Professional Documents
Culture Documents
TRM Flat Slabs
TRM Flat Slabs
INTRODUCTION
Flat slabs are two-way spanning reinforced concrete slabs supported on an orthogonal (ie at 90°)
grid of columns. The soffit is at the same level throughout, which makes them easy to construct and
avoids downstand beams conflicting with services in the ceiling below.
An important variant has a thickened area at each column formed by lowering the soffit, called a
drop panel or simply a drop, see figure 1. This thickening could be raised rather than lowered (and
concealed within a raised floor), for which the name ‘drup’ has been coined. Flat slabs both with and
without drops can be prestressed, with either bonded or unbonded post-tensioned tendons (see
TRM 136).
Flat slabs must be designed to have adequate bending strength in each of the two principal
directions separately. The range of available methods of analysis is discussed below, but all manual
methods are based on carrying out two separate designs, one in each of the two principal directions,
each for 100% of the load. It is quite wrong to treat the load as split 50-50 between each of the
two directions. BS 8110-1 permits the analysis to be based on the single load case of maximum
design load on all spans, in which case the support moments should be redistributed downwards by
20%.
There are several different ways of analysing flat slabs in bending. The method which has
traditionally been used and is covered in detail in BS 8110-1 is the equivalent frame method. A
width of slab between the centrelines of adjacent bays and centred on a line of columns is treated as
a single beam element and with the columns is analysed as a multi-bay frame, see figures 2 and 3.
The slab is generally analysed as a beam on point supports, ie the stiffness of the columns is
ignored; see below for the effect on the columns. BS 8110-1 allows for the finite width of the
supports by permitting a reduction in support moments used to determine the reinforcement of
0.15 F hc (cl 3.7.2.7).
The slab element is marked out into column strips and middle strips (see BS 8110-1 figure 3.12),
and the total moment in each panel is then divided between the column and middle strips in
specified proportions, generally 55:45 in the span and 75:25 over the supports, see table 3.18 in
BS 8110-1. Reinforcement is provided for these moments overall. In addition, two-thirds of the
column strip support reinforcement is bunched into the central 50% of the width.
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
Figure 2. Cardington flat slab showing longitudinal frames and column and middle strips
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
Figure 3. Cardington flat slab showing transverse frames and column and middle strips
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
Note that the peak moments over a column head can be as much as 2.5 times the average for the
whole width of the column strip. The effect of bunching two-thirds of the reinforcement into half the
width of the column strip is an increase of 1.33, which means that the peak moment can still be
nearly twice as much as the reinforcement is designed for.
A simplified version of the equivalent frame method called the empirical method is now largely
discredited and was dropped from BS 8110 in 1985.
BS 8110-1 permits ‘other approaches’ (cl 3.7.1.2). The main alternative manual method is
Johansen’s yield-line analysis; this is simple to analyse and gives very straightforward
reinforcement arrangements; Kennedy and Goodchild have written a helpful guide Practical yield line
design. Hilleborg’s strip method is also available. These are both theoretically correct at ultimate
load, but give no guidance as to the behaviour at service load and therefore how to distribute the
reinforcement, risking higher deflections and crack widths than elastic methods. Some adjustments
of reinforcement layout to correlate with elastic behaviour should therefore be made.
Finite element programs are now readily available, and can give good results if the mesh is defined
realistically. Although modelling the section as uncracked and unreinforced is a major over-
simplification, it generally gives a reasonable pattern of bending moments although the 20%
redistribution may need to be introduced by hand. A not-too-complex modification is to change the
3
E-value (not the thickness) so that when it is used with Ι = b h /12 it matches the E Ι of the cracked
section – calculated as in TRM 110, allowing for sustained loading as necessary. Check also that
the program adds the Wood-Armer moments (which convert torsional to orthogonal moments so that
they can be reinforced for).
A development is iterative finite element programs which model the reduced stiffness of the section
as it cracks. At least one such program designs the reinforcement simultaneously, making it
complex and expensive – and slow to run. Better to input the reinforcement from a preliminary
calculation. These methods also beg the questions of whether the section is cracked and whether
creep is allowed for, see TRM 110.
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
Grillage programs have also been used, but there appears to be little or no experience within WSP.
They are covered in CIRIA report 110, but are not thought to be entirely appropriate.
Note that BS 8110-1 states if these other methods are used, ‘the applicability of the provisions given
in this section [presumably all of 3.7] are a matter of judgment’.
