shell and to evaluate strain levels in the carbon composite shell in comparison with targets set forth in the Design Criteria w14x for the bridge. The design of the pylon leg for the bridge requires that failure will occur in the splice region by yielding of the mild steel reinforcement rather than brittle failure in a typical composite section. The final design details for the ‘typical’ splice connection in the Pylon Connection Test resulted in thirty 29 mm diameter longitudinal bars with 1.83 m splice length, confined by 13-mm diameter hoops that were spaced at 0.12 m. A photograph of the test setup is shown in Fig. 16a, while the test specimen at extreme drift is depicted in Fig. 16b. The axial load was applied using four external tendons, each composed of 27–15 mm diameter strands. The force vs. displacement hysteresis is provided in Fig. 16d w13x. Also included in Fig. 16d are the SERVICE I (395 kN), STRENGTH I (622 kN) and EXTREME I (1222 kN) demand levels which were based on AASHTO LRFD toad combinations w15x for the I-5yGilman Bridge. At the end of the test, 15 of the 30 longitudinal bars in the splice region had fractured (Fig. 16c), but the overall global capacity of the specimen only degraded approximately 20%. The Pylon Splice Connection Test was successful in demonstrating that the design satisfied the performance objectives. The compression and tension strain levels in the composite shell were well below the limit states set forth in the Design Criteria w14x. The development length provided in the splice region was sufficient in transferring the forces to the composite shell. The amount of transverse reinforcement provided by the composite shell and steel hoops in the splice region was able to resist pull-out failure of the lap splice longitudinal bars and sufficiently confine the plastic hinge region for adequate ductility capacity. Preliminary assessments of the strain levels as well as acoustic observations during testing indicate that slip of the longitudinal splice reinforcement was avoided. 5. Summary Worldwide research and application of FRP composites in civil engineering construction demonstrate the need for new construction materials in support of civil infrastructure renewal. The demonstrated advantages of FRP composites have shown that they will play a significant role in future civil engineering projects. The structural effectiveness of FRPs in the repair of con-
structed facilities, strengthening of structural walls, slabs and retrofitting of concrete columns with shear, flexure and lap-splice problems has been validated with large or full scale laboratory tests. However, the use of fiber reinforced composites for the rehabilitation of structures requires that appropriate design philosophies, guidelines and detailing be established and that the design is conducted using a methodology that ensures appropriate use of the material. With suitable design criteria for FRPs in structural rehabilitation, significant advantages can be derived from the lightweight properties of these new materials and their ease of handling and installation. Innovative FRPyconcrete bridge systems, such as the concrete filled carbon shell system (CSS) and the hybrid tube system (HTS) effectively use FRPs for new construction by combining them with conventional materials such as concrete and steel. System characterization and design studies are providing the basis for design approaches for FRP bridge systems in terms of deformation and strain limit states. It is expected that modular FRP bridge systems of this type will lead to faster construction and less traffic interruption due to their light weight, as well as lower life-cycle costs due to reduced maintenance. The extent of these FRP applications in support of civil engineering renewal will depend on (1) the resolution of outstanding issues such as reparability, fire, durability and environmental concerns; (2) the extent to which automation in the manufacturing process can reduce cost; (3) the availability of validated codes, standards and guidelines that can be used as design references and tools by the civil engineering community; and (4) the degree of quality control and quality assurance that can be developed and provided during the manufacturingyinstallation phase utilizing unskilled general construction labor. Acknowledgments The authors would like to acknowledge the United States Federal Highway Administration (FHWA), the California Department of Transportation (Caltrans), the Defense Advanced Research Projects Agency (DARPA) and the State of California for funding the research reported in this paper. Furthermore, the authors acknowledge the technical staff of the Powell Structural Research Laboratories at UCSD for their support in the construction and testing of these research projects. References w1x Weeks J, Seible F, Hegemier GA, Priestley MJN. The USTCCMAR Full-Scale Five-Story Masonry Research Building Test: Part V – Repair and Retest. Structural Systems Research Project. SSRP-94y05. University of California, San Diego, La Jolla, January, 1994.