doneux, parung comp beam col subass 11ecee

7/28/2019 Doneux, Parung Comp Beam Col Subass 11ECEE

1/13


2/13

11th European Conference on Earthquake Engineering 1998 Balkema, Rotterdam, ISBN 90 5410 982 3

2

2 DESCRIPTION OF THE SPECIMENSThe structural characteristics of the three sub-assemblages, which are shown in Figure 1, are de-fined in the following :

Specimen BR-X (Bolted Rigid - testing in X direction) has rigid, fully bolted bi-axial beam-

column connections. The column is a HEM 260 steel section, the beams are IPE 300 in one direc-tion and IPE 270 in the other direction respectively. The slab is cast on a steel sheeting parallel tobeam IPE 270. It is 120 mm thick and the basic reinforcement is a Q513 welded mesh ( A = 5.13cm/m ). The composite section is a full shear connection composite section. The headed studs are16 mm diameter and 100 mm high. Three particular measures are taken to ensure the mobilizationof the maximal effective width of the slab:- the slab is in full contact with the column;- additional reinforcement is present to permit the development of additional strut and tie mecha-

nisms in the slab. Three additional bars 10 are placed on all 4 sides of the column;- additional studs on the beams around the connection are placed to maximize the participation of

the transverse beam in the transfer of the moment.The basis of the design is described in Plumier et al. (1998) and summarized in section 6.2.

Figure 1. Description of the specimens.


3/13


3

Specimen WR-X (Welded Rigid - testing in X direction) also has rigid bi-axial beam-columnconnections, but instead of being fully bolted, they are all welded. Another difference with speci-men BR-X is the particular layout of the additional reinforcement made of10 bars placed only on2 sides of the column. This specimen should set forward the influence of the bolted connection onthe global behavior of the node and the effectiveness of a particular reinforcement design.

Specimen BF-X (Bolted Flexible - testing in X direction) has fully bolted connections as thefirst specimen, with exactly the same basic dimensions, steel sections and concrete slab. Particularmeasures are here taken to minimize the effect of the slab in the beam-column moment transfer.Additional slab reinforcement and concentrated shear studs around the column have both beenomitted. Also, to eliminate any direct contact between the slab and the column section and endplates, 2 cm thick styrofoam strips have been placed around the steel section and plates.

3 TEST SETUPAs in a moment resistant frame under lateral loads, the inflection points in the beams and columnsare located at mid-length, the sub-assemblages were designed with hinges at those inflection points.

No vertical dead loads were applied to the specimen. The beam-column moment transfer is re-alized by applying a double-acting actuator load at the top of the column and providing the neces-sary hinged support conditions at the other inflection points, as shown in Figure 2. The actuator isprogrammed to introduce displacement-controlled loads. Out-of-plane motion of the test specimenunder load was prevented by parallel guide beams and a set of rollers.

Figure 2. Test setup Figure 3. Instrumentation of specimen BR-X

4 INSTRUMENTATIONAs the specimens were to be tested under displacement-controlled cyclic loads, a potentiometerwith a range of + and -200 mm was installed to control the top displacement of the column versus areference column located on the far side of the reaction frame. For measuring the applied cyclicloads, a load cell was attached to the 100 ton actuator. In order to measure the reaction forces in thepin-ended struts, one or both struts were instrumented with strain gages.

Strain gages were placed at both sides of the top flange and in the middle of the bottom flange atdifferent sections along the loaded beams. In order to measure in-plane transverse bending effects,


4/13


4

strain gages were placed at both sides of the upper flange of one of the transverse beams. Straingages were also placed on the reinforcing bars. LVDTs were installed on the concrete slab tomeasure the concrete strains at various locations. In order to measure the angular change betweenthe beams and the column in the connection zone under loading, LVDT's were placed on each sideof the column. Inclinometers were installed to measure the rotation of the column and beams.

Detailed layouts of the instrumentation and notations of the measurements are not presented foreach specimen, because they are not basically different. Figure 3 and 4 present the instrumentationof specimen BR-X, with: forces F1 -F2, displacement transducers D1 thr. D3, strain gages at sec-tions BS1 thr. BS4 on IPE 300 and gages BT1 thr. BT10 on transverse beam (IPE 270), incli-nometers I1 thr. I7 on IPE 300, LVDT's C1 thr. C7 on slab and strain gages R1 thr. R18 on rein-forcing bars.

Figure 4. Instrumentation of specimen BR-X.

5 TEST PROCEDUREThe specimens were subjected to cyclic, displacement-controlled forces (F1). The cyclic displace-ment (D1) history was defined by the immediate test objective and aimed at establishing the per-formance of the joints up to a test displacement limit of D1 = +/-200 mm (story drift = 6.67%).

