This action might not be possible to undo. Are you sure you want to continue?
H.A. Spieth, D. Arnold, M. Davies, J.B. Mander, A.J.Carr
University of Canterbury, Christchurch, New Zealand 2004 NZSEE Conference
ABSTRACT: This paper presents results of an experimental study together with companion analytical modelling of two distinctly different precast concrete beam-tocolumn joint connections. The first consists of precast concrete beams with armoured end connections connected directly to the column, while the second is a connection offset away from the column at about the 1/8 point within the span. In both cases the beams are connected via unbonded post-tensioned high strength prestressing threadbars to a prestressed concrete column. Lateral loading tests were conducted up to ±4% drift with and without supplementary mechanical energy dissipators. The results show that the nonlinear moment-rotation performance can be accurately modelled. From this study it is concluded, that with appropriate armouring of the precast members, damage can be avoided to the connection, while the entire structure is self-centred following an earthquake. These desirable performance features lead to the possibility of immediate post-earthquake serviceability of concrete frame structures. 1 INTRODUCTION Recent earthquakes around the world, e.g. Northridge 1994, Kobe 1995, have confirmed the adequacy of conventional seismic design techniques at preventing structural collapse but have also highlighted the extensive damage levels than can be expected in the process. A desire now exists to seismically design structures to not only preserve life safety but to do so with a minimum of damage accumulation. Previous experimental work performed by Priestley and MacRae (1996), Cheok and Lew (1993), Stone et al (1995), Priestley et al (1999), has shown that precast concrete frames constructed from post-tensioned segmental elements perform notably better than conventional frames when subjected to lateral loading. Structural deformations are accommodated via by gap openings at the precast component interfaces. Similar strength and interstorey drift capacities are achievable but damage to the structural elements is significantly reduced and the frames possess a re-centring capacity that results in negligible residual displacements. This paper reports the results of an experimental investigation that utilised the post-tensioned segmental element concept as developed previously but applied a Damage Avoidance Design (DAD) philosophy, (see Mander and Cheng, (1997)), in which the only permissible damage is that which occurs to replaceable components. An 80% scale precast concrete subassemblage representative of an interior beam-column segment in a multi-storey building was tested in orthogonal directions. One direction was assumed to be resisting lateral load only and the other was assumed to be resisting lateral load and carrying the gravity load of the structure. An innovative beam-column connection detail was developed for each. The development of the frame connection systems has been dictated by the requirement for an efficient construction process that avoids the use of any onsite concrete in the frame and minimises propping requirements. The following sections outline the adopted design procedure, give a description of the prototype connection details and review some key results from the experimental testing.
Paper Number 15
the high strength threaded rods inserted through the beam-column joint and tightened. The DAD objectives require the structure to remain in an essentially elastic condition even whilst subjected to the design drift demand. 10 000 10 000 N Threaded rods in column ducts Beam rocks on steel end plate assembly Figure 2: Beam-Column connection detail and construction sequence of lateral load only system 2 . High strength threaded rods passing through the beam-column joint form the connection between the steel plate end assembly and the column. the column design drift level was assumed to be 2% with an additional reserve drift capacity before yield of approximately 1%. The other constraint on this partially prestressed solution was that the maximum permissible concrete compressive stress be kept less than 0. Special design considerations were made for the rocking zones to ensure proper spread of the high contact forces into the beams. to be designed according to DAD. It is also desirable to provide reserve column drift capacity to allow for the extreme event drift demand. The initial prestress levels were set to ensure that the rocking joint connection opening at these drift levels could be achieved without yield of the post-tensioning bars. Bolted to column onsite. 9600 9600 9600 10 000 Subassembly Seismic Frame (N-S) Seismic plus Gravity Frame (E-W) Figure 1: Prototype structure Figure 2 depicts the beam column connection detail for use in the lateral load only moment resisting frame. An advantage of this system is that all stressing operations can be performed off-site enabling an efficient on-site construction procedure. For the purposes of the experimental study. post-tensioned to steel end plate assembly offsite. Joint rocking occurs at the interface between the steel plate end assembly and the armoured beam end. the beams are craned into position. The steel end plates assemblies are fabricated and the beams precast and post-tensioned all prior to transporting to site.7 f c' . A steel plate end assemblies house the post-tensioning anchors and are post-tensioned to the precast beam end. Once on-site. This effectively controls the crack formation and ensures full crack closure upon termination of lateral loading.2 DESIGN CONCEPT Figure 1 represents the prototype structure. Steel plate armourment was cast into the beam ends to protect the rocking interfaces. Fabricated steel end plate assembly Precast beam. Sufficient quantities of mild steel reinforcing was provided to ensure that even at the maximum loading yield of the longitudinal reinforcing was prevented.
