Steel brace-to-RC frame post-tensioned bolted connection

STEEL BRACE-TO- RC FRAME POST-TENSIONED BOLTED CONNECTION

Adrian Dogariu; Sorin Bordea; Dan Dubina “Politehnica” University of Timisoara, Romania

[email protected]; [email protected]

ABSTRACT

The paper presents and evaluates the technical solution for a post-tensioned bolted connection applied to inserting infill steel braces in order to improve seismic behavior of reinforced concrete frames; in present case, buckling restrained braces (BRB) have been considered. In order to validate a simple design approach that es-tablishes the connection behavior, a complete finite element model was build. The proposed connection technique was validated through experimental tests, conducted on the entire RC frame.

1. INTRODUCTION Many of contemporary buildings located in seismic zones were erected more

than 50 years ago and being designed without a proper consideration of the seismic action present a high seismic vulnerability.

The resistant structure of these buildings is either masonry or RC frames. In case of RC frames, it is well known that 50 years ago, in the early period of seismic engineering design codification, the reinforcing details of the joints and plastic zones did not appropiately satisfy the seismic demand, and unfortunately, the subsequent important earthquakes proved that. From this reason, it is important to provide a bet-ter resistance and ductility of these structures, by appropriate consolidation meas-ures. One of the most efficient retrofitting measures, from the structural and techno-logical point of view, is based on inserting dissipative steel braces. This technique improves both the global resistance and stiffness of the structure, and the ductility. To obtain a higher capacity in terms of energy dissipation and resistance, buckling restrained braces (BRB) may be used. Even if, theoretically, the retrofitting solution itself, seems to be very efficient, the proper connection between the RC frame and the new steel elements plays a very important role. In fact, without a full interaction, the new elements cannot bring their positive effect in the global behavior. Usually, the connection between the existing RC elements and the new steel elements is made by post-installed mechanical or chemical anchors. A good behavior of this type of connection is highly dependent on the quality of the concrete, which generally, at that time was poor. The failure, if occurs in the concrete members can be fragile. It is possible to provide a local strengthening of concrete members using FRP but, in some way, this solution might disturb the installation of anchors. Present paper pro-poses a connection solution which uses post-tensioned ties and a steel detailing, which also enables for a local confinement of concrete.

2. CONNECTION SYSTEM

2.1 Technical solution The steel brace – to - RC frame connection is realized with post-tension high strength steel ties, which together with the steel plates, act like a metallic clamp. Fig-ure 1(a) and (b) presents the solutions sketch of the upper connection between the RC beam, respectively the bottom connections between the RC columns and the in-serted steel braces (BRBs). The technology is rather simple and did not require highly qualified workers or special devices other than a torch control wrench. Another important technological aspect is the possibility to remove the connection.

Figure 1. (a) Upper connection between RC beam and BRB (b) Bottom connection

between RC column and BRB.

The high strength steel ties introduce controlled a normal pressure, at the interface level between RC elements and steel plates, and activates frictional forces when relative slippage tends to occur. To obtain a good behavior of the connection, a proper post-tension level must be calibrated and introduced. In order to provide a high frictional force it would be desirable to avoid completely any relative slippage through high post-tension level. In the same time, the prestressing level, which basi-cally ensure the grip between the RC members and steel brace, also introduces a biaxial state of stress in the concrete elements, thus improving the material behavior by node confinement. There are two limitations for increasing the pre-stress level, i.e. the steel elastic re-sistance and crushing strength on concrete. In order to avoid any plastic deformation in the connection, tick plates are recom-mended to be used. Even so, in case of base plate failure, the connection will exhibit a ductile behavior. A large contact area between steel plates and RC elements is de-sired, so strongly stiffened plates are to be used, because their bending must be lim-ited. A flexible plate would bend excessive and the contact area between steel plate and RC beam might reduced. This will increase the local pressure, which might in-duce crushing of concrete. The present paper shows the application of the proposed connecting system for in-serting inverted V BRB as a seismic retrofitting technique to upgrade an existing RC frame building which was initially designed considering only the gravity loads (Nar-dini, 2008). The geometry and main dimensions of RC frame members are pre-sented in Figure 2, and all the details regarding the number and diameters of rebars in element cross section, the distance between stirrups are given in Figure 3.

(a) (b)

Figure 2. (a) Axonometric view of RC frame (b) Geometry of the planar frame (Nar-dini, 2008).