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403 w2x Laursen PT, Seible F, Hegemier GA. Seismic Retrofit and Repair of Masonry Walls with Carbon Overlays. Structural Systems Research Project. SSRP-95y01. University of California, San Diego, La Jolla, January, 1995. w3x Meier U, Deuring M, Meier H, Schwelger G. Strengthening of Structures with CFRP Laminates: Research and Applications in Switzerland. Proceedings of Advanced Composite Materials in Bridges and Structures: First International Conference, Sherbrooke, Canada, 1992;243–251. w4x Nanni A, Focacci F, Cobb C. Proposed Procedure for the Design of RC Flexural Members Strengthened with FRP Sheets. Proceedings of Second International Conference on Composites in Infrastructure, Tucson, Arizona, 1998;187–201. w5x Vasquez A. The use of carbon fiber reinforced polymer strips for the external strengthening of slabs. San Diego, La Jolla: M.S. University of California, 1999. w6x Seible F, Priestley MJN, Hegemier GA, Innamorato D. Seismic retrofit of RC columns with continuous carbon fiber jackets. J Compos Constr 1997;1(2):52 –62. w7x Karbhari VM, Seible F, Hegemier GA, Zhao L. Fiber Reinforced Composite Decks for Infrastructure Renewal – Results and Issues. Proceedings of International Composites Expo 97, Nashville, Tennessee, 1997. w8x Seible F, Karbhari VM, Burgueno ˜ R, Seaburg E. Modular Advanced Composite Bridge Systems for Short and Medium Span Bridges. Proceedings of Fifth International Conference on Short and Medium Span Bridges, Calgary, Canada, 1998. w9x Karbhari VM, Seible F, Burgueno ˜ R, Davol A, Zhao L. Damage Tolerance and Durability of an Advanced Composite Bridge
w10x
w11x
w12x
w13x
w14x
w15x
403
System. Proceedings of First International Conference on Durability of Composites for Construction, Sherbrooke, Canada, 1998;57–68. ˜ R, La Rovere H, Zhao L, Karbhari VM, Seible F, Burgueno, ¨ M, Godonou P. Experimental Investigation of ProBrostrom totype Transverse System for the Gilman Drive Advanced Technology Overcrossing. Structural Systems Research Project. SSRP 2001y04. Department of Structural Engineering, University of California, San Diego, La Jolla, May 2001. AASHTO. AASHTO LRFD Bridge Design Specifications. 2nd Edn. American Association of State Highway and Transportation Officials, Washington, D.C., 1998 (including 1999 and 2000 interims). UCSD. Design and Analysis of the I-5yGilman Advanced Technology Bridge. Department of Structural Engineering, University of California, San Diego, La Jolla, February, 2000. Hose YD, Zhao L, Oiler E, Karbhari VM, Seible F. Seismic Performance of the Concrete-Filled Carbon Shell Pylon Splice Connection for the I-5yGilman Advanced Technology Bridge. Structural Systems Research Project. SSRP-2001y20, University of California, San Diego, La Jolla, April 2002. UCSD. Design Criteria for the I-5yGilman Advanced Technology Bridge. University of California, San Diego, La Jolla, September 2001. AASHTO. AASHTO LRFD Bridge Design Specifications Modifications to Chapter 11 (abutments, walls, earth pressure). 2nd Edn. American Association of State Highway and Transportation Officials, Washington, D.C, 2001.
Use of FRP composites in civil structural applications Lelli Van Den Eindea,*, Lei Zhaob, Frieder Seiblec a
Department of Structural Engineering, University of California, 9500 Gilman Drive, San Diego, LaJolla, CA 92093-0085, USA b Department of Civil and Environmental Engineering, University of Central Florida, Orlando, FL 32816-2450, USA c Department of Structural Engineering, University of California, 9500 Gilman Drive, San Diego, LaJolla, CA 92093-0085, USA
Abstract Fiber reinforced polymer (FRP) composites or advanced composite materials are very attractive for use in civil engineering applications due to their high strength-to-weight and stiffness-to-weight ratios, corrosion resistance, light weight and potentially high durability. Their application is of most importance in the renewal of constructed facilities infrastructure such as buildings, bridges, pipelines, etc. Recently, their use has increased in the rehabilitation of concrete structures, mainly due to their tailorable performance characteristics, ease of application and low life cycle costs. These characteristics and the success of structural rehabilitation measures have led to the development of new lightweight structural concepts utilizing all FRP systems or new FRPyconcrete composite systems. This paper presents an overview of the research and development of applications of advanced composites to civil infrastructure renewal at the University of California, San Diego (UCSD). 䊚 2003 Elsevier Ltd. All rights reserved. Keywords: Advanced composites; Carbon shell system; Modular structural systems; Retrofit; Repair; Strengthening
1. Introduction The application of advanced composite materials in civil engineering has been evolving slowly, primarily due to economic reasons. Their key advantages, such as free form and tailorable design characteristics, strengthto-weight ratios that significantly exceed those of conventional civil engineering materials and a high degree of chemical inertness in most civil environments are lost in high material and manufacturing costs. Furthermore, the current practice of one-to-one component replacement of elements in conventional structural systems by advanced composite components has shown that it is difficult to justify the use of advanced composites in civil construction, not only economically, but also structurally. Several developments have changed this situation over the past few years: (1) techniques such as pultrusion, resin transfer molding, filament winding and semiautomated manufacturing of large components have led to advances in low cost FRP manufacturing; (2) reduced material demand in the high priced defense industry, expansion of a highly competitive market for these *Corresponding author. Tel.: q1-858-822-2188, fax: q1-858-8222260. E-mail address: [email protected] (L. Van Den Einde).
