You are on page 1of 3

CONCRETE AND COMPUTERS

The use of computers in the


design of concrete structures
John Morrison, Buro • a belief that computer analysis is ware writers. In particular, the method for
Happold and Tony Jones, quicker and the only way to compete in designing punching shear reinforcement
providing a low fee service can be problematic and, in the authors’
Arup Research and • low fees, leaving senior engineers with experience, most engineers end up carrying
Development insufficient time to teach approximate out this calculation by hand as computer
methods values are, at best, variable. This is not to
• lack of guidance from industry journals say that no capable software is available,

T
he aim of this article is to consider
computer use in the design of con- on practical design methods and but that the engineer needs to be sure that
crete structures and how this is approximate solutions – too much the package being used takes into account
changing the design process. It is often emphasis on highly theoretical, and all the parameters required in the code.
argued that this change is not for the necessarily computer-driven, solutions
• a belief that software calculations must Column fixity
good of design. Problems may arise
when trying to produce reinforced con- be correct and that hand calculations Some slab packages assume pins at col-
crete designs from computer output. The can only be crude approximations umn locations; this is normally accept-
article contains a checklist of items of • increasing complexity of modern codes able for slab flexural design. However, if
which the engineer should be aware deterring engineers from hand calcula- bending moments are developed, they
before relying on the output received. tions as they believe they can no longer will affect both punching shear design
Simple methods for checking the validity provide the complete answer. and column design. If the column-pinned
of computer output are highlighted and Over-reliance on computer output is assumption is made, a separate frame
the need for a body to check proprietary not only a concrete issue and has been analysis may be required to calculate col-
software is discussed. addressed in other publications. The umn moments.
Institution of Structural Engineers pub-
The current situation lished guidelines on the use of computers Boundary conditions
Today’s young engineers often appear last year(1). This report was, in part, a Owing to the inherent continuity of rein-
unwilling to undertake analyses by hand. response to concerns raised by the forced concrete construction, elemental
Whether an approximate calculation to Standing Committee on Structural Safety design packages often require restraint
confirm the viability of a concept or the (SCOSS)(2). conditions to be specified at the boundary
final design analysis, the preferred Design consultants have acknowl- edges. The forces generated then need to
approach is to use a computer. In itself edged the problem. For example, Buro be transferred to the restraining member.
this is not necessarily a bad thing and Happold has a policy where a hand calcu- The element design package will not
computers can be valuable in understand- lation must be made before the computer check the validity of these boundary con-
ing behaviour. For example, by varying analyses is started and the computer ditions and the engineer must do so. For
different parameters, the sensitivity of a model is accepted only if it corresponds example, a beam framing into a thin wall
solution to a number of variables can be to the hand calculation within defined may be closer to pinned than fixed. Many
quickly assessed. However, if the reliance limits. Arup has introduced a one-day frame programs overestimate the mom-
is such that the engineer loses the confi- course, Design without the computer, ents transferred between flat slabs and
dence to carry out simpler methods of where the objective is to teach young boundary columns. Inadequate consider-
analysis, the ability to carry out an inde- engineers how approximate answers can ation of these factors could lead to crack-
pendent check of their model is compro- be quickly calculated to check more ing in the wall, under-design of the beam
mised. In addition, in the search for an detailed analyses. Both companies have and, in the case of flat slabs, cracking and
exact answer, the nature of reinforced internal handbooks giving approximate additional deflection.
concrete construction is forgotten. The solutions and schematic element sizes.
Torsion
material is not elastic, stresses are induced
Reinforced concrete design When a structure is not dependent on tor-
through many mechanisms other than
load (shrinkage, thermal effects, creep) All these points are equally applicable to sional resistance for equilibrium, most
and the process of sensible rationalisation other structural materials. However, there codes say that torsion can be ignored.
and detailing often negates the benefits of are a number of issues specific to rein- However, if torsional stiffness is present
any perceived additional accuracy. forced concrete. It is suggested that the in a computer model, the equilibrium
There are numerous factors behind this engineer should understand how the pro- found will rely on torsion and the tor-
situation, some of which are listed below. posed computer program deals with each sional stresses developed should be
It is important to realise that all parts of the of these issues before analysis is under- designed for. Some packages deal with
industry have played their part: taken. torsions in the post-processing of results,
• the ready availability of software in the and some assume that the torsional resist-
Shear strength ance of all elements is zero. Others will
office, in universities, and free from the
internet Shear design is basically empirical and leave it to the engineer to take into
• lack of education in approximate meth- requires additional rules to define the account. Again, the engineer must under-
ods, some universities placing too boundaries under which code values are stand the assumption implicit in their
much emphasis on obtaining the exact applicable. These vary from code to code design and the computer package being
answer and can be interpreted differently by soft- used.

