Point of View

These columns of ICJ offer an opportunity to the engineering

fraternity to express their views on the current practices in
design, construction and management being followed in the
industry.
To share your opinion with our readers, you may send in your
inputs in about 1500 words via e-mail to editor@icjonline.com

Moisture correction calculations in ACI-recommended concrete mix design

Amarjit Singh

The ACIrecommended method of mix design is

commonly practiced in the United States, and is taught in
university curricula. Though immense effort might have
gone into preparing the method over many decades,
there is an obvious arithmetical error in computing and
recommending the starting point of the moisture to be
added to the mix. The error is on the liberal side, in
that more water is recommended to be added than the
exact calculations warrant. While it is realized that mix
design is an approximate science, the starting point is
based on theoretical science that can be made accurate.
The error has gone unnoticed without remark. This
paper points out where the error is, and proposes a
reasonable remedy.

Introduction

This article observes and illustrates that the calculations

for moisture correction provided in the guidelines for
Portland Cement Associations (PCA) Designing and
Proportioning Normal Concrete Mixtures, as well as those
in the recommendations of ACI 211, are flawed owing
to a simple calculation.1 The PCA process is commonly
used all over the United States and many parts of the
world, and therefore, the observation of an error comes
with some surprise. Nevertheless, the error throws off
the proper use of the proposed tables in ACI 211. In
practical mix delivery, additional water is invariably
added at the batch plant or construction site, over and
above the recommendations of the ACI method, thus
indicating that original mix designs yield stiff mixes.
However, it is evident that the recommended design

tables are based on a partially false premise. But, that

said, it is entirely feasible to make improvements to the
design recommendations.
The Portland Cement Associations (PCA) guidebook
titled Design and Control of Concrete Mixtures, Kosmatka
et al. (2002), is in its 14th edition now, having been in
use for over 75 years.2 The book is used by United
States practitioners as their primary reference, and
by universities as a textbook for their concrete mix
design classes. Consequently, the PCA guidelines are
extremely popular, not to mention they are easy to use
and apply. The PCA method is essentially similar to the
ACI method, with PCA adapting charts and tables of
ACI, such as ACI 318, ACI 211.1 and ACI 211.3. Nawy
(2001), Waddell and Dobrowolski (1993), Derucher et
al. (1998), and Mindess et al. (2003), and many others,
make reference to the ACI and PCA concrete mix design
method.3,4,5,6
However, the author discovers that the calculation
technique recommended for moisture correction in the
various manuals, guidebooks, and other books based
on the ACI procedure is technically flawed. This does
not mean that the whole design methodology should
be scrapped. Only, that the error leads to wetter mixes,
while using a proper calculation leads to stiffer mixes.
Twenty-four student trial tests conducted for various
mix strengths and environmental conditions revealed
this to be true. Nevertheless, it is quite straightforward
to see, as is described in the paper, that the error favors

SEPTEMBER 2010 The Indian Concrete Journal

53

Point of View
more water to be added to the concrete mix. For
practicing engineers and students, this may be just as
well, since workability with stiff mixes is obviously
burdensome and undesirable. Moreover, it is quite
reasonable to assume that the error went unnoticed for
all these decades because the mixes were coming out
wet and workable in the laboratory and batch plant, so
no one complained.
For a calculation flaw of this nature, tests and trials are
not necessary for illustrating the error, since the error is
in the theory. However, trials and tests help to observe
the extent of decrease in slump experienced by following
proper calculations, as was evident from instructor
supervised student trials. This paper explains where the
flaw lies, provides examples, and suggests what can be
done by way of rectification. There is no literature or
precedence on this topic. Perhaps, no one has noticed the
error, or if noticed, has not published. The central part
of this paper is in the section titled Moisture Correction
Calculations PCA Reference.

Charts and tables in concrete mix

design

The charts mentioned in Table 1 are among those

commonly used in mix designing by the ACI/PCA mix
design technology. The essential chart for determining
moisture use is chart no. 5 of Table 1, Approximate Mixing
Water and Target Air Content Requirements for Different
Slumps and Nominal Maximum Sizes of Aggregate.
Table 1. General charts for mix design recommended
by ACI and PCA
Chart
No.