materials in the sporting goods industry and prospects for large volume use in the civil sector have led to new low cost materials manufacturing; and (3) designs of these new materials in conjunction with conventional structural materials rather than individual component replacement or complete advanced composite designs, have shown that technical efficiency can be achieved within competitive economical constraints. The emerging field of renewal engineering may best describe the role of FRP composites in civil engineering. The renewal of the structural inventory, which is depicted in Fig. 1, can be divided into (1) rehabilitation, including the applications towards repair, strengthening and retrofit of structures; and (2) new construction with all FRP solutions or new composite FRPyconcrete systems. The structural effectiveness of FRPs in the rehabilitation of existing structural systems has repeatedly been demonstrated with full or large-scale structural tests at the University of California, San Diego (UCSD). Carbon fabric overlays have been used to strengthen and retrofit reinforced and unreinforced masonry walls for seismic loads, as well as to restore and more than double the displacement capacity in the repair of a fullscale five-story reinforced masonry building tested to failure under simulated seismic loads. Carbon fiber overlays and strips have also been used to strengthen
0950-0618/03/$ - see front matter 䊚 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0950-0618Ž03.00040-0
390
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 1. Use of fiber reinforced polymer composites in civil engineering.
reinforced concrete slabs (with and without openings) and to strengthen large diameter prestressed concrete pipelines to restore loss of load (water pressure) carrying capacity due to corrosion of the prestressing wires. Bridge columns have been seismically retrofitted and repaired with fiberglass, carbon and hybrid composite jackets, which were shown to be as effective as conventional steel jackets. The success of structural rehabilitation measures with advanced composite materials has led to the development of new lightweight structural concepts utilizing FRP shells and tubes to form new structural systems. FRPyConcrete composite systems have been developed for use in new lightweight bridge systems due to their simplification in construction and excellent short- and long-term structural characteristics. An overview of past and ongoing research at UCSD on application projects of advanced composites in civil engineering is provided in this paper to show the extent of current developments and to explore the feasibility of future applications in civil engineering. A prognosis for the application of advanced composites in civil engineering construction is made by outlining critical technical and implementation issues to be resolved prior to a broad acceptance by the civil engineering design and construction community. 2. Repair of constructed facilities The effectiveness of the application of advanced composite wall overlays for seismic repair and retrofitting of structural wall systems can best be demonstrated by the example of a full-scale five-story reinforced masonry building (see Fig. 2), which was tested under pseudo-dynamic simulated seismic loads at UCSD w1x. Following the initial failure loading, the crushed wall toes in the building were reconstructed with polymer concrete and the building was repaired using structural carbon composite overlays on the first two stories of the structural walls (see Fig. 2b). Only one layer of the 12 K tow uni-directional woven AS4 carbon fabricy epoxy resin composite with a nominal thickness of 1.25 mm was applied except at the toe regions, where an additional layer of the same composite was used.
The load–deflection envelopes for the original and repaired five-story building are shown in Fig. 3. The repair scheme contributed significantly to the system ductility by almost doubling the inelastic deformation capacity in the critical push direction. Measured shear deformations in the overlaid wall panels were reduced to half of those observed in the original test. Detailed information of the repair and retrofit installation and seismic response data can be found in w1x. 2.1. Strengthening of walls and slabs 2.1.1. Structural wall overlays Strengthening of structural walls can be accomplished very economically with thin advanced composite wall overlays. To date, tests have focused on (1) reducing shear deformations in structural walls; (2) retrofitting shear walls to achieve ductile in-plane behavior; (3) repairing damaged walls to increase in-plane ductility; and (4) retrofitting out-of-plane unreinforced structural walls. Seven single-story full-scale structural wall panel tests were performed at UCSD on fully grouted hollow core concrete masonry walls w2x. The overall test setup for these wall tests is shown in Fig. 4a. 2.1.1.1. Shear strengthening. Very thin overlays (only one or two layers with thickness of 0.5–1.0 mm) can achieve significant seismic improvements especially for in-plane shear wall response. Carbon fibers in the composite overlay are oriented horizontally to cross diagonal or shear cracks, while allowing horizontal or flexural cracks to open. Forces to be transferred in the composite overlays are limited by the laminar shear or principal tensile strength of the existing structural wall material, since the polymer resin typically features significantly higher tensile capacities than the concrete or masonry substrate. To improve the shear capacity of structural walls with advanced composite overlays, the following design assumptions can be made: a conservative diagonal tension crack angle of 458 can be selected and the resulting shear capacity increase can be determined based on an allowable overlay stress level derived for a maximum
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
391
Fig. 2. Five-story full-scale building test.
horizontal wall strain of 0.004, above which, aggregate interlock is assumed lost. Alternatively, stiffness criteria can be employed in the wall overlay design, limiting
shear deformations to levels that can be expected in concrete walls with conventional horizontal reinforcement (determined based on conventional design
Fig. 3. Load–displacement envelope comparison for carbon overlay repair test.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
392
Fig. 4. In-plane strengthening of structural walls.