The equivalent frame method treats the width of slab between two panel centrelines as a beam
firmly connected to the columns, whereas in practice only a narrow strip is directly connected, with
the connection from the rest being laterally via torsion; thin concrete slabs are notoriously weak in
torsion. The method therefore overestimates moments transferred to the columns. This is how it is
dealt with in BS 8110-1. Refer to figure 3.13 of BS 8110-1. Some lines are incorrectly drawn, so this
is reproduced correctly below.
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
The maximum design moment which can be transferred to an edge or corner column is
2
Mt max = 0.15 be d fcu
where be is defined in figure 3.13.
The transferred moment Mt derived from the equivalent frame analysis should not be more than
twice this limiting value - equivalent to 50% redistribution. If the moment is derived from a finite
element analysis, only 30% redistribution is permitted. This check can affect the dimensions of the
columns relative to the thickness of the slab, and needs to be carried out early in the design.
A method of getting a more realistic moment transferred to the column has been published by
Murray et al, as follows. An equivalent column stiffness Kec is adopted and is given by
(1 / Kec) = (1 / Kc) + (1 / Kt)
where Kc = stiffness of column, Kt = stiffness of torsion beam (think of 1/K as the ‘flexibility’).
The torsion beam appears to be of width cy. If there is an upstand or downstand beam, C can be the
sum of rectangles as in textbooks on torsion.
The proportion of the fixed end moment (FEM) transferred into the columns by the BS 8110 method
is
0.893 / (0.893 + 5.20) = 0.146.
Taking the 50% permitted redistribution reduces this to 0.146 x 0.5 = 0.073.
The proportion of the FEM transferred into the columns by the alternative method is
0.292 / (0.292 + 5.20) = 0.053, ie smaller but comparable.
2 2 3
Mt max = 0.15 be d fcu = 0.15 x (0.25 + 0.40) x 0.225 x 35x10 = 173 kN-m,
and FEM = 0.833 x 16.03 x 7.53 = 564 kN-m,
so maximum permissible proportion = 173 / 564 = 0.307.
This suggests that the alternative method explains the behaviour more realistically, but that both
methods produce a moment which is comfortably within the maximum allowed to be transferred.
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
Internal columns
Because of the fixity provided by the ‘wrap-around’ effect of the slab, considerably greater moments
can be transferred to internal columns. Moments obtained from an equivalent frame analysis can
therefore be used, although a modest amount of redistribution would be reasonable.
In clause 3.7.2.3 of BS 8110, it is suggested that only half the width of the slab in the equivalent
frame should be taken ‘for horizontal loading’. This implies that flat slabs may be used as stability
frames, but in practice this should be limited to low-rise structures with broad plan forms (ie with a
substantial number of internal columns).
Shear in flat slabs is called punching shear, in which the failure surface is the frustum of an
upturned cone. The BS 8110-1 method checks the punching shear on a rectangular perimeter 1.5 d
from the face of the column or wall. Note the recommendation in watch-it C&S 64 that a circular
perimeter should be used instead of a rectangular one for circular columns.
Two recent papers have drawn attention to the need to correct for large columns and elongated
columns (or ‘walls’ where length 4 x width). One by Al-Yousif and Regan concludes that the
perimeter should be the sum of four sides each of length uj = (c + 3d) [1.09 – k(c/d)], where c is the
length of the side of the column or wall, d is the effective depth and k = 0.03 for sides perpendicular
to a predominant span (or all four sides if the slab spans approximately equally in two directions),
and 0.09 for sides parallel to a predominant span. The results of this are shown in table 1. Although
aimed at internal columns and walls, it would make sense to make the same reductions for edge and
corner columns and walls.
Table 1. Values of uj = (c + 3d) [1.09 – k(c/d)] for k = 0.03 or 0.09 and c/d from 1 to 10.
k = 0.03 k = 0.09
c/d c + 3d
1.09 – k (c/d) Uj 1.09 – k (c/d) Uj
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
The bars should be sized to have the ultimate capacity to support the load from the tributary area.
CIRIA 89 (section 5.3) sets out an approach to determining the capacity of the reinforcing bars. This
turns out to be based on the same work by Rasmussen used in TRM 112 Shear capacity of dowels
and bolts cast in concrete (which includes a calculation spreadsheet).