Figure 3. Reinforcement details for the studied RC frame members

To validate the proposed technique, an isolated portal RC frame extracted from the second floor of the previous presented RC building (see Figure 2), was experimen-tally tested in CEMSIG laboratory. The RC frame was tested with and without BRB devices. The following characteristics of BRBs were obtained from the pre-experimental numerical simulations: an 1.7 m length rectangular cross section active zone (yielding segment) of 30x10 mm and a 0.4 m length transition zone of 60x10 mm. A S235 steel quality has been ordered for the BRBs, but a 335 N/mm2 yield limit has been experimentally determined. A C20/25 concrete (Rc = 20.5 N/mm2) was in-tended to be obtain, but material tests have revealed a higher characteristic resis-tance of Rc = 35.5 N/mm2. First, a Performance Based Seismic Evaluation (PBSE) of initial r.c. building frame was done in order to define the intervention objectives. After that the PBSE of the frame strengthened with BRBs was performed to quantify the demands for braces and design them accordingly. 2.2 Design procedure: simplified method Using simple assumptions and considering the full prestressed ties, an initial dimen-sion of steel plates and of the number of steel ties was determined. Based on the

application principles stated above the connection have been made by a 25 mm tick steel plates S235 steel grade and 10.9 grade for the post-tensioned M16 ties. Considering a very stiff steel plate, with a contact area of 300x504 mm with the con-crete element, the maximum force applied to all bolts, corresponding to a full preten-sion level of the ties (Ft = 109.9 kN per bolt, equivalent to a torque moment of Ms = 200Nm) creates in the concrete element at the interface with the steel plate, a nor-mal pressure σ of approximate 8.75 MPa (see Figure 4(a)) at the interface with the RC beam, which is much smaller than the compressive strength of the RC beam, i.e. 35 MPa. Consequently, the friction force (Ff) between the steel plate and the concrete ele-ment should be higher than the cumulated two horizontal components of BRB forces (see Figure 4(b)). Using a static friction coefficient of 0.45 between concrete and steel, a maximum shear force of 595 kN can be considered without any relative slip-page, corresponding to a number of 12 prestressed ties.

Figure 4. (a) RC beam – BRB connection: pressure under the steel plate due to bolts

pre-stressing (b) Load transfer and forces on the upper connection. The connection slip capacity depend on the number of prestressed ties, and the total slip force that must be withstand depend on the rigidity of the bracing system and on the imposed transient interstorey drift of the reinforced concrete frame for a certain level of performance, i.e. 1% - Immediate Occupancy, 2% Life Safety and 4% Col-lapse Prevention (FEMA 356). In a simple manner the designer have two options to determine the required shear force for which the connection must withstand. According to the first one, the de-signer must establish the desired performance level and the corresponding intersto-rey drift. At this level of displacement at interface between RC beam and the connec-tion steel plate a shear force will appear. Second option asks the designer to deter-mine the maximum shear force that can be induced into connection when the BRB attains tension capacity. Both of these approaches must give the same results be-cause the BRB cross section should be chosen according with the same imposed interstorey drift limitation. Depending on this shear force the designer can determine the necessary connecting force ( Ft, see Figure 4(a)) and obtain consequently the diameter, quality and number of prestressed ties. The connection between the RC beam and the steel elements can be considered as fully rigid, without any slippage, until the shear capacity is reached (attainment of ul-timate friction force). For the base-column connection, the dimension of the bolts have been computed in order to resist the BRB forces (both vertical and horizontal components) and the prestressing level was calibrated to avoid the slippage of the column (see Figure 5).

Figure 5. Load transfer and forces on the bottom connection.

3. EXPERIMENTAL BEHAVIOR OF THE CONNECTIONS 3.1 Experimental specimens and set-up The testing set-up includes the entire RC frames together with the BRB (see Figure 6(a)). The connection between RC column and steel BRB was introduced in a simple manner, using a very stiff steel element, in order to simulate a real RC joint (see Fig-ure 6(b)).

(a) (b)

Figure 6. (a) Experimental frame (b) experimental connection between RC column and BRB.

Displacement transducers have been used to determine all the relative displace-ments and rotations that can appear in the connections. Four displacement trans-ducers were used on the bottom of each column, to measure the vertical displace-ment (see Figure 7) the horizontal displacement of the connection and the load transfer from the connection to the column and bottom part (see Figure 8).

Figure 7. Measurement of the vertical displacement of the connection between the

BRB and the column.

Figure 8. Measurement of the horizontal displacement of the connection between the

BRB and the column. Two displacement transducers have been placed on the RC beam for monitoring the slippage of the upper connection (see Figure 9).

Figure 9. Measurement devices distribution: horizontal on the RC beam.

Monotonic and cyclic tests were considered. A modified ECCS cyclic loading proto-col was applied in the cyclic tests (ECCS, 1986). 3.2 Experimental results All the connections between BRB system and RC frame (BRB to column and BRB to beam) showed a very good behavior and no slippage was recorded. No important slippage between steel plates and concrete columns was emphasized (see Figure 10).