requirements). To determine the amount of horizontal overlay fabric Aoh, the equivalent required horizontal steel reinforcement Areq sh is scaled based on the ratio between the steel stiffness and overlay stiffness as shown in Eq. (1). Aohs
Esh req Ash Eoh
(1)
where Esh and Eoh represent the horizontal steel and composite overlay laminate moduli, respectively. Three full scale 2=2 m structural wall panels were constructed with 152 mm thick fully grouted and reinforced concrete block masonry using reinforcement that was 19 mm in diameter in the vertical direction and 9.5 mm in diameter in the horizontal direction. Both vertical and horizontal reinforcement had a spacing of 400 mm and a nominal yield stress of 414 MPa (see Fig. 4b). The walls were tested in the (1) as-built condition; (2) repaired condition following diagonal shear failure; and (3) retrofitted condition. The repair of the original wall panel consisted of epoxy injection of the diagonal shear
failure crack, reconstruction of the crushed compression toe with polymer concrete and horizontal carbon composite overlays with a layer of 490 gym2 unidirectional carbon fabric on both sides of the wall. The retrofit specimen was a new as-built wall with two layers of the same horizontal carbon composite overlays only on one side of the wall. Thus, the amount of horizontal carbon fabric was the same in both cases (see Fig. 4c). For the horizontal cyclic load test, the wall panels were first loaded with a 1.8 MPa axial or vertical load and subsequently restrained against rotation at the wall top. The failure mechanism for both the repaired and retrofitted walls were similar and consisted of separation of the face shell from the core at the compression toe regions, which resulted in the bulging out and delamination of a large region of the carbon overlay and crushing of the masonry units. Envelopes for the load– deflection hysteresis curves for the three tests, which are depicted in Fig. 4d, show that both the double sided repair and single sided retrofit with horizontal carbon overlays resulted in twice the deformation capacity of the as-built wall specimen. The improved performance
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
393
Fig. 5. Out-of-plane retrofit of unreinforced walls.
of the repaired wall specimen (over the other two tests) can be attributed to the enhancement of the wall toe that was rebuilt with polymer concrete, which has a higher compression strength, following the initial failure test. 2.1.1.2. Flexural strengthening. For out-of-plane wall strengthening, the key design aspects are (1) to achieve force transfer from the substrate to the overlay material in regions of high moment; and (2) to avoid the potential risk of buckling and delamination of the thin and stiff overlays on the compression side of the flexural member. Two out-of-plane flexural wall tests were also performed at UCSD on 2=2 m unreinforced fully grouted masonry wall panels, see Fig. 5a, with the same geometric and axial load conditions as the shear walls described in Fig. 4. To provide flexural capacity to the wall, a single vertical layer of carbon fabric overlay was developed at the base of the cantilever wall. End anchorage or fabric development was achieved with four different anchorage types. The two concrete anchorage solutions both incor-
porated bolted base beams and used different wraparound curvature for the carbon fabric (see Fig. 5b), while the two steel anchorage solutions consisted of using either a bolted steel angle or individual steel straps (‘Simpson ties’), as shown in Fig. 5c. The steel angles were placed on top of a carbon fiber overlay while the steel straps were sandwiched between a carbon base strip and a carbon fiber overlay. The cantilever load–deflection envelopes for the wall retrofit schemes with the four different connection details are depicted in Fig. 5d. In the first quadrant, the dashed line represents the ‘Concrete Curve’ case, while the dotted line represents the ‘Steel Angle’ case. In the third quadrant, the dashed line represents the ‘Concrete Fillet’ case and the dotted line represents the ‘Simpson Tie’ case. By far, the largest deformation capacity was obtained using individual steel straps on the tension side as anchorage, since the right-angle-bent steel straps (‘Simpson Ties’) straighten out with increasing cantilever loads. Lateral load capacities do not depend on the anchorage type as long as the full carbon fabric capacity
394
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 6. CFRP strengthening of large-scale slabs with openings.
can be developed. Test results show that the design of flexural carbon overlays is controlled by strain limits on the compression side, which is less than half their coupon tensile capacity. As a general rule of thumb, a limit of one-third the ultimate coupon tensile strain, ´ou, should be applied. 2.1.2. Slab strengthening The use of Carbon FRP (CFRP) composites in the form of strips or overlays, offers a cost-effective technique for the flexural strengthening of substandard reinforced concrete slabs. The efficiency of this application has been demonstrated through numerous experimental investigations and field applications, e.g. w3,4x. A recently completed experimental test program at UCSD investigated the effectiveness of the CFRP strip strengthening method on large scale slabs weakened by the introduction of a rectangular hole at the center of the slab w5x. A series of four tests on large-scale slabs with a rectangular cutout in the center retrofitted with CFRP strips was conducted to evaluate the use of this strengthening method for typical slab upgrades. Of the four test units that were considered for this project, two slabs were used as base-line references, one loaded in three
point bending and the other in four point bending. The remaining two specimens were reinforced with carbon strips in the two principal directions and were tested in the two respective loading configurations. The dimensions of the slabs considered were 6 m in length, 3.2 m in width and 0.18 m in thickness. A rectangular opening measuring 1 m in length and 1.6 m in width and centered with respect to the slab was later introduced. The dimensions of the cutout were chosen as half the width of the slab such that, the main reinforcement ratio was reduced by half and 1 m in the longitudinal direction to maintain a reasonable strength capacity. The test setup for the two flexural slabs tested in three-point bending is shown in Fig. 6a. A steel distribution beam over the cutout allowed introduction of load at mid-span. An overview of the test results for the three- and four-point bending tests is provided by the load–displacement curves in Fig. 6c and d, respectively. The figures demonstrate that the ultimate strength of the slab was increased to more than twice that of the weakened slab with cutout. Under both loading conditions, the strengthening measure was successful in recovering the initial strength of the previously tested slab without the cutout w5x, while also increasing the stiffness. In addition to a higher load carrying capacity, the CFRP strength-
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
395
Fig. 7. Carbon fiber jacket shear column retrofit test.