The limited research that has been carried out into punching shear failure suggests that the top steel
can resist 25% of the shear, therefore the bottom steel can be designed for 75% of the total shear.
In many cases passing two bottom bars designed for flexure in the span through or into the column
will satisfy the above equation. If separate bars are used, they should be well lapped with the main
bottom bars.
Eurocode EN 1992-1-1:2004 clause 9.4.1 (3) requires two or more bars in each orthogonal direction
passing through each internal column. It is not clear why perimeter columns are excluded, and the
guidance above should be followed.
Concrete Society Technical Report no 64 (published 2007) clause 5.3 states ‘A minimum of two
bottom bars, capable of taking a design force of 120 kN, in each orthogonal direction should be
provided at all columns and these should pass through the column’. No derivation of the 120 kN is
given. Assuming the capacity is based on tension, the area required is
3 2
120 x 10 / (500/1.15) = 276 mm .
2
The best fit with two bars is 2-B16 giving 402 mm , capacity = 175 kN. The guidance above is
preferred.
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
Flat slabs are less successful than other structural forms in accommodating holes and openings.
Holes close to columns are considered with punching shear above. Elsewhere, larger openings can
be accommodated, although the BS 8110-1 rules seem unduly restrictive, especially the requirement
for many of them to ‘be completely framed on all sides with beams to carry the loads to the columns’.
Provided openings are properly considered, eg by reducing the stiffness in the analysis and by
appropriate additional reinforcement, they can be satisfactorily accommodated in most locations.
The way round BS 8110-1 is to use a different method of analysis, eg yield line or finite element.
Deflection
The combination of thin sections and two-way spanning means that flat slabs are prone to quite
large deflections. The deflection at mid-span can be visualised by going from a column along the
grid line to the halfway point, and then turning 90° to the mid-point of the bay. This means that the
mid-bay deflection will be about twice the one-way deflection. Don’t forget that the reinforcement is
in two layers, so either use an average effective depth or calculate the two directions separately.
Remember also that key sections will probably be cracked, which means that conventional elastic
methods will usually seriously underestimate deflection. See TRM 110 for guidance on appropriate
section properties for calculating deflection.
The limiting span/depth ratios for beams and one-way slabs in BS 8110-1 have to be reduced by 0.9
for flat slabs. However, although the 0.9 reduction will reduce the deflection by 15-20%(?), given
that the mid-bay deflection will be approximately double that for a one-way span, even if the span is
treated as the distance between diagonally opposite columns, ie 1.4 L for a square grid, the target
(presumably L/250) could still be missed by a substantial margin, perhaps 20-25%. A greater
reduction, say 0.8, is recommended.
Vibration
With their longer spans and more deflection-prone behaviour, flat slabs are thought to be more at
risk of uncomfortable vibrations, although there have been few or no reported problems. Checks
should certainly be carried out before using flat slabs for dancing or aerobics.
Although drop panels improve the punching shear capacity, their principal purpose is to provide extra
strength and stiffness at the column heads where the moments are highest. For this reason,
BS 8110-1 requires drops to have their smaller dimension at least one-third of the smaller dimension
of the surrounding panels. The regions outside the drops can then be made thinner, thus using less
concrete overall and increasing the space for services in the ceiling. However, the thinner section
must still meet the span/depth rules (but not with the 0.9 reduction).
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
BS 8110-1 permits flat slabs with drops to be analysed taking into account the greater stiffness of
the drop areas. This has the effect of attracting more negative moment to the column heads, where
the extra thickness provides the additional capacity required. The 20% redistribution associated with
the all-spans-loaded load case should still be applied as it is not a redistribution but a means of
getting a realistic moment envelope.
Note that the extra stiffness will attract more elastic shear and thus increase the loads on interior
columns.
TRM 136 introduces post-tensioned flat slab solutions for offices and car parks, and the Concrete
Society has produced helpful guidance (eg TR 43). Bonded tendons are preferred to unbonded; with
unbonded tendons failure of an anchorage, eg in fire, would be catastrophic and the behaviour at
ultimate load capacity is difficult to understand. However, design and construction is better put in the
hands of specialist contractors, see TRM 100 Contractor designed items.
COLUMN GRID
Internally, ie away from the perimeter, the column grid is best kept regular, although the equivalent
frame method can be manipulated to accommodate changes of grid; yield line is an easier method.
A different grid on the perimeter (usually smaller to align with cladding or to double up as wind posts)
can be advantageous, and can be analysed very simply.
