Read view_file.cfm?file=JL-02-JULY-AUGUST-6.pdf text version

Behavior of Welded Plate Connections in Precast Concrete Panels Under Simulated Seismic Loads

Christian L. Hofheins, P.E.

Engineer JM Williams and Associates Salt Lake City, Utah

Lawrence D. Reaveley, Ph.D., P.E.

Professor Department of Civil & Environmental Engineering University of Utah Salt Lake City, Utah

Tests were performed on precast wall panels with typical loose-plate connectors located in the vertical joint between panels. The tests were performed to investigate the performance of the connectors under simulated seismic loads. Inplane lateral cyclic loads were applied to the wall panels, which applied tension-shear and compression-shear forces to the loose-plate connectors. The paper describes the experimental program and results for the welded plate connections in ten precast concrete wall panel assemblies. Design assumptions and simplified design models are also examined. The research shows that the connection possesses little ductile capacity and, therefore, is not suitable for use in high seismic regions (Zones 3 and 4). However, based on the observed failure modes, minor modifications to the connection are suggested that will increase the ductility of the connection.

Chris P. Pantelides, Ph.D., P.E.

Professor Department of Civil & Environmental Engineering University of Utah Salt Lake City, Utah

T

his paper addresses the behavior of a specific looseplate welded connector under applied cyclic loading. This type of connection is widely used in the United States. Due to the limited number of tests performed, no specific design parameters were considered in this study. The objectives of this investigation were to: (a) Quantify the performance of the connection in terms of force-deflection and ductility. (b) Check the validity of design values that are currently used for loose-plate welded connections in hollow-core precast concrete wall panel construction.

PCI JOURNAL

22

Fig. 1. Details of hollow-core wall panel. Note: 1 ft = 0.3408 m.

(c) Model the connection and provide preliminary recommendations based on observed failure modes. Precast concrete has largely been used in parts of the world where seismic issues play a small role in design. As a result, many common precast concrete connections are generally not designed to provide the desired ductility in seismic resistant structures. Presently, there is not an adequate set of seismic code requirements for the design of loose-plate connections in precast wall panels. Most looseplate connections currently specified by engineers are designed with static models that are not supported by test data. The design of precast connections for high seismic areas must address the need for design strength, displacement ductility, or both. One strategy is to design a ductile connection that is weaker than the precast concrete wall panels. This enables the connection to be at a location of ductile inelastic deformation and the precast wall panels to remain elastic under seismic response. As a result, overall costs decrease because the precast concrete wall panels do not need to be designed for ductility. Ductile connections allow lateral forces to be redistributed to all connectors. Another attractive feature of this system is that some ductile connections can be replaced after a seismic event, resulting in considerable savings in repair costs. A loose-plate connection typically comprises a steel plate welded to steel embeds cast into the concrete. The majority of loose-plate connections used in current practice have not been subjected to thorough testing. Consequently, there is little experimental

July-August 2002

validation upon which design procedures can be based. Most precast connectors were developed through field experience by individual precast manufacturers. These connectors are not supported by sufficient test data to determine their strength and deformation capacity. Standard test methods may be required in the future, because design codes will likely define design criteria in terms of performance objectives. A performance objective is the combination of a specific seismic hazard and a desired performance level. In this scenario, all components of a structure will be required to undergo rigorous testing to determine its performance level.

LITERATURE REVIEW

During the last 40 years, several studies have been carried out on a variety of wet and dry precast wall panel connections. A wet connection is made by cast-in-place concrete between the precast concrete panels; a dry connection consists of steel embedded plates, angles, or other steel elements that are welded together by a steel plate. The continuity of precast, prestressed double tee floors was investigated in a series of tests.1 Intermediate grade deformed bars were placed across the supports, and concrete was placed in the space between adjacent ends of the double tees to form transverse diaphragms. The primary objective was to investigate the structural soundness of the continuity connection, which was found to be adequate. Additional testing was performed to determine the flexural resistance of cast-in-place insulated walls. Three types of metal shear connectors be-