ening reduced the strain levels in the steel reinforcement and enhanced the development of a more evenly distributed crack pattern. Failure of the strengthened test units was due to delamination or peeling-off of the strips (see Fig. 6b). 3. Seismic column retrofit Recent earthquakes in California such as the Wittier 1987, Loma Prieta 1989 and Northridge 1994 have repeatedly shown the vulnerability of existing bridge columns built before the 1971 San Fernando Earthquake. Tests on 0.4 scale bridge columns at UCSD have shown that carbonyepoxy jacket retrofits can be just as effective as comparable steel jacket retrofits w6x. As an example, the dimensions and reinforcement layout of one shear column test unit are shown in Fig. 7a together with the
carbonyepoxy jacket design of variable thickness. A jacket thickness of only 0.4 mm was required over the shear critical center region of the column to prevent brittle shear failure. Experimental load–deflection hysteresis loops for the carbon fiber retrofitted shear column are shown in Fig. 7b. Completely stable hysteresis loops up to a displacement ductility level of mDs10.5 were achieved at which point the test was terminated due to test setup limitations rather than retrofit failure. Comparative load–deflection envelopes for (a) the unretrofitted or ‘as-built’ shear column; (b) a companion steel jacket retrofitted column with 5 mm jacket thickness; and (c) the carbonyepoxy retrofitted column are depicted in Fig. 7c. Clear improvement of deformation capacity can be seen for both the steel and carbon retrofitted cases over the ‘as-built’ case, which failed in brittle shear at a displacement ductility level of mDs2.0.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
396
Fig. 8. Modular FRP deck test.
The test results in Fig. 7b and c show that the column retrofitted by a steel jacket exhibited a slightly higher initial stiffness and a small increase in lateral load carrying capacity with increasing displacement levels due to the isotropic nature of the steel jacket, resulting in a more concentrated plastic hinge and more strain hardening at the column ends. Both stiffness and capacity increases are not necessarily desirable in bridge column retrofits, since typically higher seismic force levels are transmitted to adjacent structural elements. Thus, the carbon jacket with only horizontal or hoop directional strength and stiffness can better accommodate the requirements for minimal stiffness or strength increase than steel jackets. FRP jackets have also been shown to be effective for rectangular columns. For most column retrofit designs, the stiffness of the overlaying jacket material controls the design. In general terms, the required jacket thickness follows the relationship found in Eq. (2).
tjf
D C Ej
(2)
where D is the column dimension in the loading direction, Ej is the jacket or overlay modulus and C is a general coefficient w6x depending on the retrofit demand. While FRP jacket retrofitting clearly provides sufficient structural effectiveness based on short-term load tests, appropriate reduction factors for differences and uncertainties in the materials and lay-up systems, jacket curing and durability considerations need to be applied for actual retrofit application.
4. Structural replacement—new bridge systems 4.1. All FRP systems The advantages of FRP composites make them attractive for use in replacement decks or in new bridge systems as well. In addition to the potential lower lifecycle costs, FRP decks would be significantly lighter resulting in large savings in column and foundation costs and enabling higher live load levels through the replacement of heavier decks. FRP deck systems also have a high application potential in areas where longer unsupported deck spans are necessary or where lower weight would translate into lower seismic demands. Research at UCSD on full-scale FRP deck elements has indicated that appropriate performance levels can be achieved at significantly lower weights w7x. The challenge is to optimize the configuration and use of advanced composite material to match both the performance and cost of reinforced concrete decks. Initial tests on a number of different configurations show that FRP decks are capable of much higher failure loads with comparable stiffness at service demand. Fig. 8 shows the testing configuration and gives representative loaddeformation results of the testing program conducted at UCSD on modular E-glassyvinylester deck panels composed of pultruded cores with hand lay-up face sheets. Additional tests have been performed at UCSD on FRP decks with cores made of (1) endgrain balsa; (2) foam cubes made of triangular and trapezoidal prisms; (3) corrugated premanufactured glass sheets; (4) pultruded I and tubular shapes; and (5) aluminum honeycomb. Manufacturing methods include, but are not limited to pultrusion with hand lay-up top and bottom
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
397
Fig. 9. Concrete filled Carbon Shell System (CSS).
