Unchanging concepts - Concrete Mix Design, Quality Control and Specification - page 258

There are two important provisos that should be made in advocating
cash penalties and greater emphasis on standard deviation:
1. The standard deviation (and the mean strength, but that is much eas-
ier) must be accurately determined.
2. The cash penalties (which may be described as “liquidated damages”
or “provision for reduced durability” or formatted as a bonus clause
rather than a penalty) should be very moderate, only about twice the
cost of the additional cement that would have avoided any penalty
(i.e., about 10 kg/m 3 per MPa [12 lb/yd 3 per 100 psi] of deficiency so,
in Australia, about a $2 penalty per MPa of deficiency).
The requirement for an accurate SD is easily satisfied under a cash pen-
alty system because it is not necessary to identify which concrete is slightly
understrength, only how much and how defective. Therefore the penalty
can be levied on the concrete represented by 30 consecutive results with
great accuracy.
Does anyone have a convincing counter argument? If not, how long do
you think it would take to implement this proposal? 5, 10, 20 years? It may
be of interest that the outline of this argument was advanced in papers pub-
lished by Day in 1959 and 1961.
11.3 HOW SOON IS SOON ENOUGH?
The first edition contained a 21-page account of an investigation using a
massive computer analysis of synthetically generated data to clearly estab-
lish the superiority of cusum analysis over any other system known for the
early detection. That section of the first edition is not repeated here but is
available on the website.
The two most significant points arising were as follows:
• No computer analysis is as efficient and reliable as the eye examina-
tion of a cusum graph in detecting a small change in mean strength.
• The mathematical significance of a downturn in a single variable (i.e.,
strength) is in any case immaterial when the significance of the down-
turn is correlated to simultaneous changes in other variables such as
slump, temperature, and density.
The economic value of a more efficient analysis system is briefly com-
pared to that of other factors affecting the attainment of the desired con-
crete quality, such as better equipment, more skillful personnel and higher
testing frequency. It is pointed out that a more efficient detection system
is equivalent to a higher testing frequency in achieving early detection.