tween the concrete shells were included: a truss, a ladder, and an expanded metal shear connector. The truss and ladder shear connectors were found to be satisfactory.2 A variety of wet joints were studied to determine their ultimate shear strength.3 The research proved that wet joints used for vertical joints in panel structures effectively resist high shear forces. Although the joint installation is labor intensive, the joint can be very ductile if properly designed. Originally, dry joints were mostly composed of headed studs welded to the back of a steel plate. In one such headed stud connection,4 it was found that shear loads are transmitted through the embedded plate to the surrounding concrete by three distinct mechanisms: (a) Friction between the embedded plate and concrete. (b) Bearing of the end of the embedded plate on concrete. (c) Interaction between studs and concrete. These headed stud connections provide good shear resistance, but have a low ductile capacity. The PCI-sponsored Precast Seismic Structural Systems (PRESSS) research program has taken the lead on research and design recommendations for precast concrete structures in areas of high seismicity. Among other topics, the PRESSS program has performed research on a variety of welded connections for precast wall systems. The initial goal of the research was to develop ways of classifying and evaluating connection details.5 The National Institute for Standards and Technology (NIST) investigated the seismic performance of horizontal and vertical joint connections in

23

Fig. 2. Embedded angle assembly for welded connection tested.

precast walls.6 The connections were designed to be ductile, and to be the major location of inelastic response of the structure. Vertical joint connections included different designs of welded loose-plate and bolted ductile connections. The connections took advantage of the interaction between the embed and concrete by incorporating flexural yield, tension/compression yield, shear yield and friction sliding concepts. The behavior of a six-story precast concrete office building under moderate seismicity was investigated.7 It was concluded that uneven shear distribution in a precast system causes a high ductility demand in the panelto-panel joint connections. The uneven distribution drives the connection elements into the inelastic range. Therefore, connection details that can be easily replaced should be used in precast concrete structures. As part of the PRESSS five-story precast concrete building test, a structural wall system consisting of precast concrete panels was tested under simulated seismic loading.8 The precast concrete panels were connected to each other and the foundation by unbonded vertical post-tensioning, using threaded bars. A horizontal connection across the vertical joint was provided by stainless-steel energy-dissipating U-shaped flexure plates, welded to embed plates in both adjacent wall panels. In addition to providing energy

24

dissipation, these plates provided additional resistance by shear coupling between the structural walls. The structural response of the building under simulated seismic loads was extremely satisfactory. The ability of precast double tee floor diaphragm and wall systems to perform adequately under in-plane seismic forces has been studied in terms of: (a) The behavior of connections between double tees. (b) The analytical modeling of connectors, diaphragm, and wall systems. (c) The development of design guidelines for double tee diaphragms and wall systems.9 It was found that the interaction between shear and tension forces in a flange connection between double tees could be significant. The connector's ductility should allow the diaphragm to redistribute the force among individual connectors; this ensures that all connectors reach their full strength.9 In an experimental study of 3/8 in. (9.52 mm) stud-welded deformed bar anchors subject to tensile loads, it was found that a number of specimens fractured at the weld. Based on the test results, quality control procedures and revised settings were recommended for stud welding of deformed bar anchors.10 The strength and ductility of several tilt-up concrete wall panel connections were investigated in a se-

ries of monotonic and cyclic tests.11 Most of the connectors tested did not show sufficient ductility to be used in areas of high seismicity. Even when a connection possessed some ductility, extensive damage to the surrounding concrete was observed. Presently, there is no adequate set of seismic code requirements for the design of loose-plate connections in hollow-core precast wall panels. Many of the loose-plate connections currently used in construction are proportioned using design models that are seldom backed up with test data. The truss analogy, currently being used to describe the performance of the connection under consideration, leads to a conservative design. This paper addresses the behavior of a specific loose-plate welded connector for hollow-core precast wall panels under cyclic loading; this type of connection is widely used in high seismic regions of the United States. The primary objective of this research was to quantify the performance of the connections between precast concrete panels using loose-plate connectors and to assess the feasibility for their use in regions of high seismicity. Due to the limited number of tests performed, no specific design parameters have been considered in the study. The assemblies had variations which commonly occur in practice. These included the width of the welded plate, the length of the weld, the vertical unevenness of the embedded angles between adjacent panels, and the misalignment of the three wall panels in the out-of-plane direction. This paper presents the experimental results, analytical models of the connections, and the details of a proposed new welded connection.