Horizontal displacement [mm] Vertical displacement [mm]

Figure 10. Behavior of connection between BRB and left RC column during cyclic test.

When the left side BRB failed in tension, the horizontal displacement recorded at the connection between BRB and the RC was limited to 5 mm (see Figure 11)

BRB (left side)

-200

-150

-100

-50

0

50

100

150

200

-6 -4 -2 0 2 4 6

Beam Connection Displacement [mm]

MR

F Fo

rce

[KN

]

Figure 11. Connection between the BRB and the beam during cyclic test – hysteretic

curve.

4. FINITE ELEMENT MODELS OF CONNECTIONS 4.1 Description of finite element models Considering the number of contact zones of the specimens (between RC frame, steel ties and steel plates), ABAQUS Explicit product has been preferred. This prod-uct has the advantage of an explicit formulation of the equilibrium, which is well-suited to simulate and to effectively handle severely nonlinear behavior, such as con-tact. This product was build mainly for dynamic events, but using proper amplitude and mass scaling of the model, quasi-static loading response can be obtain with very good accuracy and in a shorter computing time. For the preliminary verification of the connecting systems prior to the experimental tests, two numerical models of the both connections have been made. Using ABAQUS software, two detailed models have been made. Due to the symme-try of the model and to the load conditions, for the base-column connection between the RC column and BRB only half of the specimens have been made; with respect to the symmetry boundary conditions (see Figure 12). 3D deformable finite elements (C3D8R) have been used. No prestressing of ties have been applied.

Figure 12. Finite Element Model of the connection between BRB and RC column.

In the case of upper connection between BRBs and RC beam a more complex nu-meric model was used. To replicate with higher accuracy the behavior of the connec-tion and the loading conditions, the entire RC frame model was build (see Figure 13). The tested frame has a 4.5m span and 3.2m total height, with rectangular cross sec-tion of 30 by 30 cm for beam and column. 3D wire elements (B31) for RC columns

and RC beam corners, and 3D solid elements (C3D8R) for the middle part of the beam in the connection area for introducing the contact properties at their interface, to allow development of frictional forces, have been applied. The BRB members were modeled also as wire having only axial properties (k = 28000N/mm). The sup-ports were considered pinned, the behavior of the hinge connection between RC col-umn and BRB being considered perfect-ideal without any significant slip or deforma-tion. The experimental tests confirmed these assumptions. The bolt prestressing load cannot be introduce in the model as a load option in case of explicit procedure, instead connecting elements loaded with connector force to simulate the prestressing, can be used (Montgomery, 2002). Considering the small bending rigidity of steel ties, an axial type connector was chosen to model the bolt rod. These types of connector provide a translational connection between two nodes that acts along the line connection the nodes. Elastic behavior only was considered (k = 145000N/mm), no plasticity or damage behavior options were define, because during the desire behavior mode, i.e. no relative slip occurs, the elongation in the connector element must be almost zero so the load level in the connector must re-main almost constant. The bolt head and the nut were model using solid elements (C3D8R) to allow the occurrence of local forces between bolt head and steel plates and bolt rod and plates holes. All the steel plates from the connection were modeled as solid elements (C3D8R). The contact behavior was define as “hard” contact with the possibility of separation after contact in case of normal behavior and for the tangential behavior a penalty fric-tion formulation with a 0.45 static friction coefficient was used. Because the purpose of this paper was only the study of the slip-resistant connection behavior, the plastic behavior of the reinforced concrete elements was neglected. However, the level of interface stress was monitored to avoid the normal pressure to exceed the concrete compressive resistance value of 35 Mpa, and the equivalent von Misses stress in the steel plates. The analysis was conducted in two steps. First the pretensioning load was applied in the connectors, and secondly a linear horizontal displacement of RC frame corner was imposed (see Figure 13).

First step: connector pretensioning load

Second step: horizontal imposed displacement

Figure 13. Finite Element Model of the entire experimental frame with the boundary

and load conditions.

4.2 FE results The numerical simultation enabled to observe the effectiveness of friction contact in-between the concrete and steel plates. Both normal and tangential stresses have been considered. At the end of the first step a tension stress of 700 MPa was obtained in the bolts, while the total clamping force Ft was 1300 kN,this corresponds to approx 110 kN in each connector see figure 14(a). A slight bending only of the steel plate was ob-served (see figure 14(b).

(a) (b) Figure 14. (a) Pretension forces in steel ties and (b) vertical deformation and bend

shape of the steel plate. The numerical simulation shows the steel plate does bend in such way to introduce to high pressure on the concrete beam corners and cause excesive local damages (see Figure 15). The upper steel plate stress level (see Figure 16) slightly exceeds the yield limit value in limited area.

(a) (b) Figure 15. The bending of the upper steel plate (a) (10x) and the crush zones in con-

crete beam (b).