54

Reference table

ACI
reference

Moisture correction calculations PCA

reference
The mistake in the PCA guide observed by the author is
one of simple arithmetic, but of considerable consequence
to how mix design can be logically understood and
performed. Take the example of Kosmatka et al (2002)
of the Portland Cement Association, pp. 163-165, using
the Absolute Volume Method of mix designing. Their
empirical requirements for coarse and fine aggregate,
both ovendry, are 1674 lb (759 kg) and 1236 lb
(560 kg), respectively; the amount of water indicated
for the special conditions is 270 lb (122 kg). These
empirical requirements using the design charts of Table
1 are satisfactory, so far. The moisture conditions of the
aggregates in the sample considered are:
Aggregate type

Specific
gravity

Total
moisture

Moisture
absorption

Moisture
available
for mixing

Coarse aggregate

2.68

2%

0.5%

1.5%

Fine aggregate

2.64

6%

0.7%

5.3%

Now, weight correction for moisture requires that

we batch the following quantities of coarse and fine
aggregate, mentioned correctly by Kosmatka et al. as
Coarse Aggregate (CAgg): 1674 x 1.02

= 1707 lb (774 kg)
Fine Aggregate (FAgg): 1236 x 1.06

= 1310 lb (594 kg)
So far, so good! However, the mistake creeps in at this
junction. If we are to batch 1707 lb (774 kg) of CAgg and
1310 lb (594 kg) of FAgg, the surface moisture that will
be available to the mix will be directly related to the
amounts we batch. Now, 1.5% of the CAgg of 1707lb
(774 kg) is surface moisture, while 0.5% is absorbed by
the aggregate as moisture based on its absorptivity and
is not available for mixing. Similarly, 5.3% of the FAgg
is available for mixing, while 0.7% is absorbed owing
to its absorptivity. It is easy to see that the weight of
the water absorbed and free comes from the total
amount batched at the weight station. Hence, the logical
amounts of mixing water to be excluded from the mix
should be as follows:

Maximum w/c material ratios and minimum ACI 318

design strengths for various exposure
conditions

Requirements for concrete exposed to

sulphates in soil or water

Relationship between w/c material ratio and ACI 211.1

compressive strength of concrete
and 211.3

Bulk volume of coarse aggregate per unit

volume of concrete

ACI 211.1

Approximate mixing water and target air

content requirements for different slumps
and nominal maximum sizes of aggregate

ACI 211.1
and ACI
318

Recommended slumps for various types of

construction

ACI 211.1

Minimum requirements of cementing

materials for concrete used in flatwork

ACI 302

CAgg: 1707 x 0.015 = 25.6 lb (11 kg)

Cementitious materials requirements for

concrete exposed to deicing chemicals

ACI 318

FAgg: 1310 x 0.053 = 69.43 lb (31 kg)

ACI 318

The Indian Concrete Journal SEPTEMBER 2010

Point of View
Therefore, the actual amount of water to be added to
the mix should be
270 25.6 69.43 = 174.97 lb (79 kg)
However, Kosmatka et al. make the error of stating that
the amounts of water to be excluded should be
CAgg: 1674 x 0.015 = 25.11 lb (11 kg)
FAgg: 1236 x 0.053 = 65.51 lb (29 kg)
Resulting in a recommendation for the amount of water
to be added to be
270 (1674 x 0.015) - (1236 x 0.053) = 270 - 25.1 - 65.51
= 179.38 lb (81 kg)
Hence, there is a difference of 4.41 lbs (2 kg) of water
between what Kosmatka et al calculate and what is the
correct calculation. Kosmatkas calculation makes the
mix wetter and more workable. This difference is the
root cause of the difficulties encountered in accurate mix
designing. This is an error of iteration that ignores the
fact that when you increase the amount of CAgg from
1674 lb (759 kg) to 1707 lbs (774 kg), you bring in extra
surface water as a basic component of those additional
33lbs (14 kg). The extra amount of free moisture
available as a result of those additional 33 lbs (14 kg)
is 33 x 0.015 = 0.495 lb (0.224 kg). The same reasoning
goes for FAgg, where we can see that on increasing the
weight of FAgg from 1236 lb (560 kg) to 1310 lb (594 kg),
we add 74 x 0.053 = 3.92 lb (1.77 kg) of free moisture for
mixing. As a consequence of increasing the batched
weights of CAgg and FAgg, we have an additional
4.41 lb (2 kg) of free moisture available for mixing. The
reader will see that this 4.41 lb (2 kg) is directly equal to
the difference of 179.38 lb (81 kg) of water to be added,
calculated by Kosmatka et al, and the 174.97 lb (78 kg)
of water calculated by the author.
The fact of the matter is that the weight measured at
the batch plant is not for ovendry aggregates. If it were
to be so, the recommendations of Kosmatka et al (2002)
could be acceptable. However, were it to be so, concrete
companies would have to go to great lengths to ovendry
their aggregates for mass consumption, an exercise that
would be costly and time consuming. Consequently,
ovendrying is NOT done in practice. Hence, Kosmatka
et al (2002) ignore the practical dimensions of the
problem.