EXPERIMENTAL PROGRAM

Tests were performed by applying a quasi-static cyclic load to three precast hollow-core wall panels connected together with two loose-plate connectors at each vertical joint. Ten wall panel assemblies were tested, all using the same loose-plate welded connection. Description of Precast Wall Panel Assemblies

PCI JOURNAL

Typically, hollow-core precast panels are 8 ft wide, 12 to 24 ft high (2.44 x 3.66 to 7.32 m) and have six hollow cores as shown in Fig. 1. The overall thickness of the panels is 8 in. (203 mm). Panels 12 ft (3.66 m) high and 4 ft (1.22 m) wide were used for testing due to space constraints in the load frame. Panels 4 ft (1.22 m) wide were fabricated by cutting an 8 ft (2.44 m) panel in half. The two center hollow cores of the 8 ft (2.44 m) panels were filled with concrete. These solid cores were required to form a pin connection at the two outside panels at the supports of the wall assembly. The average 28-day compressive strength of the concrete wall panels was found to be 7150 psi (49 MPa) with a standard deviation of 190 psi (1.3 MPa). Description of Welded Connections Two welded connections were located between panel pairs in vertical joints. Each welded connection comprises two embedded angle assemblies and a loose plate. Each embedded angle assembly consists of a 11/2 x 2 x 1/4 in. (38 x 50.8 x 6.4 mm) x 6 in. (152 mm) long angle, with three 3/8 in. (9.5 mm) diameter weldable steel deformed anchor bars. The bars are 12 in. (305 mm) long, and are stud welded to the back of the angle as shown in Fig. 2. Fig. 3 shows the details of the embedded angle assemblies. Each wall panel assembly consists of three hollow-core wall panels joined together with four welded connections. Two welded connections are placed 3 ft (914 mm) from the top and bottom of the wall panels in each vertical joint found in between the wall panels, as shown in Fig. 4. The width of the loose plate varied in some wall panel assemblies. Eight assemblies used 3 in. (76 mm) wide plates, and two assemblies used 2 in. (51 mm) wide plates. Test results showed that the plate width had no effect on the maximum force or displacement sustained by the wall assemblies. The loose plate was 1/ 4 to 3/ 8 in. (6.4 to 9.5 mm) thick A36 steel, and it was welded to the embedded angle

July-August 2002

Note: 1 in. = 25.4 mm.

Fig. 3. Details of embedded angle assembly.

assembly with two 3/16 in. (4.8 mm) fillet welds that ran along the 5 in. (127 mm) vertical edge of the plate as shown in Fig. 5. All welds were performed by certified welders with an E70 electrode and a 7018 rod. Test Setup A total of ten wall panel assemblies were tested in a load frame at the Structures Laboratory at the University of Utah. A steel belt enclosed the wall panel assembly and was connected to a hydraulic actuator with a force link. The panels were welded together in the vertical position after being placed in the load frame. The entire wall assembly was pushed or pulled by a 150 kip (667 kN) hydraulic actuator through the force link and the steel belt. The steel belt transferred the force from the hydraulic actuator to the wall panel assembly without restraining the panels. The panels were supported by two pin connections placed at the two bottom corners of the wall panel assembly as shown in Fig. 4. The pin used in this connection was a 2 in. (51 mm) diameter steel rod. The pin supports

supported the wall assembly 1.5 in. (38 mm) above the bottom of the test frame, making the pins the only support for the wall assembly. This allowed the walls to rotate at the pins and transfer the applied cyclical force between the panels in a symmetrical manner. A 1.5 in. (38 mm) thick steel plate was placed under each corner of the center panel as shown in Fig. 4. These plates raised the center panel up to the same height as the outside panels. This aligned the embedded angles to facilitate the placement of the welded plate. A more detailed description of the loading system and the wall assembly supports can be found in other publications from the University of Utah.12,13 Test Procedure and Instrumentation A force was applied to the top left corner of the wall assembly with a hydraulic actuator in a quasi-static manner. The test was carried out in a force-controlled mode at a rate of approximately 1 kip (4.5 kN) per second. Loading steps began at 10 kips (44.5

25

Fig. 4. Setup and instrumentation of typical wall assembly. Note: 1 in. = 25.4 mm.

kN) and increased by 5 kips (22.2 kN) until the welded connections failed. Each loading step consisted of three cyclic load increments to simulate the effects of an earthquake. Strain gauges

were placed on welded plates to form a three-element rectangular rosette. Displacement transducers were used in all of the tests to measure the displacements at various locations of

the wall panel assembly (see Fig. 4).