Specimen BR was first tested under cyclic, increasing, top-column displacements in the x-direction (to study the IPE300 moment transfer) up to maximum displacements of +/-48 mm. Thespecimen was then turned 90 and subjected to the same displacement history. Finally, the speci-men was turned back to its original position and loaded up to maximum top displacements of +/-200 mm. It was showed that the x-direction test results were not affected by the intermediate y-direction testing. Consequently, specimen BF was directly tested in the y-direction up to maximumtop-column displacements of +/-50 mm. Subsequently, the specimen was turned 90 and tested inthe x-direction cyclically to maximum top displacements of +/200 mm. Specimen WR was testedonly in the x-direction up to the maximum top displacement of +/200 mm, because the reinforcingbars layout was specifically designed to calibrate formulas under a one directional earthquake.


5/13


5

6 EXISTING MODELS OF THE FORCES TRANSFER IN THE SLABSeveral presumed models exist to describe the behavior of the slab in composite joints or rein-forced concrete joints under vertical or horizontal loads. This section summarizes the different ap-proaches to which reference could be made during the analysis of the test results.

6.1 Eurocode 4Eurocode 4 deals with composite structures under vertical loads. The proposed annex J (Cost C11997) deals in particular with the behavior of composite joints. In case of unbalanced negativemoments, the additional tension at one side is equilibrated by additional compression on the col-umn at the other side of the joint through a strut and tie mechanism supported by the column. Inthis case, the longitudinal re-bars are designed to exclude failure in transverse reinforcement and toexclude brittle failure in concrete. Figure 5 illustrates this behavior.

In case of horizontal loading, this scheme is not directly applicable because the unbalance is of a

much higher order.

Figure 5. EC4 approach under unbalanced negative Figure 6. Anchorage at the sides of the column.

Moments (Cost C1 1997)

6.2 Assumptions used in the design of the tested specimensThe different specimens are defined to validate or invalidate design concepts developed in thecompanion paper Plumier et al. (1998). These design concepts aim at having very minor damagesin the slab, yielding being essentially taking place in the bottom flange of the steel section.

Additional transverse reinforcing bars are placed to allow the formation of an additional strutand tie mechanism at the positive bending side of the beam. The concept consider compressivestruts anchored not only at the face of the column, but also at the sides of it (Fig. 6).

The role of the transverse beam is assumed to be important, and is tried to be maximized byputting on it a number of studs greater than required by the bending resistance.

6.3 The concept of tension flange in reinforced concrete jointsOn the basis of several tests on reinforced concrete units, Paulay & Priestley (1992) postulated amechanism for in-plane shear transfer in a slab of reinforced concrete frames to enable flange ac-tion to develop. He demonstrates by global and local equilibrium considerations that the enhance-ment of the resistance of the beam is mainly coming from the reinforcing bars of the slab in ten-


6/13


6

sion, from which the so called tension flange. The corresponding resisting mechanism is drawn atfigure 7. At the negative bending side of the joint, the tension in the re-bars are transferred througha strut and tie mechanism to the core of the beam and equilibrated by the compression of the bot-tom part of the beam. These forces do not act in the same horizontal plane, allowing an additionalbending moment to be resisted at that beam section. At the positive bending side, the anchored ten-

sion coming from the negative moment side is equilibrated by the same resisting scheme, but the

Figure 7. Behavior of the slab as a tension flange - Maximum effective width of the compression flange

compression is in the same horizontal plane and does not bring more bending resistance. The trans-verse beam is assumed to play a very small role in the transfer of forces and is even neglected.

Paulay admit a small enhancement of the positive bending moment but not significant, becausehe assumes that the width of the compression flange is very limited around the column (Fig. 7).

Another important consequence of the postulated mechanism is that the development of tensileforces in a (cracked) slab requires the simultaneous development of tensile transverse forces ofcomparable amplitudes. From which the idea that it is not possible to anchor additional forces

coming from additional compression of concrete if the longitudinal and transverse reinforcementare of similar section. The detailed description of the model is given in Cheung et al. (1989).

7 EXPERIMENTAL RESULTS7.1 Cyclic Behavior - General CharacteristicsThe global behavior of the different specimens will be discussed on the basis of the behavior of theleft beam of each specimen tested in the x direction (i. e. bending of the IPE300). This choice wasmade necessary because of early failures in the right beam-to-column connection due to a missingfillet weld connecting the web of the IPE 300 web to the beam end-plate for specimen BR and a

lack of penetration of the bottom flange weld for specimen WR.Figure 8 presents the left moment versus global rotation curves for each specimen. The globalrotation is the ratio of the controlled top column displacement by the height of the column. It hasbeen considered as a good estimation of the global rotation of the beam because the joints havebeen designed to dissipate energy in the beams (column and panel zone remain elastic).