Laboratory tests

Tests performed by groups of students in 24 class

exercises for varying concrete mixes as lab exercises,

gave extremely stiff mixes of 0-0.5 slump for the authorrecommended version of calculations, compared to the
PCA expectation of slump of 1-4 (refer chart no. 6 of
Table 1). The actual average of all those exercises was
0.15, with half the exercises yielding no slump at all.
Upon commensurately adding the extra water (which
would be 4.41 lbs (2 kg) in this example), the slumps
increased considerably enough to make a difference
between a somewhat workable mix and non-workable
mix. Slumps increased by an average amount of 0.6
for the 24 exercises, and only two exercises eventually
yielded 0.25 slump.
This confirms the theoretical underpinnings of the
problem. While the ACI and PCA mix designing
methods recommend multiple iterations for adding
water at the batch plant, and further recommend the
initial amount of water to be added only as an entry
point, this initial amount of water does not satisfy the
practical dimensions that arise from weigh batching the
fine and coarse aggregates.

Moisture correction calculations ACI

reference

This error has been perpetuated by other authors and

researchers. Mindess (2003), referring to the ACI Method
of mix design, pp. 236-239, makes the same assumption
as Kosmatka et al. Basically, the incremental weight of
water owing to the incremental weight of aggregates
is ignored in the same manner. A summary of his
calculations versus what they should be are given in the
following Tables 2 through 4:
Table 2. Aggregate properties
Aggregate type

Specific
gravity

Total
Moisture
moisture absorption

Moisture
available
for mixing

Coarse aggregate

2.68

1.5%

1.0%

0.5%

Fine aggregate

2.63

5.5%

1.3%

4.2%

Table 3. Aggregate weight calculations

Aggre- Weight Weight Weight Incremental Incremental
gate
recomm. SSD, batched,
weight,
free
type
OD,
lb (kg) lb (kg)
lb (kg)
moisture,
lb (kg)
lb (kg)
CAgg

1701
(771)

17183
(779)

17271
(783)

265
(11)

0.136
(0.06)

FAgg

1097
(497)

11114
(503)

11582
(525)

61
(27)

2.56
(1.16)

Wt. of Water = 305 lb (138 kg); OD = Ovendry; SSD = Saturated Surface Dry
1
= 1701 (OD) x 1.015; = 1718 (SSD) x 1.005
2
= 1097 (OD) x 1.055 = 1157.3 ~ 1158; = 1111 (SSD) x 1.042 = 1157.6 ~ 1158
3
4
= 1701 (OD) x 1.01
= 1097 (OD) x 1.013 = 1111.2 ~ 1111
5
6
= 1727 1701 = 26
= 26 * 0.005 = 0.13 lb

SEPTEMBER 2010 The Indian Concrete Journal

55

Point of View
Table 4. Moisture deduction approach
Moisture deduction, lb (kg)

Final weight of water added, lb (kg)

Mindess (2003)

1718 x 0.005 + 1111 x 0.042 = 55.25 (25)

305 55.25 = 249.75 ~ 250 (113)

Author

1727 x 0.005 + 1158 x 0.042 = 57.27 (26)

305 57.27 = 247.7 (112)

Thus, it is seen that there is a difference of over 2 lb

(0.907 kg) in the amount of extra water to be added,
which equals the total incremental free moisture found
in Table 3 (excluding multiple rounding errors). The one
difference between Kosmatka et al (2002) and Mindess et
al (2003) is that the former takes the ovendry weights as
the basis for making moisture correction values, while
the latter takes saturated surface dry weights as the basis
for making moisture correction values. In that regard,
the cumulative error of Mindess et al is less than that of
Kosmatka et al. However, they both fail to account for
all the additional incremental moisture that comes as a
result of the incremental weight increase for batching.