EXPERIMENTAL RESULTS

The tests revealed the following characteristics for the connection studied in this research: (a) The connection can resist relatively high shear loads. (b) The connection possesses little ductile capacity. (c) The connection should be designed as elastic due to insufficient ductility. Failure Mechanism Cracking around the connections began near the 20 kip (89 kN) load cycle. Cracking was initiated by the embedded angle pushing into the surface of the concrete. As soon as the concrete crumbled away from around the connection (see Fig. 6), the deformed anchor bars on the back of the embedded angle assemblies quickly tore away from their welds. Figs. 6(a)

PCI JOURNAL

Fig. 5. Details of welded loose-plate connection. Note: 1 in. = 25.4 mm.

26

(a)

(b)

Fig. 6. Welded connections for Assembly 4 at failure: (a) top right connection, and (b) bottom right connection.

and 6(b) illustrate the typical failed connections. The following is a description of the typical mode of failure for this connection: (a) The concrete around the embedded connections begins to crack. (b) The bearing capacity of the deformed anchor bars and embedded angle is severely decreased. (c) The deformed anchor bars quickly tear free from the embedded angles as soon as the concrete crumbles around the embedded angle assemblies. (d) The load carrying capacity of the connection is lost. The welds connecting the looseplate to the embedded angle assemblies for nine of the ten wall assemblies were not damaged. A weld in one wall panel assembly failed due to poor penetration of the weld onto the connecting plate. In general, the weld did not contribute to the failure of the connection. Vertical displacement transducers DT2 and DT3 (see Fig. 4) recorded very small relative movement between panels, until the connections failed. Therefore, the wall assembly moved as a relatively rigid body until the first connection failed. Force-Displacement Relationship of Wall Panel Assemblies

July-August 2002

The hysteretic behavior of Assembly 8 is typical of all wall assemblies and is shown in Fig. 7. The shape of the hysteresis loops demonstrates that they were stable and did not degrade until sudden failure. The assembly allowed a displacement drift of only 0.5 percent, and did not demonstrate any appreciable ductile behavior. The hysteresis envelope for every wall assembly was approximated by a general component behavior curve as described in FEMA 273.14 The general component behavior curve for the

ten assemblies tested is shown in Fig. 8. The general component behavior curve is able to define the hysteresis curves into important design criteria. As defined by FEMA 273, QCE is the expected strength of the welded connection of the wall section, and QCL is the lower-bound estimate of the strength. Table 1 contains a summary of the test data that was used to create the general component behavior curve of every wall assembly. The mean elastic force, QCL, equals 28.4 kips (126.3 kN), and the mean

Fig. 7. Hysteresis curve for Assembly 8. Note: 1 kip = 4.448 kN; 1 in. = 25.4 mm.

27

Table 1. Summary of test results for wall assemblies.

WallElasticforce,QCLElasticdisplacementUltimateforce,QCEUltimatedisplacement a ssembly (kips) (kN) (in.) (mm) (kips) (kN) (in.) (mm) 1 26.3 117.0 0.44 11.2 28.1 125.0 0.53 13.5 2 24.8 110.3 0.52 13.2 28.8 128.1 0.71 18.0 3 27.1 120.5 0.51 12.9 30.2 134.3 0.62 15.7 4 31.0 137.9 0.63 16.0 35.0 155.7 0.74 18.8 5 32.3 143.7 0.57 14.5 35.0 155.7 0.79 20.1 6 30.2 134.3 0.54 13.7 33.1 147.2 0.71 18.0 7 30.5 135.7 0.55 14.0 34.5 153.5 0.62 15.7 8 23.2 103.2 0.57 14.8 28.2 125.4 0.70 17.8 9 29.9 133.0 0.69 17.5 33.2 147.7 0.80 20.3 10 28.3 125.9 0.40 10.2 30.7 136.6 0.56 14.2

