Role of Specification, Mix Design,
and QC in Developing Durable,
Economical and Sustainable
Concrete Construction
Ken W Day
FACI, FICT, Hon CIA
WHAT NOT TO DO!
• Specify concrete by strength supplemented by
a generous minimum cement content or a
maximum W/C ratio “for durability”
• Require a trial mix to be approved and not
departed from without a further trial
• View cement replacement materials, if
permitted at all, as diluting cement content.
WHY?
• It has been clearly established that materials such
as fly ash, slag, silica fume, rice hull ash, superfine
silica etc are far more effective than a low W/C
ratio in reducing permeability and therefore
improving durability.
• Specifying a minimum cement content destroys
any incentive for a concrete producer to become
competent – it explains why many US producers
are more than 20 years behind Australian
producers in developing good quality control.
WHAT ELSE IS IMPORTANT?
• The best Quality Control is that which is able
to detect any change in mix performance at
the earliest possible moment and quickly
compensate for it.
• Any low 28day strength should have been
predicted and compensated for WEEKS before
it occurs
• It is ridiculous to wait for enough low 28day
results to statistically confirm a downturn!
WHAT TO SPECIFY
•It is desirable that you should NOT specify a
unique set of requirements unnecessarily
•If a suitable mix is already in production,
advantage can be taken of history to date.
•At a given strength level there are still many
properties to consider:
Shrinkage, Pumpability, Heat generation,
Early strength for demoulding or prestressing,
Low bleeding for good (formed) surface finish,
A short-term test for long-term durability is
urgently needed (RCP?)
MIX DESIGN
• Mix design has recently changed significantly
• W/C ratio is no longer useful in design
• The important aspect of Grading is no longer the
sand % and grading curve but the packing density
of the fine materials
• Some producers may need assistance in arriving
at a suitable design
• But ultimately the producer must “own” the
design in the sense of being responsible for fine
tuning of the mix in production
COMPETITIVE MIX DESIGN
• It is very easy to design a mix over the telephone
if what is needed is good “ordinary” concrete and
there is no competition
• It gets hard when you are competing with other
producers and a large variety of materials is
available
• In time, producers will have experimented with
many mixes and materials and must be
encouraged and allowed to use the materials and
proportions found best economically when they
can demonstrate satisfactory results
“SUPER” CONCRETE
•
•
•
•
High strength? 100MPa fairly easy, 200 limit?
Pumpability: limit about 500m?
Durability??? Permeability the criterion?
For all these attention is needed to the
grading/packing density of cement with finer
materials, slag, pfa, silica fume, RHA, superfine
calcium carbonate, nano silica
• I do not have a mix design system to give a
definitive answer for this and experimentation is
necessary
QUALITY CONTROL
• A margin has to be provided between the
required properties and their average
• Strength may not be what matters most, but it
is the best means of detecting change
• The financial benefit of reducing variability is
easily calculated and can be substantial
• Other benefits such as smoother working and
better appearance are less calculable but can
be even greater
PROBLEM DETECTION
• A major benefit of good control is earlier
detection of problems
• To put a number on it, a downturn of half a
standard deviation is detectable, a wide range
since SD can be anywhere from 2 to 6MPa
• But consider the effect on pumpability,
appearance of formed surfaces, variable
finishing times, wear resistance and surface
tolerance of floors................
HOW LONG TO DETECT DOWNTURN?
• Specifications the world over deal with analysing
a number of 28day results
• Action really needs to be taken BEFORE there are
ANY low 28day results!!
• A downturn in mean strength can usually be
detected in 3 or 4 7day (or 3day) results using
cumulative sum graphing
• The cause may be detected even earlier than this
using cusum graphs of 24hr density/unit wt,
slump and temperature, batch plant error or sand
grading
WHO SHOULD INITIATE CONTROL
ACTION?
• It is obvious that a consulting engineer cannot
demand action on the slender grounds
proposed
• The concrete producer should be the one
initiating mix revision
• He may do this to avoid future trouble, but I
think a clear (small) cash penalty is desirable -- the one thing that almost no-one in the
world agrees with me!!