sheets, vacuum assisted resin transfer molding (VARTM) and complete hand lay-up w7x. The high material costs of the advanced composites virtually limit the material choice for replacement bridge decks to EGlass and even then cost competitiveness with conventional concrete or orthotropic steel bridge decks is difficult to achieve. 4.2. Innovative FRP-concrete composite systems The development of new structural concepts and systems, that combine the directional and tailorable characteristics of FRP composites in tension with the dominant characteristics of conventional structural materials such as concrete in compression, show great potential for advances in the design and construction of new civil engineering structures. 4.2.1. Carbon shell system (CSS) The first modular FRP-concrete composite structural system consisting of concrete filled carbon shells has been developed at UCSD w8x and is schematically shown in Fig. 9. The concrete filled Carbon Shell System
(CSS) concept creates new structural elements by using prefabricated filament-wound carbonyepoxy thin shells filled on-site with concrete. The shell serves the dual function of longitudinal and circumferential confinement reinforcement and stay-in-place formwork for the concrete core. The concrete provides compression force transfer, stabilizes the thin shell against buckling and provides anchorage of connection elements. The design of the carbon tubes consists of carbon fibers in the longitudinal tube direction ("108 from the member longitudinal axis due to manufacturing limitations) and hoop fibers (908) in the transverse direction. Circumferential ribs are provided on the inside of the carbon shell for full force transfer between the concrete infill and the shell. The concept is suitable for both columns and girders. For the development of bridge superstructure components, the concrete filled carbon shells are combined with a structural deck system, which may consist of either a conventional cast-in-place reinforced concrete (RC) slab or a modular FRP deck system as shown in Fig. 10. The connection between the deck and the carbon shell girders is accomplished by embedding shear
Fig. 10. Modular CSS beam and slab assemblages.
398
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 11. Beam-and-slab assembly test (‘3-girder’ system test).
connector dowels into the carbon shell system during grouting. In the slab, the dowels are either cast directly into the RC deck or anchored in polymer concrete filled sections of the cellular E-glass deck systems. The modular FRP E-glass deck consists of pultruded trapezoidal elements bonded together and overlaid with continuous face sheets under factory conditions. Typical composite decks weigh 1–1.5 kNym2 or 1y4 of a conventional reinforced concrete deck, resulting in significant weight savings for construction and mass reductions for seismic zones. 4.2.1.1. ‘3-girder’ system test. To validate the overall performance of the CSS system, a full-scale superstructure system consisting of three carbon shell girders and a FRP deck system was constructed and tested in the laboratory at UCSD w9x. The test specimen simulated one half of the width and the distance between the support and the inflection point for one span of the prototype Kings Stormwater Channel Bridge in California. Characterization of the system durability in terms of structural integrity and damage tolerance was investigated by subjecting the bridge superstructure first to stiffness characterization tests up to service load levels
and then to 2 million cycles of fatigue service loading. A photograph of the test setup is shown in Fig. 11a. The loading pattern consisted of four servo-controlled actuators applying simultaneously a load of 56 kN each at a frequency of 1 Hz. This load level duplicates the shear force demand on the prototype bridge at the girder–deck interface under full service loads. Test results show that the strength of the structure did not degrade during fatigue loading. This conclusion was further validated when the stiffness characterization tests were repeated after completion of the 2 million fatigue cycles without any noticeable change in structural response. The load–displacement response of the system up to failure at mid-span is shown in Fig. 11b. The system remained primarily linear up to failure, which occurred in the top face sheet of the deck panel. At the peak load (490 kN per actuator), which is equivalent to approximately 8.8 times the service demand level, the face sheet of the deck delaminated from the core and the core material buckled near its connection with the inclined web (see Fig. 11c,d). After failure, the specimen was unloaded and then cycled three times with a load of 445 kN force per actuator. The final permanent settlement at the comple-
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
399
Fig. 12. Hybrid tube bridge system.
tion of the test was less than 5 mm. No visual damage to other components, such as girders, end blocks and connections was observed w9x. 4.2.2. Hybrid tube system (HTS) Another modular FRP bridge system has been developed based on pultrusion and hand lay-up of a carbony glass hybrid material system w8x. The hybrid tube system (HTS) beam-and-slab bridge proposes the use of hollow E-glassycarbon hybrid beams that are connected along their tops with a polypropylene fiber reinforced concrete deck system as shown in Fig. 12. The girders consist of
Fig. 13. Hybrid tube bridge system characterization studies.
pultruded or hand laid-up E-glassyvinylester rectangular sections, which are reinforced along the bottom flange with unidirectional carbon fibers. An FRP panel is snaplocked to the pultruded girders providing flexural reinforcement to the concrete slab between girders as well as stay-in-place formwork for the slab. The end hooks of the panel are anchored by filling polymer concrete in the ‘dovetail’ pockets of the top portion of the transverse girder. Prefabricated carbonyvinylester snap-in stirrups are also snapped into the grooves to provide horizontal shear transfer between the concrete deck and the girders. A parametric study of a hybrid tube bridge system was performed by evaluating the maximum simply supported span length of the system under a set of allowable design criteria for service and factored loads. The bridge system under consideration consisted of seven transversely spaced HTS girders supporting a 13m wide road surface. The results of the parameter study as shown in Fig. 13, indicate that the deflection under live load governs the design, which is common for FRP structures. 4.2.2.1. Two-girder test. A two-girder assembly test was also conducted at UCSD to verify the potential application of the HTS concept. The test specimen was simply supported and the hollow rectangular girders were locally filled with concrete near the pin supports, FRP deck panels were installed between the two girders and a polypropylene fiber reinforced concrete slab was poured on top, as discussed in Section 4.2.2 w10x. The load cases for the test were designed to evaluate the response of the system under loads equivalent to the demands of a prototype bridge and to assess potential
400
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
Fig. 14. Two-girder hybrid tube system (HTS) test.