ultimate force, Q CE (mean value of peak on all hysteresis), equals 31.7 kips (141.0 kN). The mean elastic displacement is 0.54 in. (13.7 mm) and the mean ultimate displacement is 0.68 in. (17.3 mm). Thus, the range for the inelastic displacement was only 0.14 in. (3.6 mm). According to FEMA 273, the wall panel assembly would be defined as a force-controlled action due to the small plastic range. Strain gauges oriented in a threeelement rosette pattern were applied to several welded plates on the wall panel assemblies as shown in Fig. 9. This rosette pattern was chosen so that the principal stresses and their directions could be determined. There was insufficient instrumentation to determine the force in each plate directly from the strain gauges. Fig. 10 shows the principal strains recorded by the three-element rosette on the plate of the bottom right connection of Assembly 2. Although the plate yielded in the last loading cycle, the connection failed immediately thereafter. As a result, the ductility of the connection was not significantly increased by the yielded plate.

Fig. 8. General component behavior of ten wall assemblies. Note: 1 kip = 4.448 kN; 1 in. = 25.4 mm.

ANALYTICAL RESULTS

Fig. 9. Threeelement strain gauge rosette applied on loose-plate connector.

28 PCI JOURNAL

A structural analysis of the wall assembly was performed using the structural analysis program SAP 2000.15 The purpose of the analysis was to find the forces across each welded connection of the wall panel assembly, and compare them to the commonly used design methodologies. The precast concrete wall panels were modeled as rigid frame elements with a diaphragm constraint on each wall panel (as shown in Fig. 11). The wall panel connections were modeled as rigid pins, which is a reasonable assumption given their brittle mode of failure. The nodes located at the supports of the wall panel assembly were assigned pin restraints. The shim supports under the center panel (see Fig. 4) were not considered in the model. Vertical displacement transducers revealed that the center panel rose vertically, whether the wall assembly was being pushed or pulled. These displacements were a result of vertical movement occurring at the pin supports, and the rigid body motion of the wall panel assembly. The holes in the panels for the pin

Fig. 10. Principal strains at bottom right plate connection of Assembly 2.

supports were oversized for ease of erection in the load frame. The oversized holes allowed the entire assembly to rise and move as a rigid body. As a result, the bottom corners of the middle panel never touched the shims

during loading cycles. Consequently, the shim supports did not restrain the panel assembly, and were not included in the model. The weight of each wall panel was applied as a point load at four differ-

Fig. 11. Structural analysis model of wall panel assembly. Note: 1 kip = 4.448 kN.

July-August 2002 29

Fig. 12. Results of structural analysis for maximum lateral load applied to the wall assembly. Note: 1 kip = 4.448 kN.

ent nodes on each wall panel (see Fig. 11). The average maximum force at failure, 31.7 kips (141.0 kN), was applied as the lateral load at the top left corner of the wall panel assembly to find the capacity of each welded connection. The structural analysis results are shown in Fig. 12. For the above conditions, the shear force at failure of the welded con-

nections was 15.0 kips (66.7 kN) on the two left connectors, and 16.6 kips (73.8 kN) on the two right connectors. This is significant because the capacity of this connection typically used in design is equal to 8 kips (35.6 kN). Structures built with these welded connections were safely designed with an approximate factor of safety of 1.9. Using this design value, the connection

will safely stay in the elastic range. Force-Displacement Relationship of Welded Connection The force-displacement relationship of each welded connection was found by plotting the relative vertical displacement of two adjoining wall panels versus the shear force across the welded connection. The relative displacement of two adjoining wall panels in the vertical direction was found by subtracting data retrieved from displacement transducers DT2 and DT3 (see Fig. 4). The shear force across each connection was found as follows: the force applied by the hydraulic actuator on the wall assembly was multiplied by the ratio of the average maximum force at failure of the welded connection, or 16.6 kips (73.8 kN), to the average maximum force at failure of the wall panel assembly, or 31.7 kips (141.0 kN). This assumption is reasonable because the connections behave in a linear elastic manner. The hysteresis curve for the connectors of eight wall panel assemblies, was approximated by a general component behavior curve as described in the "Guidelines for the Seismic Rehabilitation of Buildings," FEMA 273.14

PCI JOURNAL

Fig. 13. General component behavior of the welded connectors of eight wall assemblies. Note: 1 kip = 4.448 kN; 1 in. = 25.4 mm.