7/13


7

Figure 8. Moment at the end of the left beam versus global rotation of the node

An excellent ductile behavior can be observed in general with stable cyclic hysteretic loops forall the three specimens. Under positive bending, specimen BR-X with the most rigid slab and thepresence in the transverse direction of big plates (Christmas tree) shows the highest positive mo-ment capacities. Specimen WR-X shows lower but also good moment capacities, bringing to thefore the good influence of the big plates in specimen BR-X. In specimen BF-X, no additional rein-forcing bars, no studs around the column are present and styrofoam has been placed around thecolumn section and end plates, included those of the Christmas tree to which the IPE 270 beamswere connected. The positive resisting moments are logically lower than for the first two speci-mens. Under negative bending, the three specimens experienced flange buckling as illustrated bythe flattening of the curves after several cycles. The low cycle fatigue of the lower flange, betweencompressive buckling and high tension, finally lead all the specimens to tension cracks in the lowerflange of the steel beam. Crushing of the concrete slab in direct contact with the column is visibleon the rigid specimens only. Specimen BF-X showed also a light buckling of the upper flange -which was not connected to the slab in this case.

For the moment-rotation diagrams of each of the tested beam-to-column joint, a series of ex-perimental quantities have been computed: the conventional positive and negative plastic moments

Mp, which are deduced in accordance with the ECCS procedure (Mp corresponds to the intersec-tion between the line of elastic behavior and the plastic line, tangent to the monotonic envelopeof the cyclic diagram and having a slope of 1/10 of the elastic one.), the corresponding elastic rota-tions p, the elastic conventional joint stiffness Kc = Mp /p and the ductility defined as u /p.

Besides these experimental values, theoretical resistant bending moments have been computed,with actual yield or failure stresses. Mref.steel is the plastic moment of the steel section.

Mref.comp+ is the positive plastic moment of the composite beam with an effective width ob-tained using the concepts of Eurocode 4. The distance between the zero moment points lo is takenequal to 1.6 m and beff = 0.4 m. Mref.comp- is the negative plastic moment of the composite beam.

Two different effective width are considered; the Eurocode 4 effective width beff = 0.6 m cal-culated on the basis of lo = 1.5L = 2.4 m; and the total width of the specimen. Because of ductilityrequirements, Eurocode 4 disregard the contribution of welded mesh in the calculation of resistant

moments. Within the scope of capacity design, all the reinforcing bars present in the effectivewidth (including the mesh) are taken into account.The material properties and the computed plastic moments of beams are presented in tables 1

and 2. The results of the test data reduction described above are summarized in tables 3 to 5.

Table 1. Material properties

Concrete fc cylinder 24.5 N/mmStructural steel fy 287 N/mmReinforcing bars fs 570 N/mm


8/13


9/13


10/13


10

Figure 10. Measurements on transverse re-bars versus moment.

Figure 11. Spread of the forces in the slab of a composite section by shear mechanisms

Figure 10 illustrates the strains measured on the transverse re-bars versus the moment in thebeam. Whatever the specimen is, the studied transverse re-bars are mainly in tension. In a compos-ite element of the type studied here, it is the normal situation to have tension in the transverse re-bars under positive bending moment as well as under negative bending moment, as they allow forthe spread of the force in the slab by shear transfer (Fig. 11).

In specimen WR, where additional re-bars are only put in the transverse direction, one may ob-serve that the measured strains under positive bending moment are higher or equal to the strainsobtained under negative moment. This could validate the idea of an additional anchorage of con-crete struts on the sides of the column under positive bending.

In specimen BR where the re-bars are the same in both direction, the measured strains aresmaller or equal to the strains under negative bending, showing no additional anchorage, certainlybecause the forces in the longitudinal re-bars are directly equilibrated by the same forces in thetransverse re-bars and that there is no possibility if additional anchorage of concrete struts.

In specimen BF, no additional re-bars were placed. The measurements are made on the mesh.The re-bars are also the same in both directions. The tendency is the same as for specimen BR

but the strains under negative bending are far greater than those under positive bending. This dif-

ference is not yet explained.

7.2.3 Measurements on longitudinal reinforcing barsThe types of measurements on the longitudinal re-bars in one particular section of the two fullyconnected specimens (BR WR) are described at Figure 12.


11/13


11

Figure 12. Measurements on the longitudinal re- bars Figure 13. Presumed applied forces on the trans- inone section of specimen BR-X at D1 = +/-120 mm verse beam for the rigid specimens

One can observe that under positive bending, some bars are in compression and some are not. Itseems to invalidate Paulays tension flange postulate. However, its existence is possible if super-posed with an important contribution of the compression flange in the surrounding of the column.