If additional support is required, it is best achieved with upstand beams as these do not interrupt the
flat soffit formwork.
STABILITY
It is recommended that stability is achieved by shear walls or other stiff members and not by frame
action, although at least one seven-storey building has been designed in this way. The reason is
that the inherent flexibility of flat slabs (especially the connection between the slab and perimeter
columns) means that the movements under lateral load will be larger than in other structures of
similar size.
TOOLS
The RCC (Reinforced Concrete Council) suite of spreadsheets, which are available on the intranet,
include a spreadsheet for the design of flat slabs by the equivalent frame method of BS 8110-1.
Elastic finite element methods are included in CADS, MasterSeries and other software packages.
Non-linear finite element methods are being pioneered by Integer.
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
REINFORCEMENT DETAILING
Correct choice of layer is important. If one span is longer than the other, put the long span bars in
the B1 and T1 layers. If the spans are equal pairing them into B1 and T2 one way, B2 and T1 the
other, is usually best.
Reinforcement has conventionally been detailed as loose bars in accordance with the rules in
BS 8110-1. However contractors specialising in concrete construction are moving towards more
labour-efficient systems, including pre-formed mats of various types.
The contractual arrangement in which the contractor details the reinforcement is preferable, but in
any case the contractor should be consulted before the reinforcement is detailed.
This list has been adapted from the list in CIRIA report 110, and only applies to non-prestressed flat
slabs with the same solid thickness throughout.
1. Assess required depth of slab from span/effective depth limitations factored down by at least 0.9.
Decide which direction of reinforcement will go in which layer (or use the average for equal
spans both ways), and allow adequate cover.
2. Check whether slab at a typical internal column needs shear reinforcement. Consider effect of a
200 mm square hole close to the column.
3. Calculate typical top reinforcement at an internal column and check possible congestion of
reinforcement.
4. Calculate typical mid-span reinforcement, check span/depth assumptions.
5. Check moments and shears at typical edge and corner columns. Ensure that the junction is
capable of transferring the required moment.
6. Check effects of large columns and rigid corners such as core walls.
7. Check likely position and effect of holes and openings. Decide whether they comply with the
BS 8110 rules, and how to design them if they don’t. Re-check punching shear if necessary.
Al-Yousif and Regan, Punching resistances of rc slabs supported by large and/or elongated
columns, The Structural Engineer, 4 March 2003, pp 30-34.
Source of guidance on large and long columns.
Concrete Society, Guide to the design and construction of reinforced concrete flat slabs, Technical
Report no 64, 108 pp, 2007 (library ref E/294).
Supersedes CIRIA 110, presumably more up-to-date and authoritative.
WSP Group
TRM 140
TECHNICAL REFERENCE Rev 9
Concrete Society, Post-tensioned concrete floors – design handbook (second edition), Technical
Report no 43, 110 pp, 2005.
The standard work for post-tensioned slabs, with Appendix G on vibration serviceability updated.
Kennedy G and Goodchild C, Practical yield line design, Reinforced Concrete Council, 2003.
Very easy to follow comprehensive guide to the subject.
Moss R M, Approaches to the design of reinforced concrete flat slabs, Building Research
Establishment, BR 422, 2001.
Easy to read guide prompted by the Cardington work, but biased towards buildability and yield line
design.
Murray K A, Cleland D J, Gilbert S G and Scott R H, Improved equivalent frame analysis method for
flat plate structures in vicinity of edge columns, ACI Structural Journal, Jul-Aug 2003, pp 454-464.
Source of alternative method for moments transferred into edge columns.
Regan P E, Behaviour of reinforced concrete flat slabs, 89 pp, CIRIA report 89, 1981.
Showing its age, but valuable nevertheless. Detailed treatment of anti-progressive collapse
reinforcement.
Teng et al (Singapore), Effect of large openings and column rectangularity on punching shear
strength of slabs, Dundee conference, 2002.
Source of guidance on long columns and allowing for holes.
Whittle R T, Design of reinforced concrete flat slabs to BS 8110, 66 pp, CIRIA report 110, second
edition, 1994.
Superseded by CSTR 64. Was the recognised guide, very comprehensive and clear. The scheme
design check list is particularly helpful.
KEYWORDS
Analysis; concrete; deflection; design; drops; flat slabs; post-tensioned; prestressed; progressive
collapse; punching; reinforced; vibration.
WSP Group