30

Fig. 13 shows the general component behavior curve for the connectors of eight wall assemblies. The average elastic force on the connectors was 14.7 kips (65.4 kN), and the average ultimate force was 17.1 kips (76.1 kN). The force-displacement relationship is linear until the connection fails in a brittle manner. This connection should be designed to remain elastic due to its brittle mode of failure and limited ductility. Analytical Model of Welded Connection The probable resisting mechanisms of the connector under consideration are bearing and tension actions in the deformed anchor bars, as well as bearing of the angle section. Many designers currently model this welded connection with the truss analogy as described in the PCI Design Handbook.16 Fig. 14 is an illustration of the truss analogy. The following equations are used to describe this model: CU = TU = As fy VRU = (CU + TU )cos (1) (2)

Fig. 14. Truss analogy model for welded connection.

where CU = compression force TU = tensile force = capacity reduction factor = 0.9 = angle of deformed anchor bar = 45 degrees As = area of 3/8 in. (9.5 mm) di-

ameter deformed anchor bar = 0.13 sq in. (71 mm2) fy = yield strength of mild steel reinforcement [= 60 ksi (420 MPa)] VRU = vertical shear force resisted by connection The equations from the truss analogy yield a vertical shear resistance of 8.4 kips (37.4 kN) for each connection. The analysis indicates that the average capacity of this connection is between 15.0 and 16.6 kips (66.7 to 73.8 kN). The truss analogy is a conservative design methodology when applied to this connection. The following is a list of some of the differences between the truss analogy and the connection under consid-

eration: a. The angle for this connection equals zero, not 45 degrees (see Fig. 2). b. The deformed anchor bars are bent 90 degrees into the back of the angle (see Figs. 2 and 3). The bars will not be able to develop the full tensile capacity as described in the truss analogy. The deformed anchor bars act more as 3/8 in. (9.5 mm) studs with ineffective tails rather than bars in tension. c. The truss analogy does not account for the bearing of the angle assembly into the concrete. Angle bearing is one of the main force resisting mechanisms of the connection. Fig. 15 illustrates that the deformed

Fig. 15. Statics of deformed anchor bar at failure for current connection. Note: 1kip = 4.448 kN, 1 in. = 25.4 mm; 1 k-in. = 133 N-m.

July-August 2002 3

anchor bar cannot fully develop in tension due to the eccentric load from the bend in the bar. Assuming the force taken by each vertical deformed anchor bar is 8.3 kips (36.9 kN) (half of the total vertical shear force taken by the connection), the maximum shear and moment taken by each vertical deformed anchor bar is 8.3 kips and 10.4 kip-in. (36.9 kN and 1.17 kN-m), respectively. The eccentric load causes the deformed anchor bars to quickly tear free from their welds as soon as the concrete crushes around the connection.

PROPOSED NEW WELDED CONNECTION

The most effective way to improve this connection is to provide a larger surface area for concrete bearing and to minimize eccentric loads from the deformed anchor bars. Fig. 16 is a drawing of a proposed new embedded angle assembly. The angle is replaced by a 6 in. (152 mm) long ST2x3.85 to create a greater bearing area in the concrete. One continuous deformed anchor bar replaces the two vertical deformed anchor bars of the previous connection. The vertical deformed anchor bar is attached to the back of the embedded angle assembly with a 4 in. (102 mm) long, 3/16 in. (4.8 mm) fillet weld. The vertical deformed anchor bar is bent at 5 degrees to minimize eccentric loads and to ensure adequate concrete cover. The strength of this fillet weld can be described by Eq. (3), and the strength of the base metal can be described by Eq. (4), as:17 Rn = 0.75te(0.6Fexx) Rn = 0.75t(0.6Fu) (3) (4)

Fig. 16. Details of proposed new embedded angle assembly. Note: 1 in. = 25.4 mm.