are used to provide the rocking joint with moment resistance and balance gravity loading. representing the behaviour in both directions. Progressive cycles up to 4% column drift were performed. the dissipators were bolted over the rocking connections to the steel plate armourment but performed poorly due to the small tolerances in the assembly. The bolt bars can then be snug tightened with the full prestress applied at a later convenient stage in the construction process.The beam-column connection detail assumed for use in the gravity carrying direction is illustrated in Figure 3. The precast beam units with post-tensioning bars already in position and lightly pre-stressed are dropped into place between multi-storey column units. Draped post-tensioning bars. anchored in the columns at floor level. The lateral load only frame beam-column connection detail was tested first then the column was rotated 90 degrees and the gravity carrying frame beam-column connection detail tested. 1000 kN MTS Ram 600 1112 2842 4050 1170 560 Strong Floor 4000 4000 Figure 4: Experimental lateral load only subassemblage test setup (Davies (2004)) 3 1000 . In the prototype building this would be the location of a hinging floor plate mechanism. Figure 4 and Figure 5 show the experimental subassemblage test setup. Interlocking shear pintels on the beam and column stub faces locate and support the beam units removing any need for propping. The anticipated construction sequence is as follows. The design strength of the dissipators was governed by the criteria that the joint closing moment provided by the prestress be sufficient to overcome the resisting moment applied by the work-hardened dissipators. In later tests the dissipators were welded to the steel plate armourment and the energy dissipation characteristics improved significantly. Initially. Armoured Rocking Zone Anchor Plate. These were designed to dissipate energy either by flexural yielding or by repeated yielding in tension and compression. This test included the application of a gravity load via two floating hydraulic rams. Nut and Coupler Figure 3: Beam-Column connection detail of lateral load and gravity load system To increase the level of damping provided by the rocking joints various external mild steel energy dissipators were trialled. The threaded prestressing “bolt bars” are then inserted through ducts in the column and connected to the beam prestressing bars via couplers in the beam ends. 3 EXPERIMENTAL SETUP The system behaviour was tested on two subassemblies. The rocking beam-column interfaces are located approximately one member depth away from the column face.
Figure 6: RUAUMOKO model of lateral load only subassembly test setup 4 . The adjacent nodes of the post-tensioned springs were connected to the beam by newly developed friction elements which model the friction between the post-tensioning bars and the ducting that houses them. The bending stiffness of the cracked beam and column sections was adjusted according to Priestley et al (1996).1000 KN Ram 300 KN Gravity Load Ram Beam Roller Strut 2845 1905 8000 1905 Figure 5: Experimental lateral and gravity load subassemblage test setup (Arnold (2004)) 4 COMPUTATIONAL MODELLING In parallel to the experimental investigations a computational modelling strategy was developed using the RUAUMOKO finite-element code (Carr (2004)). Giberson beam elements with material properties determined from the experimental study were used to model the beam and column members. The rocking contact zones between the column and beams were modelled by a newly developed multispring contact element (Spieth et al (2004)). This allowed the accurate modelling of the compression zone formation and the position of the neutral axis at the rocking joints. The supplemental energy dissipation devices were modelled as hysteretic springs with an adjacent gap element to take into account slop in the connections. The models of the two subassemblies are shown in Figure 6 and Figure 7. For the case of the lateral and gravity load system the posttensioning springs were placed in agreement with the position of the draped post-tensioned bars in the experiment. The post-tensioning bars were modelled using spring elements. For the case of the lateral load only system the post-tensioning springs elements ran parallel to the centre line of the beams.