40 May 2003 CONCRETE


Stiffness cracking. This invalidates deflection beam add up to wl2/8; use of standard
assumptions, and at worse causes failure end restraints will also allow a loose
In many packages, stiffness is simply the
of restraining elements, e.g. shear failure check on the magnitude of redistribu-
elastic section properties of the concrete.
of end columns in long continuous struc- tion assumed
This is usually adequate for an ultimate
tures. These effects are unlikely to be • checking the load increase (and face
limit state design, but will tell the engi-
considered in the computer analysis and shear) in a column at any given floor is
neer little about the likely serviceability
the danger is that over-reliance on the approximately equal to the load on the
performance. Stiffness is affected by the
computer output will result in the engi- floor area notionally supported by the
elastic modulus of the concrete, which
neer also ignoring them. column
varies because of creep. A simple short-
Several of the above issues come • checking the moments at flat slab inte-
term/long-term modulus may not be rel-
together in flat slab design, where there rior columns comply with where m and
evant for the case considered. It should
is a trend to use thinner and longer spans. m / are the moment capacities in hog-
also be noted that for a C40 concrete to
This has, in part, been encouraged by the ging and sagging at the column loca-
BS 8110-2: 1985 Structural use of con-
reduction in the safety factor for rein- tion and P is the column reaction
crete – Code of practice for special cir-
forcement and by the use of high- • checking selected members with the
cumstances(3) gives a range of elastic
strength concrete. As a result, flat slab charts in BS 8110-3: 1985 Structural use
modulus of 22–34kN/mm2. This can be
design tends to be controlled by punch- of concrete – Design charts for singly rein-
refined if the aggregate type is known.
ing shear and the serviceability limit forced beams, doubly reinforced beams and
However, if an accurate prediction of
state. There is concern that inadequate rectangular columns(5); although the tables
serviceability behaviour is required, rely-
description of stiffness is leading to a sit- still incorporate a 1.15 safety factor on the
ing on default values could be problem-
uation where some flat slabs will not per- steel, they should still highlight any gross
atic.
form as intended in the long term. As a errors
Stiffness is also affected by the tensile
minimum, reinforced concrete design • checking that span-to-depth ratios are in
strength of the concrete that dictates the
packages should include warnings when accordance with the code, unless a deflec-
extent of cracking. In practice, there are
code span-to-depth ratios are being tion analyses acknowledging the non-lin-
large variations in tensile strength for any
exceeded. ear behaviour has been carried out.
given compressive strength, and this ten-
These checks may be considered by
sile strength varies with age. Normally, it Hand checks some as overly simplistic. However, it is
is the strength at the time of cracking that
The term ‘hand check’ is perhaps wrong. the authors’ opinion that such simple
is important. Once the concrete is
What is required is a progression from checks can locate most gross errors intro-
cracked, tension stiffening comes into
simple analysis that can be easily verified duced by using computers. The checks
effect, and a number of methods can
through to final analysis, with each step can be valuable in understanding the
model this. It is important to understand
being validated by the one before. accuracy of design. For example, there
the limitations of any model. In terms of