Eq. (3) yields the strength of the fillet weld as 4.2 kips per in. (0.74 kN/ mm), and Eq. (4) yields the strength of the base material as 8.4 kips per in. (1.47 kN/mm). A 4 in. (102 mm) long weld gives a strength of 16.8 kips (74.7 kN), which is significantly higher than the allowable shear resistance of the welds in the tested connection. In addition, the concrete will not easily break away from the connection due to the increased bearing area with the web of the structural tee embedded into the wall.

DISCUSSION OF TEST RESULTS

Engineers prefer the panel connections, not the panels themselves, to be the weak link in the system. This investigation has shown that the connections are in fact the weakest link. Although the loose-plate connection used in this research effectively transferred the applied shear forces, the connection failed in a brittle manner. The small displacement ductility exhibited by the welded connections is lost as soon as the deformed anchor bars on the back of the embedded angle fracture from their welds. Fail-

where Rn = strength of fillet weld or base material Fexx = strength of electrode = 70 ksi (483 MPa) Fu = tensile strength of base material = 60 ksi (420 MPa) te = 0.707a a = weld size = 3/16 in. (4.8 mm) t = thickness of base material = 5 /16 in. (7.9 mm)

32

ure occurs before shear yielding can take place in the welded plate. These tests reveal that hollow-core precast concrete panels can be used in seismic regions provided that the connections can be improved. To this end, a new welded connection is proposed; ductility may be restored to the system by increasing the surface area for concrete bearing and by reducing the eccentric load in the deformed anchor bars. If the connection is a location of ductile inelastic deformation, the precast concrete panels will remain elastic under seismic response. Damage to the overall structure will be reduced and repair of the structure will be less costly. Ductility in shear will allow the force to redistribute among individual connectors. Ductility will enable all connectors to reach their full strength, thereby increasing the overall force resisting capability of the structure. For existing connections of the type tested in this investigation, a seismic retrofit option has been studied using a carbon fiber composite connection, which will be published shortly.

CONCLUSIONS

Simulated seismic load tests of

PCI JOURNAL

loose-plate vertical connections between precast concrete wall panels were performed. Based on the results of this investigation, the following conclusions can be drawn: 1. The loose-plate connection commonly used in precast construction can resist relatively high shear forces. 2. The connection fails in a brittle manner when the deformed anchor bars tear free from the embedded angles, which occurs as soon as the concrete crumbles around the embedded angle assemblies; as a consequence, the connection possesses little ductile capacity. 3. The connection should be de-

signed to remain elastic; in its current form, the connection is not suitable for use in areas of high seismic regions (Zones 3 and 4). 4. The design methodologies commonly used for this connection are conservative. 5. The connection can be modified to increase its ductile behavior by providing more surface area for concrete bearing, and by minimizing eccentric loads in the deformed anchor bars.

ACKNOWLEDGMENT

The authors would like to acknowledge the funding provided by XXsys

Technologies, Inc., and the Center for Composites in Construction at the University of Utah. The authors wish to express their gratitude to Eagle Precast Company (Monroc, Inc.), for providing the precast wall specimens. The authors would like to thank Vladimir Volnyy and Professor Janos Gergely for their assistance with the tests. In addition, the authors are grateful to Philip Richardson and Carl Wright of Eagle Precast Company for their suggestions. Lastly, the authors want to express their appreciation to the PCI JOURNAL reviewers for their thoughtful

REFERENCES

and constructive comments.