STARTING ON QC
• Good QC has to start with accurate testing and
accurate batching
• It has to cover the variability of input materials,
especially sand, cement and coarse aggregate
also important but (hopefully) less likely to vary
• When these items are under control, a large
proportion of day-to-day variability is usually
due to variations in water content
WATER CONTENT VARIATION
• This is usually caused by one of: temperature,
slump or sand grading
• Increased water reduces density so it can be
detected by weighing cubes/cylinders at 24hr
• So if your slump or temperature shows an
increase and 24hr density is down, you can
already be sure that your 28day strength will
be lower
TESTING ACCURACY(1)
• Defective testing will obviously invalidate QC
• It can be detected by a high pair difference
between 28day specimens or a variable 7 to
28day strength gain but this is very late.
• It can be due to variable compaction or curing,
defective capping of cylinders, defective cube
moulds or more than a dozen other causes
• My experience is that the best testing is usually
achieved by the concrete producer, who is the
one who suffers financially if it is sub-standard
TESTING ACCURACY(2)
• So we should be continually watching average
pair differences and 7 (or3) to 28day gain
since they can be automatically available, but,
if detected, urgent action is needed
• Average pair difference should not exceed
1MPa and individual differences over 2MPa
should be investigated and the lower of the
pair discarded.
CUSUM CONTROL GRAPHING
• So now we know what items must be
controlled in order to obtain low variability
concrete
• “Change points” must be distinguished from
random variation and CUSUM (Cumulative
Sum) graphs are the answer.
• What is graphed is the cumulative sum of
differences from a target value
MULTIGRADE, MULTIVARIABLE,
CUSUM QUALITY CONTROL (MMCQC)
• The breakthrough is to use the current average value
of a variable as the target, this has two major
benefits over a fixed target:
• a) you don’t need to calculate and revise a target
• b) several grades of concrete can appear on the same
graph of each variable, each automatically having its
own target, but differences being accumulated as if
all from the same target
• There is not time here to explain the technique in
detail, some explanation is in the written paper, but
for a full account you need to go to my website
www.kenday.id.au
AVAILABLE MMCQC PROGRAMS
• There are at least two versions of this program:
• One is the very comprehensive one shown on the
website as Conad but now owned and marketed by
Command Alkon as Command QC. This copes with
the dozens of variables which can affect concrete.
• What I am presenting here is a small free program
“KensQC” which covers only temperature, slump,
density and strength at two ages (3 or 7 and 28 days)
This little program will very promptly detect the
existence of any problem but can only detect the
cause if it is one of the monitored variables or substandard testing. It works well if batching error, and
sand and cement quality variation, are already under
good control.
USE AND EFFECT OF MMCQC PROGRAM
• The small program can be operated by a nontechnical clerical person who simply enters results
each day and raises an alarm if the (automatic)
graphs show a change of slope.
• The essential user is the concrete producer who
must react to change and identify cause
• The program is also suitable for use by a consulting
engineer who wants to assure himself, with very little
effort, that QC is being well done.
• In the initial stages control could be duplicated by an
international consultant taking only a few minutes a
day
CONCLUSION AND RECOMMENDATION
• The objective of this presentation has been no less
than to transform how concrete is specified,
produced and controlled
• It presents the view that the only satisfactory
situation is for concrete producers to develop
expertise in the design and control of concrete mixes
• They will not do this unless they can see the prospect
of profiting by acquiring the necessary expertise.
• This requires that specifications be based on
required properties, allowing freedom in
proportioning and selection of materials.
• But which comes first, the chicken or the egg?
ACHIEVING THE TRANSFORMATION
• Under a suitable specification (perhaps initially
providing an additional margin) some producers
would very rapidly acquire initial expertise by
engaging consultants.
• The KensQC program could be used to enable both
specifiers and producers to examine current
performance at little cost prior to change.
• A big question is QC field and lab staff and facilities.
It is desirable that the concrete producer should at
least be the employer of the field staff and perhaps
establish labs– does India have the equivalent of the
Australian NATA to assess their performance?