The system re-centred without residual drift after release of the applied load. In the first stage the energy dissipators were bolted to either side of the rocking connections. No hysteretic energy dissipation can be seen. drift of subassembly lateral load only system without supplemental energy dissipators. The numerical simulation shows good agreement with the experimental results. No damage or loss in stiffness or strength of the system was observed.5 -50 -100 -150 -200 -250 Drift (%) Experiment Simulation 0 0. 5. The joints start to open at approximately ±80 kN shown by the change in stiffness.1 Lateral Load Only System Figure 8 shows the system force vs. The bolt holes 5 . maximum drift 2 %.Figure 7: RUAUMOKO model of lateral and gravity load subassembly test setup 5 RESULTS The following section shows the results of the experimental investigations and the computational model in comparison. drift of the lateral load only subassembly without supplemental energy dissipators. During fabrication of the bolted connections every effort was made to ensure a tight fit.5 -1 -0. comparison of experimental results with RUAUMOKO simulation The subassemblage experiments with the tension-compression energy dissipators were run in two stages.5 2 Figure 8: System force vs. The displacement cycles from +2% to –2% drift were repeated several times. 250 200 150 System Force (kN) 100 50 0 -2 -1. The bilinear behaviour can be clearly seen.5 1 1.
2 mm. Confirmation of the sensitivity of the subassemblage response to even small tolerances in the connection was confirmed by the numerical model by placing a 1mm gap element in series with the dissipator springs. drift of subassembly lateral load only system.5 -1 -0. maximum drift 2 %. a bolt movement of nearly 1 mm due to play in the bolt thread and flexibility of the bolt in single shear was observed. The numerical simulation took this into account by setting the gap spring length to 0 mm. The numerical model displays a more abrupt change in stiffness than the experimental response as the gap element fails to capture the gradual flex of the bolt assembly. drift of subassembly lateral load only system with welded supplemental energy dissipators.5 1 1. Before the dissipator can re-activate in tension it must decompress and straighten out any buckling that may have occurred as a result of being previously yielded in compression. The deviation of the numerical simulation from the experimental response in the pre-gap opening phase is due to the presence of an initial compression force in the dissipator from the previous cycle. Three loading cycles were then applied to the subassembly. First one cycle to a maximum drift of 2 % was performed (Figure 9 (b)) followed by two cycles to 3 % drift (Figure 10). Despite this. This effectively eliminated the slop in the connection.5 2 Experiment Simulation Experiment Simulation (a) system with bolted energy dissipators (b) system with welded energy dissipators Figure 9: System force vs. The effect of this movement on the rocking joint performance can be seen in Figure 9 (a). system with supplemental energy dissipators.5 -100 -200 -300 -400 Drift (%) 0 0.75 % drift. In the second stage of testing with the tension-compression dissipators the dissipators were welded to the specimen. comparison of experimental results with RUAUMOKO simulation 400 300 200 System Force (kN) 100 0 -3 -2 -1 -100 -200 -300 -400 Drift (%) 0 1 2 3 System Force (kN) 400 300 200 100 0 -3 -2 -1 -100 -200 -300 -400 Drift (%) 0 1 2 3 Experiment Simulation Experiment Simulation (a) first cycle (b) second cycle Figure 10: System force vs.in the beams units were match drilled with the dissipator bolt holes to a tolerance of 0.5 2 System Force (kN) 400 300 200 100 0 -2 -1. maximum drift 3 %. (compare Figure 9(a) with Figure 9(b)).5 -1 -0. comparison of experimental results with RUAUMOKO simulation 6 .5 -100 -200 -300 -400 Drift (%) 0 0. The hysteretic energy dissipation is significantly improved. In the first cycle to 2 % drift it can be seen that the effect of slop in the dissipator connections has been eliminated. 400 300 200 System Force (kN) 100 0 -2 -1. Large pinching of the hysteretic response significantly reduced the energy dissipation characteristics with little or no energy dissipation below ±0.5 1 1.
comparison of experimental results with RUAUMOKO simulation Figure 12 shows the experimental response and numerical simulation of the subassembly cyclically 7 . 250 200 150 System Force [kN] 100 50 0 0. To take into account the effect of friction in the numerical model a new friction element was developed. This energy loss is due to friction between the unbonded prestressing thredbars bars and the plastic ducting that houses them.00 -0. the response loses stiffness due to yield of the post-tensioning bars.50 1.00 1. 5.00 -1.25 % drift and that the experimental response exhibits a small amount hysteretic energy dissipation. Beyond 2% drift. experience a relative displacement to the ducting as they are stretched by the rocking joint opening. maximum drift 2 %.00 -50 -100 -150 -200 -250 Drift [%] Experiment Simulation -2. The draped prestressing bars. This leads to a reduction in the post-tensioning force in the subsequent cycle.50 2. Further discussion on this element may be found in (Spieth et al (2004)). drift of subassembly lateral load and gravity load system without supplemental energy dissipators.00 Figure 11: System force vs. drift of subassembly lateral load and gravity load system without supplemental energy dissipators. which bear against the top surface of the ducting with considerable force.2 Lateral Load and Gravity Load System Figure 11 shows an initial test without supplementary energy dissipators to 2% column drift of the subassemblage with the lateral load and gravity carrying beam-column connection detail. comparison of experimental results with RUAUMOKO simulation 300 300 200 200 System Force [kN] 0 -4 -3 -2 -1 -100 0 1 2 3 4 System Force [kN] 100 100 0 -4 -3 -2 -1 -100 0 1 2 3 4 -200 Experiment Simulation -200 Experiment Simulation -300 Drift [%] -300 Drift [%] (a) first cycle (b) second cycle Figure 12: System force vs.50 -1.50 0. It can be seen that the simulation shows a good overall agreement with the experimental results. maximum drift 4 %.In the first cycle to 3% drift (Figure 10 (a)) the loading curve up to 2% drift is comparable to that of the previous cycle shown in Figure 9 (b). It can be seen that gap opening occurs at approximately 0.