failure mechanisms and sequences in the system assembly. Fig. 14a shows a photograph of the test under a simulated wheel load condition. Results from one of the flexural loading cases are provided in Fig. 14b. The two-girder system showed significant strength reserves over the demand of the proposed prototype bridge. The system and its concrete slab behaved linearly under wheel and global flexural loads up to factored load levels. Even after the concrete slab was severely damaged at multiple locations (from punching shear tests) and showed signs of slippage (bilinear point in Fig. 14b at 60 mm), the system was still able to carry loads greater than the demand with minimal stiffness loss. The girder-to-deck connection displayed good integrity up to 150% of the factored shear demand. The deflection and strains in the deck between girders were compared with those obtained from simply supported
deck tests w10x. It was concluded that under an AASHTO HS 20 truck wheel load w11x, a significant strength reserve existed in the deck system. 4.3. Advanced technology bridge To demonstrate the application of the advanced composite bridge systems presented above, a 137 m long cable-stayed bridge supported by a 46 m high A-frame pylon was designed to be constructed at UCSD utilizing advanced composites. The bridge structure is designed for two 3.7 m vehicular lanes, two bicycle lanes, two pedestrian walkways and a utility service tunnel. Based on a preliminary study of different structural systems, the design presented in Fig. 15 was selected. The structure proposes the use of the carbon shell system (CSS) for the pylon legs and edge longitudinal girders,
Fig. 15. The I-5yGilman Advanced Technology Bridge.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
401
Fig. 16. Pylon splice connection test.
supported by a dual cable plane system. In the transverse direction, partially grouted carbon tube cross-beams are employed, which in turn support longitudinally spanning prefabricated E-glass or reinforced concrete deck panels w12x Prior to construction of the I-5yGilman Advanced Technology Bridge, prototype component and system
validation testing of all FRP members is required. The following section describes results from one of the fullscale tests that have already been performed. 4.3.1. Pylon connection test The A-frame pylon in the proposed I-5yGilman Advanced Technology Bridge is composed of 1.5 m
402
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403
diameter, 9.75 m long filament-wound cylindrical carbon composite tubes filled with concrete that are spliced together up the pylon height. As part of a Validation Test Program for the bridge, a Pylon Splice Connection Test was conducted w13x to validate the design and analysis methods used to develop the connection splice reinforcement details, to examine the ductile failure of the mild steel reinforced splice section, to assess the confinement effects of the carbon shell and to evaluate strain levels in the carbon composite shell in comparison with targets set forth in the Design Criteria w14x for the bridge. The design of the pylon leg for the bridge requires that failure will occur in the splice region by yielding of the mild steel reinforcement rather than brittle failure in a typical composite section. The final design details for the ‘typical’ splice connection in the Pylon Connection Test resulted in thirty 29 mm diameter longitudinal bars with 1.83 m splice length, confined by 13-mm diameter hoops that were spaced at 0.12 m. A photograph of the test setup is shown in Fig. 16a, while the test specimen at extreme drift is depicted in Fig. 16b. The axial load was applied using four external tendons, each composed of 27–15 mm diameter strands. The force vs. displacement hysteresis is provided in Fig. 16d w13x. Also included in Fig. 16d are the SERVICE I (395 kN), STRENGTH I (622 kN) and EXTREME I (1222 kN) demand levels which were based on AASHTO LRFD toad combinations w15x for the I-5yGilman Bridge. At the end of the test, 15 of the 30 longitudinal bars in the splice region had fractured (Fig. 16c), but the overall global capacity of the specimen only degraded approximately 20%. The Pylon Splice Connection Test was successful in demonstrating that the design satisfied the performance objectives. The compression and tension strain levels in the composite shell were well below the limit states set forth in the Design Criteria w14x. The development length provided in the splice region was sufficient in transferring the forces to the composite shell. The amount of transverse reinforcement provided by the composite shell and steel hoops in the splice region was able to resist pull-out failure of the lap splice longitudinal bars and sufficiently confine the plastic hinge region for adequate ductility capacity. Preliminary assessments of the strain levels as well as acoustic observations during testing indicate that slip of the longitudinal splice reinforcement was avoided. 5. Summary Worldwide research and application of FRP composites in civil engineering construction demonstrate the need for new construction materials in support of civil infrastructure renewal. The demonstrated advantages of FRP composites have shown that they will play a significant role in future civil engineering projects. The structural effectiveness of FRPs in the repair of con-
