Civil Engineering Reference
In-Depth Information
view that the main contributor to shear resistance is aggregate interlock
(Fenwick and Paulay, 1968; Taylor, 1968; Regan, 1969). This is because only
through aggregate interlock can the cracked web be the sole contributor to the
shear resistance of an RC T-beam, as specified by current code provisions e.g.
BS 8110 (British Standards Institution, 1985). The concept of the shear
capacity of critical sections is itself a prerequisite for the application of the
truss analogy because it is the loss of the shear capacity below the neutral axis
that the shear reinforcement is considered to offset.
It appears therefore, that the aggregate interlock concept, although not
explicitly referred to, forms the backbone of current concepts that describe
the causes of shear failure. And yet this concept is incompatible with
fundamental concrete properties; a crack propagates in the direction of the
maximum principal compressive stress and opens in the orthogonal direction
(Kotsovos, 1979; Kotsovos and Newman, 1981a). If there was a significant
shearing movement of the crack faces, which is essential for the
mobilisation of aggregate interlock, this movement should cause crack
branching in all localised regions where aggregate interlock is effected. The
occurrence of such crack branching has not been reported to date.
The inadequacy of the concepts currently used to describe the causes of
shear failure has been demonstrated in an experimental programme
(Kotsovos, 1987a, b). The programme was based on an investigation of the
behaviour of RC beams, with various arrangements of shear reinforcement
(
Figure 2.8)
,
subjected to two-point loading with various shear span to depth
ratios
(a/d).
The main results of this programme are given in
Figure 2.9
which shows the load-deflection curves of the beams tested.
On the basis of the concept of shear capacity of critical sections, all
beams that lack shear reinforcement, over either their entire shear span or a
large portion of its length, should have a similar load-carrying capacity.
However, beams C and D were found to have a load-carrying capacity
significantly higher than that of beams A which had no shear reinforcement
throughout their span. Beams D, in all cases, exhibited a ductile behaviour,
which is indicative of a flexural mode of failure, and their load-carrying
capacity was higher than that of beams A by an amount varying from 40 to
100% depending on
a/d
. These results indicate that such behaviour
cannot
be explained in terms of the concept of shear capacity of critical sections
and the failure of the beams cannot be described as a shear failure as defined
by this concept.
The evidence presented in Figure 2.9 also counters the view that
aggregate interlock makes a significant contribution to shear resistance. This
is because the large deflections exhibited by beams D, in all cases, and
beams C, in most cases, led to a large increase of the inclined crack width
and thus considerably reduced, if not eliminated, aggregate interlock. In fact,
near the peak load, the inclined crack of beams D had a width in excess of 2
mm which is an order of magnitude larger than that found by Fenwick and
Paulay (1968) to reduce aggregate interlock by more than half. It can only be