Consider the design of an irregular are at least three ways you can design a
stiffness, reinforced concrete is far from
flat slab. Initial sizing and reinforcement stocky column to BS 8110 – equation 39,
an elastic or even elastic–plastic material,
quantities may be determined from span- the charts in Part 3, and by considering
and models for accurate prediction of
to-depth ratios and the moment tables in plane sections. For the same column, a
serviceability behaviour need to be thor-
BS 8110: Part 1: 1997 Structural use of range of 24–32mm bars is dependent on
oughly checked and the underlying
concrete – Code of practice for design the method. In this particular case, the
assumptions considered. It is certainly
and construction(4) or yield-line analysis. simplest method (equation 39) was actu-
not acceptable to predict the deflection of
This may then progress to a two-dimen- ally found to be more accurate than the
members by assuming uncracked elastic
sional frame of a more typical bay, and Part 3 charts. Discrepancies were easy to
properties, particularly with slender
finally to a grillage or finite element explain and added to the understanding
members where the cracked stiffness is
analysis, if required. At the extreme, if of those involved.
significantly reduced.
deflections are critical, the finite element There must be a wealth of similar or
Redistribution analysis may need to be non-linear. Each better approximate checks/design meth-
Codes generally allow redistribution layer builds on the previous, discrepan- ods known to CONCRETE readers, and
from the moments calculated using gross cies can be considered and, if not readily it is hoped that this note will prompt sug-
concrete properties. Non-linear analysis explainable, the models checked. It gestions. If enough contributions are
will ‘automatically’ allow some redistri- should be noted that even for a regular made, these may be combined into a
bution, due to cracking. If further hand flat slab, there is a tendency to move to a paper on simplified methods. It would
redistribution of the moments is under- finite element analysis. With current soft- also be interesting to hear whether, and
taken, greater overall redistribution than ware this may be quicker than a two- how, universities teach approximate
that assumed by the code will be implied. dimensional frame, but any assumption methods. It is clear from examples of
In any case, suitability of code rules deal- that the answers are any more accurate is coursework and exam papers provided
ing with detailing should be carefully not necessarily correct. With experience, by University College, London, that
considered when non-linear analysis has it will soon be found that the hand meth- some universities require the use of
been used for the ultimate limit state. ods give results that are reliable if not simple methods in developing a design.
better than the computer approach. As so little design is analysis, students
Intrinsic movements Other more straightforward checks should understand that the computer is
Unlike structural steel, concrete changes include: where the design is refined not necessar-
in volume through shrinkage and heat of • checking that the sum of reactions ily begun.
hydration. The amount is often signifi- equals the applied load; this check is
cant when compared with normal in- not only valid for the model but also Validation
service temperature variations. If such for individual elements within it The final issue to be considered is the
movements are restrained, significant • checking that the support and mid- validation of software. Some practices
stresses can build up and may cause span design moments in a continuous validate their own software formally,

