Quality control - Concrete Mix Design, Quality Control and Specification - page 239

dangerously defective concrete does not occur, but if unsatisfactory
concrete is supplied with impunity, dangerously defective concrete is
likely to follow.)
It is obvious that these principles, and the requirement of early detection,
are not met by the current practice of statistically analysing a number of
28-day test results and following up any marginally unsatisfactory result
with coring or nondestructive testing (NDT).
What is required is to examine early-age results and to follow up any
predicted departure from the required mean strength with an immediate
mix adjustment or other effective action to restore the required mean. It
may be good practice to make slightly more than the required increase in
the case of a shortfall and slightly less than the calculated saving in the case
of a higher strength than necessary, that is, to ensure that any shortfall is
definitely immediately remedied and to approach the required mean with
caution. This kind of immediate reaction to observed early-age variation
cannot be imposed by a purchaser or supervising engineer, and must be in
the hands of the producer.
Another question is the extent to which mixes should be deliberately
varied at the time of batching in an attempt to avoid changes in strength.
For example, it is known that a higher concrete temperature at the time of
batching will increase water requirement and therefore reduce strength,
yet almost any set of traditional test results will show a strength reduc-
tion in early summer and a strength increase in early winter. It takes some
traditional control systems more than a month to react to this situation
and adjust mixes. There can be a distinct difference between concrete
temperature on a cold morning and a hot afternoon and Day has written
about concrete on a cold morning being rejected as of excessive slump
when it might in fact have a lower water to cement (w/c) ratio and a
higher eventual strength than lower slump concrete on the same after-
noon. However, the whole question of whether it is ever reasonable to
reject an individual truck of concrete on slump or appearance grounds
needs careful consideration.
There may be genuine reasons for requiring a higher slump (or a longer
delivery time) on some deliveries and rather than having water added on
site the concrete could be officially supplied at a higher slump, probably
increasing the admixture dosage rather than adjusting water and cementi-
tious quantities.
Variation in cement, admixture, or aggregate properties may also be
known in advance of batching in some cases. Of course this possibility
of “just-in-time” mix variation would require close regulation and skilled
staff if it is not to lead to even worse variability. Any such policy should be
introduced on a well-scrutinised trial basis and continued only if shown to
reduce test result variability.