Figure 16. Von Misses stress on the upper connection after the application of the

pretension in ties (deformation scale factor of 10).

It was also checked the relative slippage at the steel-concrete interface (see Figure 18) at the expected level of force in order to observe if the connection works as a full shear joint. The slip (see Figure 17) occurs at a 37.48 mm horizontal displacement, correponding to a total applied force of 757 kN. The numerical analysis showed that the global force at which the friction resistance is exceeded, which represents the failure criteria of the upper connection, is much higher that the capacity of retrofitted RC frame.

Global behavior curve of RC frame

0.E+00

1.E+05

2.E+05

3.E+05

4.E+05

5.E+05

6.E+05

7.E+05

8.E+05

0 10 20 30 40

Top displacement (mm)

Bas

e re

actio

n fo

rce

(N)

Friction connection behavior

0.E+00

1.E+05

2.E+05

3.E+05

4.E+05

5.E+05

6.E+05

7.E+05

8.E+05

9.E+05

-1.5 -1 -0.5 0 0.5 1 1.5

Relative displacement - slip (mm)

Bas

e fo

rce

reac

tion

(N)

(a) (b)

Figure 17. Behavior curves of RC frame (a) and friction connection (b)

(a) (b)

Figure 18. (a) Relative slippage of the RC beam (b) von Mises stress in steel plates when the friction force is exceeded

The higher friction force can be explained by the higher force in the connectors and the slight clamping of connection plates around concrete beam (see Figure 19). The induced force in the BRB is 640 kN, and taking into account that its plastic capacity is 105 kN, and the ultimate resistance is 135 kN, it means, at this level the BRB core is already broken.

Figure 19. Vertical displacement and the clamp of upper connection at the slip (10x)

For the bottom connection (with the RC column), the stress level in the ties, bolts and in the concrete elements was also analyzed (see Figure 20).

Figure 20. Von Misses stress on the bottom connection (15x).

The results of the numerical simulations have been used to validate the geometry of the specimens, obtained considering the simplified assumptions.

5. CONCLUSION AND FURTHER DEVELOPMENT The connections between the steel braces and the RC members performed very well under both monotonic and cyclic loading. The connection devices used for installing BRB’s within the frame took benefit from the friction resistant forces induced by the post-tensioned ties and showed a very good behavior. However, related to the moment when the BRB internal stress are transferred to the ties, these are, in fact, pre-stressed. Very reduced slippage only was observed, with very small influence on the hysteretic loops of BRB system. The experimental results recommend the application of this connecting system for such interventions. Moreover, in the case of multi-storey frames (see Figure 21), such connecting systems provide also a beneficial confining effect at the frame joints, en-hancing both strength and ductility of the MRF+BRB system. The portal frame tested experimentally could not take such benefit.

Figure 21. Bottom connection in case of multi-storey frames.

This type of connection may be used successfully for all types of retrofitting tech-niques for RC moment resisting frames based on inserting dissipative steel bracings. For a global analysis, both connections can be considered fully rigid. It is on interest to develop a more complex nonlinear model based on calibrated FE models in agreement with the experimental results. This model should consider not only the behavior until the attainement of the bearing capacity, but also the real rigidity of the connections and the post-slippage behavior. As a further study, the behavior of the connections on dynamic loading conditions should be considered.

ACKNOWLEDGMENT The proposed techniques were developed and tested in the frame of STEEL-

RETRO project (RFCS-CT-2007-00050 Steel Solutions for Seismic Retrofit and Up-grade of Existing Constructions) at CEMSIG Research Center from “Politehnica” Uni-versity of Timisoara.

REFERENCES Bordea, S. (2010), “Dual frame systems of buckling restrained braces”, PhD Theses,

Ed. Politehnica, Timisoara, ISBN 978-606-554-059-0 ABAQUS Version 6.8 Documentation (2008) © Dassault Systems Simula Montgomery (2002), “Methods for modeling bolts in the bolted joint”, ANSYS User

Conference, 2002 FEMA 356 (2002), Guidelines for Seismic Rehabilitation of Buildings, Vol 1, Wash-

ington DC, 2002 Nardini L., Braconi A., Osta A., Salvatore W. (2008), “Definition of the reinforced

concrete benchmark building for the execution of comparative performance analysis between steel intervention techniques”, STEELRETRO benchmark

ECCS (1986), “Recommended testing procedures for assessing the behaviour of structural steel elements under cyclic loads”, ECCS Publication N.45, 1986

PROHITECH (2003), FP 6 INCO-CT-2004-509119/2004 „Earthquake Protection of Historical Buildings by Reversible Mixted Technologies 2003

STEELRETRO (2006), RFSR-CT-2007-00050 Steel solutions for seismic retrofit and upgrade of existing constructions, 2006