CONCRETE May 2003 41


Figure 1: Final
CONCRETE AND COMPUTERS analysis of the
base to an off-
shore concrete
gravity struc-
ture. Initial
NEWS from
design was car-
ried out by
hand using
standard tables and assumptions of uniform bearing
pressure. During the design process, soil structure
interaction and actual applied load distribution were
taken into account. The initial design was then
checked against this final analysis. Association of Concrete Industrial
Flooring Contractors
others relying on a network of users to highlight prob-
33 Oxford Street, Leamington Spa, CV32 4RA, UK
lems. What is clear is that there is little or no exchange
Tel: +44 (0)1926 833 633 Fax: +44 (0)1926 423 236
of validation information across the industry. This is
E-mail: acifc@hotmail.com Website: www.acifc.org.uk
costly and could be perceived as wasteful. Possible
solutions are:
• An independent organisation to verify software. This The current situation
would need careful planning in terms of liabilities if
a validated program is shown to have an error. Two years of intensive debate and discussion about concrete industrial
• The development of a set of validation models for floors led to the launch, on 5 March, of Concrete Society Technical Report
checking software. These would need to be secret to 34 Concrete industrial ground floors – a guide to design and construction.
avoid ‘tuning’ of the software and always run the risk ACIFC is now undertaking a new round of problem-
of being obsolete, as new developments require vali- solving, setting the pace for higher standards, contractor accreditation and
dation. It would, however, be possible to define the renewed industry partnerships.
extent of validation clearly. There is continuing Government indifference to the retention system. Led
• The setting up of a voluntary network of users within by the National Specialist Contractors Council, representation is being made
the industry so that problems can be circulated and on this matter and in relation to the increases in insurance premiums affect-
discussed. ing the entire construction industry.
All these have their difficulties, and further discus- The practicalities of floor construction and finished slab performance
sion is welcomed. A further plea is for software houses remain the concern of specialist contractors. In particular, poor-quality sub-
to produce detailed documentation on the technical strata are always a challenge, especially as demand for redevelopment of
assumptions made. brownfield sites increase.
Floor contractors and foundation specialists are combining to achieve
Conclusions greater predictability in relation to geotechnical factors. The factors
enabling achievement of a successful floor are seen as a partnership
Computer packages greatly increase the speed of the
between designer, main contractor, specialist contractors and their suppli-
design of concrete structures. However, there is concern
ers, together with an improved design and construction code.
that there is an over-reliance on the output and insuffi-
cient education on approximate methods of checking/- Piled floors
designing. There is a need for users to understand the
The ACIFC working party is seeking assistance in the provision of published
assumptions behind the software they are using, and an
references relating to design, construction and performance monitoring of
even greater need for engineers to learn simplified or
suspended ground floor slabs, thereby speeding preparation of a fundamental
approximate methods. Similarly, software validation
guidance document.
needs to be co-ordinated. Comment on all these aspects
To date, most information has been sourced from steel-fibre-reinforcement
is welcomed, to enable better guidance to be produced
specialists. It is hoped that those involved with steel bar and mesh alternatives
and perhaps show how the industry can move forward
will make their views known. While the new TR34 covers broad design of this
with software validation. ■
construction type, much better knowledge is needed to help flooring and main
Contact John Morrison on +44 (0)1255 320 600. contractors understand the components of a good piled floor.
E-mail: john.morrison@burohappold.com Web tendering
The method of tendering for large floor construction has developed electroni-
References:
cally, with supermarket Tesco seeking bids for new floor construction over
1 THE INSTITUTION OF STRUCTURAL ENGINEERS. the internet. While this process increases pressure for lower prices, the main
Guidelines for the use of computers for engineering
calculations, 2002. contractor has already identified the preferred bidders.
2 STANDING COMMITTEE ON STRUCTURAL SAFETY. This has two effects: the bidder has no knowledge of competitors, but as
Structural safety 1997–99: Review recommendations: 12th all pricing has to be received at a given moment, Tesco identifies the lowest
Report of SCOSS, The Institution of Structural Engineers,
London. 2001. bid and gives a half-hour reflection time, during which any bidder can revise
3 BRITISH STANDARDS INSTITUTION. BS 8110-2: 1985 the submission.
Structural use of concrete – Code of practice for special
circumstances.
The second effect is to increase collaboration between the designer, spe-
4 BRITISH STANDARDS INSTITUTION. BS 8110-1: 1997 cialist contractor and supply chain to enable partnership on floor detailing,
Structural use of concrete – Code of practice for design and allowing submission of a revised bid.
construction.
5 BRITISH STANDARDS INSTITUTION. BS 8110-3: 1985
Structural use of concrete – Design charts for singly
Annual golf
reinforced beams, doubly reinforced beams and rectangular The ACIFC Annual Golf event will take place on 5 June at Hawkstone Park,
columns.
near Shrewsbury. ■

42 May 2003 CONCRETE

You might also like