The lack of damage and re-centring ability of an unbonded post-tensioned frame allows for the immediate post-earthquake serviceability of the structure.A. Seismic Design and Retrofit of Bridges. Seismic Damage Avoidance Design of Beam-Column Joints Using Unbonded PostTensioning: Theory. Stanton. 44. Seismic resistance of bridges based on damage avoidance design. J. G.J. Spieth.. Yield of the post-tensioning bars results in a drop in prestress leading to a lower joint opening force on the second cycle. Carr. Department of Civil Engineering. Vol. NCEER Technical Report 97-0014. D. 1997. 2004. 2004. Murahidy.. 64-81 Mander.. Modelling of PostTensioned Precast Reinforced Concrete Frame Structures with Rocking Beam-Column Connections. 6. PCI Journal. H. Seible.B. 6 CONCLUSIONS Based on the experimental investigation and computational modelling of two distinctly different precast concrete beam-to-column joint connections constructed in accordance with the principles of Damage Avoidance Design. Conley... January-February 1996. University Of Canterbury. PCI Journal. The post-joint opening stiffness is consistent with that observed in the tests to 2% column drift till yield of the post-tensioning occurs.. This can be seen as a softening of the response at approximately 3. University of Canterbury. Master Thesis.. and MacRae. pp. A.S. Computer Program Library. Pampanin. The numerical simulation captures this behaviour over the two cycles very accurately. No. Christchurch.N. Experiments and Design Example. Preliminary Results and Conclusions from the PRESSS Five-Storey Precast Concrete Test Building. “Seismic Tests of Precast Beam-to-Column Joint Subassemblages with Unbonded Tendons”. pp 80-92. Vol. Such advanced modelling techniques will be essential ingredient in the verification of future structural designs that incorporate the principles of damage avoidance. Priestley. ACI Journal. Calvi. December 10 Priestley. Christchurch. M. Master Thesis in Preparation. 38. Priestley. 8 . M. New Zealand. the following specific conclusions are drawn: • • • Appropriate armouring of the precast members can prevent damage from occurring to rocking beam-to-column connections. March-April.S. Department of Civil Engineering Department.C. J.S. 1995.. pp 43-67. John Wiley & Sons. S. 1996. Model Precast Concrete Beam-to-Column Connections Subject to Cyclic Loading. Mander. A.M. Davies.B. 2. Cheok. Carr.J. M. No.J. G. Christchurch. New Zealand. Lew.A.J.displaced to 4% column drift. Vol. PCI Journal.G. It should be noted that despite the excessive drift demands placed on the subassemblage no damage to the subassemblage was observed..R.. S. 1996. REFERENCES: Arnold. Cheok. ACKNOWLEDGMENTS This work has been generously supported by the German Scientific Society (Deutsche Forschungsgemeinschaft DFG). 2004 NZSEE Conference Stone. 1999. “Ruaumoko 3D: User Manual”. F. New Zealand.. W.5% drift. A. 2004.N. No. Sritharan. D. Cheng. Department of Civil Engineering Department. J. Performance of Hybrid Moment-Resisting Precast Beam-Column Concrete Connections Subjected to Cyclic Loading. Davies.. 2003. G.N. H. Arnold.N.. G. University Of Canterbury. M. The developed modelling approach provides accurate predictions for moment-rotation response of the proposed connection details. J. The posttensioned bars were re-tensioned to the original level and further experiments showed comparable behaviour of the subassembly as before the 4 % test. 4.91.F. M..J. 1993.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.