1. Rostasy, F. S., "Connections in Precast Concrete Structures ­ Continuity in Double-T Floor Construction," PCI JOURNAL, V. 7, No. 4, 1962, pp. 18-48. 2. Scoggin, H. L., and Pfeiffer, D. W., "Cast-in-Place Concrete Residences with Insulated Walls-Influence of Shear Connectors on Flexural Resistance," Journal of the PCA Research and Development Laboratories, V. 9, No. 2, 1967, pp. 2-7. 3. Abdul-Wahab, H. M. S., "Ultimate Shear Strength of Vertical Joints in Panel Structures," ACI Structural Journal, V. 88, No. 2, March-April 1991, pp. 204-213. 4. Spencer, R. A., and Neille, D. S., "Cyclic Tests of Welded Headed Stud Connections," PCI JOURNAL, V. 21, No. 3, May-June 1976, pp. 70-81. 5. Stanton, J. F., Hawkins, N. M., and Hicks, T. R., "PRESSS Project 1.3: Connection Classification and Evaluation," PCI JOURNAL, V. 36, No. 5, September-October 1991, pp. 62-71. 6. Schultz, A., Tadros, M. K., Juo, X. M., and Magana, R. A., "Seismic Resistance of Vertical Joints in Precast Shear Walls," Proceedings, XII FIP Congress, Washington, DC., May 29 June 2, 1994. 7. Low, S.-G., "Behavior of a Six-Story Office Building Under Moderate Seismicity," University of Nebraska, Lincoln, NE, May 1995. 8. Priestley, M. J. N., Sritharan, S., Conley, J. R., and Pampanin, S., "Preliminary Results and Conclusions from the PRESSS Five-Story Precast Concrete Test Building," PCI JOURNAL, V. 44, No. 6, November-December 1999, pp. 42-67. 9. Pincheira, J. A., Oliva, M. G., and Kusumo-Rahardjo, F. I., "Tests on Double-Tee Flange Connectors Subjected to Mono10. tonic and Cyclic Loading," Research Report, University of Wisconsin, Madison, WI, 1998. Strigel, R. M., Pincheira, J. A., and Oliva, M. G., "Reliability of 3/8 in. Stud-Welded Deformed Bar Anchors Subject to Tensile Loads," PCI JOURNAL, V. 45, No. 6, November-December 2000, pp. 72-82. Lemieux, K., Sexsmith, R., and Weiler, G., "Behavior of Embedded Steel Connectors in Concrete Tilt-Up Panels," ACI Structural Journal, V. 95, No. 4, July-August 1998, pp. 400413. Pantelides, C. P., Reaveley, L. D., Gergely, I., Hofheins, C., and Volnyy, V., "Testing of Precast Wall Connections," University of Utah, Department of Civil and Environmental Engineering, Report UUCVEEN 97-02, 97-03, 98-01, Salt Lake City, UT, 1997-98. Hofheins, C., "Welded Loose-Plate Connections for HollowCore Precast Wall Panels," M.Sc. Thesis, Department of Civil & Environmental Engineering, University of Utah, Salt Lake City, UT, May 1999. Building Seismic Safety Council, "NEHRP Guidelines for the Seismic Rehabilitation of Buildings," FEMA Publication 273, Federal Emergency Management Agency, Washington, DC, October 1997. SAP2000 Analysis Reference, Computers and Structures, Inc., V. I, Berkeley, CA, 1997. PCI Committee on Industry Handbook, PCI Design Handbook: Precast and Prestressed Concrete, Fifth Edition, Precast/Prestressed Concrete Institute, Chicago, IL, 1999. Salmon, C. G., and Johnson, J. E., Steel Structures Design and Behavior, Fourth Edition, Harper Collins College Publishers

11.

12.

13.

14.

15. 16.

17.

APPENDIX A -- NOTATION

Inc., New York, NY, 1996.

Ab As CU Fexx fs Fu fy n

= = = = = = = =

area of reinforcing bar area of deformed anchor bar compression force strength of electrode steel stress tensile strength of base material yield stress of reinforcement number of reinforcing bars

QCE QCL Rn t te TU VRU Vs

= = = = = = = = =

expected strength lower-bound strength strength of fillet weld or base material thickness of base material effective area of weld tensile force vertical shear force resisted by connection shear strength of connection angle of deformed anchor bar

33

July-August 2002

= capacity reduction factor

34

PCI JOURNAL

Information

13 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

560002


You might also be interested in

BETA
DG01.pdf
This Specification Index is based on the Construction Specifications Institute Master List of Titles and Numbers. Omissions from sequential numbering of Sections is intentional. The organization of the Specification into Divisions and Sections is not int
Microsoft Word - COMENTARIOS Bibliografia.doc