structed facilities, strengthening of structural walls, slabs and retrofitting of concrete columns with shear, flexure and lap-splice problems has been validated with large or full scale laboratory tests. However, the use of fiber reinforced composites for the rehabilitation of structures requires that appropriate design philosophies, guidelines and detailing be established and that the design is conducted using a methodology that ensures appropriate use of the material. With suitable design criteria for FRPs in structural rehabilitation, significant advantages can be derived from the lightweight properties of these new materials and their ease of handling and installation. Innovative FRPyconcrete bridge systems, such as the concrete filled carbon shell system (CSS) and the hybrid tube system (HTS) effectively use FRPs for new construction by combining them with conventional materials such as concrete and steel. System characterization and design studies are providing the basis for design approaches for FRP bridge systems in terms of deformation and strain limit states. It is expected that modular FRP bridge systems of this type will lead to faster construction and less traffic interruption due to their light weight, as well as lower life-cycle costs due to reduced maintenance. The extent of these FRP applications in support of civil engineering renewal will depend on (1) the resolution of outstanding issues such as reparability, fire, durability and environmental concerns; (2) the extent to which automation in the manufacturing process can reduce cost; (3) the availability of validated codes, standards and guidelines that can be used as design references and tools by the civil engineering community; and (4) the degree of quality control and quality assurance that can be developed and provided during the manufacturingyinstallation phase utilizing unskilled general construction labor. Acknowledgments The authors would like to acknowledge the United States Federal Highway Administration (FHWA), the California Department of Transportation (Caltrans), the Defense Advanced Research Projects Agency (DARPA) and the State of California for funding the research reported in this paper. Furthermore, the authors acknowledge the technical staff of the Powell Structural Research Laboratories at UCSD for their support in the construction and testing of these research projects. References w1x Weeks J, Seible F, Hegemier GA, Priestley MJN. The USTCCMAR Full-Scale Five-Story Masonry Research Building Test: Part V – Repair and Retest. Structural Systems Research Project. SSRP-94y05. University of California, San Diego, La Jolla, January, 1994.
L. Van Den Einde et al. / Construction and Building Materials 17 (2003) 389–403 w2x Laursen PT, Seible F, Hegemier GA. Seismic Retrofit and Repair of Masonry Walls with Carbon Overlays. Structural Systems Research Project. SSRP-95y01. University of California, San Diego, La Jolla, January, 1995. w3x Meier U, Deuring M, Meier H, Schwelger G. Strengthening of Structures with CFRP Laminates: Research and Applications in Switzerland. Proceedings of Advanced Composite Materials in Bridges and Structures: First International Conference, Sherbrooke, Canada, 1992;243–251. w4x Nanni A, Focacci F, Cobb C. Proposed Procedure for the Design of RC Flexural Members Strengthened with FRP Sheets. Proceedings of Second International Conference on Composites in Infrastructure, Tucson, Arizona, 1998;187–201. w5x Vasquez A. The use of carbon fiber reinforced polymer strips for the external strengthening of slabs. San Diego, La Jolla: M.S. University of California, 1999. w6x Seible F, Priestley MJN, Hegemier GA, Innamorato D. Seismic retrofit of RC columns with continuous carbon fiber jackets. J Compos Constr 1997;1(2):52 –62. w7x Karbhari VM, Seible F, Hegemier GA, Zhao L. Fiber Reinforced Composite Decks for Infrastructure Renewal – Results and Issues. Proceedings of International Composites Expo 97, Nashville, Tennessee, 1997. w8x Seible F, Karbhari VM, Burgueno ˜ R, Seaburg E. Modular Advanced Composite Bridge Systems for Short and Medium Span Bridges. Proceedings of Fifth International Conference on Short and Medium Span Bridges, Calgary, Canada, 1998. w9x Karbhari VM, Seible F, Burgueno ˜ R, Davol A, Zhao L. Damage Tolerance and Durability of an Advanced Composite Bridge
w10x
w11x
w12x
w13x
w14x
w15x
403
System. Proceedings of First International Conference on Durability of Composites for Construction, Sherbrooke, Canada, 1998;57–68. ˜ R, La Rovere H, Zhao L, Karbhari VM, Seible F, Burgueno, ¨ M, Godonou P. Experimental Investigation of ProBrostrom totype Transverse System for the Gilman Drive Advanced Technology Overcrossing. Structural Systems Research Project. SSRP 2001y04. Department of Structural Engineering, University of California, San Diego, La Jolla, May 2001. AASHTO. AASHTO LRFD Bridge Design Specifications. 2nd Edn. American Association of State Highway and Transportation Officials, Washington, D.C., 1998 (including 1999 and 2000 interims). UCSD. Design and Analysis of the I-5yGilman Advanced Technology Bridge. Department of Structural Engineering, University of California, San Diego, La Jolla, February, 2000. Hose YD, Zhao L, Oiler E, Karbhari VM, Seible F. Seismic Performance of the Concrete-Filled Carbon Shell Pylon Splice Connection for the I-5yGilman Advanced Technology Bridge. Structural Systems Research Project. SSRP-2001y20, University of California, San Diego, La Jolla, April 2002. UCSD. Design Criteria for the I-5yGilman Advanced Technology Bridge. University of California, San Diego, La Jolla, September 2001. AASHTO. AASHTO LRFD Bridge Design Specifications Modifications to Chapter 11 (abutments, walls, earth pressure). 2nd Edn. American Association of State Highway and Transportation Officials, Washington, D.C, 2001.