This contribution seems to be more important in the case of this composite specimen than in thecase of the concrete units tested by Paulay. The bars near the column are still in compression (ef-fective width of the positive moment) while the bars near the edge of the specimen are in tension(anchorage of the bars in tension at the negative bending moment side tension flange). Figure 13is a sketch of the forces imagined on the transverse beam with the longitudinal strain measurementspresented here. The distance between two points of zero strain could be imagined to derive effec-tive width under positive moment at a certain level of loading in the studied section.

For the more flexible specimen BF, the strains under positive bending are always in compres-sion, showing that no tension field develops in this specimen. In Figure 13, the tension Tx have tobe replaced by compression and the transverse beam is expected to be more loaded.

7.2.4 Stress field in the slabThe moment transfer mechanism is very dependent on the detailed design, which directly influ-ences the magnitude of stresses in the reinforcing bars and in the slab.

In specimen BR, it is obvious that yielding is located in the steel part of the beams, and conse-quently no slab failure mechanisms are clearly observed.

In specimen WR, the smaller concentration of re-bars in the main direction make possible theformation of concrete struts at 45. The strain measurements on longitudinal re-bars show tensionand compression under positive bending moment, giving substance to a mechanism superposingthe tension flange resistant scheme coming from the negative moment side and a compression ofthe slab in the surrounding of the column.

Specimen BF did not follow this resistant scheme. It is not illogical because around the column,everything is done to minimize the effect of the slab. However, it is now sure that the slab plays arole (see Mp comp exp/Mp steel = 1.16 or 1.29), maybe linked with the role of the transverse beam.

7.3 Transverse beamThe strain measurements on the transverse beam are aimed to quantify the participation of thetransverse beam to the moment transfer from the beam to the column. The upper flange of thetransverse beam is instrumented with couples of strain gages at 5 different sections. If the trans-verse beam is loaded by membrane forces in the slab, the induced torsion of the IPE profile willmainly introduce bending in its upper flange (Plumier et al. 1998). Under elastic loading, the meas-ured couples of strains may be divided into normal and flexural strains. Assuming that the resistantsteel section is the upper flange, an elastic bending moment and a normal force are computed.

The bending moments are presented in Figure 14 for the 3 specimens for a same load at the top


12/13


12

of the column F1 = 165 kN, corresponding to the top displacement D1 = 30-25-60 mm for speci-men BR-WR-BF respectively.

Figure 14. Moment in the upper flange of the transverse beam

The moment observed in the transverse beam of specimens BR-X and WR-X remains near zero.It means that the mechanism of struts and ties imagined with the transverse beam to contribute

to increase the positive bending moment is not activated, or far less than in the specimen BF. Thiswas well suggested by Figure 13. The presence of big plates on the transverse beam of specimenBR gives greater moments in the Christmas tree plate near the column than in the IPE270 beam ofspecimen WR (no Christmas tree). But the difference is not important enough to derive a quantita-tive conclusion at this stage.

In comparison, the transverse beam of specimen BF-X is strongly loaded and yield strains areeven observed. The fact that there are no additional reinforcing bars and no additional studs leadsto a greater flexibility of the specimen because the other ways to anchor struts and carry the loadare not effective. The absence of additional devices makes the activation of the transverse beammechanism possible. The greater stresses in the transverse beam of specimen BF-X may also beexplained by the absence of studs 50 cm around the column. So, if compressed struts rest againstthe studs, the level of arm will be greater in this specimen than in the specimen BR-X and conse-quently, for a same load, the stresses will be greater.

These interesting observations remind that all the resistance mechanisms play a different rolefollowing their relative rigidity in the general resistant scheme and that one particular layout willpromote one mechanism and weaken the other.

8 CONCLUSIONAt present, one practical conclusion can already be drawn from the tests described here: the design

relations presented in Plumier et al. (1998) gave a safe design, bringing the intended yieldingscheme for a value of bending moment which can rather accurately be computed. However, thecontribution of a slab in its role, as a tension or compression flange of a beam, carrying membraneforces, are complex phenomena, which still need to be deeper studied.

In this paper, we have described tests on three composite beam-to-column sub-assemblages: onebolted connection with detailed rules for rigid slab, one welded with rigid slab and one bolted witha more flexible slab. This description provides indications on the field of stresses existing in theslab for various layout of re-bars, studs and contact between concrete and column. On the basis ofthe indications, attempts to understand the exact mechanism of force transfer in the slab in the con-nection zone have been made. Further analysis of the data and the development of a very detailednumerical modeling of the connection zone, now under way, should allow in a near future a com-


13/13

doneux, parung comp beam col subass 11ecee

Documents