The property of shear connector most relevant to design is the relation-ship between the shear force transmitted, P, and the slip at the interface, s This load-slip curve should ideally be found from tests on composite beams, but in practice a simpler specimen is necessary Most of the data on connectors have been obtained from various types of push-out or push test. The flanges of a short length of steel I-section are connected to two small concrete slabs. The details of the standard push test of Eurocode 4 are shown in Fig 2.9 .The slabs are bedded onto the lower platen of a compression-testing machine or frame, and load is applied to the upper end of the steel section slip or frame, and load is applied to the upper end of the steel section. Slip between the seel member and the two slabs is measured at several Slip between the steel member and the two slabs is measured at several points, and the average slip is plotted against the load per connector. A typical load-slip curve is shown in Fig .2.10, from a test using composite slabs(23)
In practice, designers normally specify shear connectors for which strengths have already been established, for it is an expensive matter to carry out sufficient tests to determine design strengths for a new type of connector. Lf reliable results are to be obtained. The test must be specified in detail, for the load-slip relationship influenced by many variables, including:
(1) number of connectors in the test specimen,
(2) mean longitudinal stress in the concrete slab surrounding the connectors,
(3) thickness of concrete surrounding the connectors,
(4) thickness of concrete surrounding the connectors,
(5) freedom of the base of cach slab to move laterally, and so to impose uplift forces on the connectors,
(6) bond at the steel concrete interface.
(7) Strength of the concrete slab, and
(8) Degree of compaction of the concrete surrounding the base of each connector,
The details shown in Fig .2.9 include requirements relevant to items 1 to 6. The amount of reinforcement specified and the size of the slabs are greater than for the British standard test, (22) which has barely changed since it was introduced in 1965.The Eurocode test gives results that are less influenced by splitting of the slabs, and so give better predictions of the behaviour of connectors in beams.(15)
Tests have to be done for a range of concrete strengths, because the strength of the concrete influences the mode of failure, as well as the failure load. Studs may reach their maximum load when the concrete surrounding them foils but in stronger concrete, they shear off. This is why the design shear resistance of studs with hld ≥ 4is given in Eurocode 4 as the lesser of two values:
the enlarged diameter at the weld collar at the base of the stud, shown in Fig 2.6; but it is clear that the effective compressive strength is several times the cylinder strength of the concrete.
This very high strength is possible only because the concrete bearing on the connector is restrained laterally by the surrounding concrete, its reinforcement, and the steel flange. The results of pushtests are likely to be influenced by the degree of compaction of the concrete, and likely to be influenced by the degree of compaction of the concrete, and even by the arrangement of particles of particles of aggregate, in this small but critical region. This is thought to be the main reason for the scatter of the results obtained.
The usual way of allowing for this soatter is specify that the characteristic resistance PRk be taken as 0%below the lowest of the results from three tests, and then corrected for any excess of the measured strength of the results from three tests, and then corrected for any excess of the measured strength of the connector material above the minimum specified value.
The load-silp curve for a connector in a beam is influenced by the difference between the longitudinal stress in a concrete flange and that in the slabs in a push test ,Where the flange is in compression the load/slip ration(the stiffness)in the elastic range exceeds the push-test value ,but the ultimate strength is about the same .For slabs in tension (e.g. in a region of hogging moment),the connection is significantly less stiff(24) but the ultimate shear resistance is only slightly lower .This is one reason why partial shear resistance is only slightly lower .This is one reason why partial shear connection (Section 3.6) is allowed in Eurocode 4 only in regions of sagging bending moment.
There are two situations in which the resistance of a connector found from push tests may be too high for use in design .One is repeated loading, such as that due to the passage of traffic over a bridge. This subject is covered in Chapter 10(Volume 2). The other is where the lateral restraint to the concrete in contact with the connector is less than that provided in a push test, as in a haunched beam with connectors too close to a free surface (Fig.2.12).For this reason, the use of the standard equations for resistance of connectors is allowed in haunched beams only where the cross-section
of the haunch satisfies certifies certain conditions, In Eurocode 4, there are that the concrete cover to the side of the connectors may not be less than 50 mm (line AB in Fig .2.13),and that the free concrete surface may not lie within the line CD, which-runs from the base of the connector at an angle of 459with the steel flange .A haunch that just satisfies these rules is shown as EFG.
There are also rules for the detailing of reinforcement for haunches, which apply also at the free edge of an L-beam.
Tests show that the ability of lightweight-aggregate concrete to resist the high local stresses at shear connectors is slightly less than that of normal-density concrete of the came concrete of the same cube strength. This is allowed for in Eurocode 4 by the lower value of Ecm that is specified for lightweight concrete. For concrete of density 1750 kg/m3,the resistance given by equation (2.15) is only 73% of that for normal-density concrete .This is considered in the UK to be too low, the corresponding ratio in BS 5950(14) is 90%.
