Troubleshooting - Concrete Mix Design, Quality Control and Specification - page 272

the substrate or within the element. It is important to know the likely
cause of cracking: internal or external restraint.
Autogenous shrinkage—Autogenous shrinkage should not be forgotten
as a cause of early cracking in low water to cementitious materials
(w/cm) ratio mixes. Hydration removes free water in the concrete
causing tensile forces to develop within the capillary network. Unlike
drying shrinkage, it can occur in spite of any measures taken to
reduce or prevent evaporation. It is also a through section shrinkage
that can exacerbate thermal shrinkage problems. Autogenous shrink-
age can be a particular concern with low w/cm ground-granulated
blast-furnace slag (GGBS) concrete.
As autogenous shrinkage is due to tensile stresses developing in the
capillaries, which requires the formation of a meniscus. Therefore
ponding with water during early curing period is a good way to limit
autogenous shrinkage, at least in the surface layer.
12.1.4 Excessive variability
The first thing is to establish whether the variability is in the concrete or in
the testing. Two places to look are the average pair difference in the 28-day
results and the range of densities of test specimens from the same sample
of concrete. The average pair difference should desirably be below 1.0 MPa
and densities should not have an average range exceeding 50 kg/m 3 .
However, calculated densities may vary through inaccurate measurement
of specimens rather than variable compaction or segregation, and this
would have no effect on strength variability.
A second place to look is at multivariable cusum graphs of strength and
other variables. If slope change points in strength correlate with those of
other variables, the cause will be clear. Direct plots of multiple variables
will show whether individual high or low results have an explanation.
If there is no explanation, and especially if 7- and 28-day results do not
correlate, testing would be suspect.
Having established that the variability is actually in the concrete and
not just the testing, batch quantity records should be available if batch-
ing is by a computer-operated plant. It should not be overlooked that the
correct quantities may be weighed out but may be insufficiently mixed to
give uniformity. There have also been examples of short central mixing
times (prior to further mixing by agitator trucks) that have not permit-
ted time for all the metered admixture to enter the mixer. Similarly part
of a particularly critical ingredient such as silica fume may hang up
in the batching skip from time to time and finish up in the next load.
Alternatively silica fume may not be properly dispersed greatly reducing
its effectiveness.