Moisture correction calculations No

slump concrete, ACI 211.3R-97

ACI guidelines for the design of no-slump concrete,

ACI 211.3R-97, makes the same mistake in moisture
correction (Guide, 1997). On page 7 of their guide,
they provide an example: The dry weights calculated for
coarse and fine aggregate, and their moisture properties
are provided in Table 5.
Table 5. Weight requirements and moisture properties
of aggregates
Aggregate
type

Dry weight

Total
Moisture
moisture absorption

Moisture
available
for mixing

Coarse
aggregate

2403 lb
(1089 kg)

1.0%

0.5%

0.5%

Fine aggregate

914 lb
(414 kg)

5.0%

0.7%

4.3%

Recommended water content = 243 lb (110 kg)

On this basis, the amount batched should be as follows,

and ACI and author are in agreement with this:
CAgg: 2403 x 1.01 = 2427.03 lb (1100 kg)
FAgg: 914 x 1.05 = 959.7 lb (435 kg)

56

The Indian Concrete Journal SEPTEMBER 2010

Of these weights, 0.5% is free moisture for CAgg, and

4.3% is free moisture for FAgg. The real calculations,
then, should be as follows:
Free moisture from CAgg = 2427.03 x 0.005

= 12.13 lb (5.5 kg)
Free moisture from FAgg = 959.7 x 0.043

= 41.26 lb (18 kg)
Total free moisture = 53.39 lb (24 kg)
Therefore, water to be added = 243-53.39

= 189.61 lb (86 kg)
However, ACI 211.3R-97 gives the following calculations
for the total free moisture:
Free moisture from CAgg = 2403 x 0.005
= 12.01 lb (5.44 kg) [~ 12 lb (5 kg) as used by ACI]
Free moisture from FAgg = 914 x 0.043
= 39.30 lb (17 kg) [~ 39 lb (17 kg) as used by ACI]
Total free moisture
= 51.31 lb (23 kg) [~ 51 lb (23 kg) as used by ACI]
Therefore, water to be added
= 243 - 51.31 = 191.69 lb (87 kg)
Hence, the ACI calculations recommend that 2.08 lb
(0.9 kg) of more water should be added, in contrast to
what the calculations actually indicate. Considering that
their recommendation is for no-slump concrete, they are
not recommending a mix relatively dry enough.

Observations and discussion

Though mix design is an approximate science one that

is engineered rather than follow a perfect technique it
is nevertheless better to base ones charts and tables
on calculations as accurate as possible rather than
have to increase the level of approximation and trial
batches necessary. We observe that the PCA-method of
moisture correction exemplified by Kosmatka et al; the
ACI-method of mix design exemplified by Mindess et
al; and ACI 211.3R-1997 for no-slump concrete are all
partially flawed on the matter of moisture correction.
This flaw arises from a defect in ACI 211 itself, which
fails to give due recognition in free moisture correction

Point of View
to the incremental aggregates added during design for
the moisture carried by the aggregates. While this error
might be small, it does not excuse the error, especially
given that small differences in moisture added have
a significant effect on slump. PCA has apparently
perpetuated the error of ACI.
Such an error may possibly be managed and contained
on industrial sites, since mix designing is an approximate
science, anyway. However, such a difference can be
troublesome to researchers hunting for accurate answers,
and to students, who, while learning concrete mix
designing, expect logical rigor.

Analogy

A simple analogy should suffice to explain the reason

behind the flaw: a sponge having weight z0, for instance,
can hold n% of its own weight in water. A same type
of sponge of size 1.1 x z0 will still hold n% of its weight
in water, thereby holding 1.1 times more water than
sponge of size z0. So it is with porous aggregates. The
arithmetic and rationale is as simple as this probably
high school level but ACI and PCA and thousands
of engineers over many decades in USA seem to have
either missed this point, or thought the flaw to be of no
consequence, or overlooked it deliberately in order to
not make any waves.

Query and discussion

A general query and discussion on this matter was

conducted between Steven Kosmatka, the author of the
PCA book on concrete mix designing, and the author.
In reference to the authors suggestion to him that the
moisture correction should be made in relation to the
total amount batched, Steven Kosmatka (2003) replied
that the moisture correction should be made in relation
to the dry weights. His reply was
The book has the correct calculation 270 - (1674
x 0.015) - (1236 x 0.053) = 179. Always make the
moisture adjustments relative to the dry weights. If
not, the base value continually fluctuates.
There is nothing wrong in making moisture adjustments
relative to the dry weights, provided the commensurate
moisture adjustments are also made. The commensurate
moisture adjustments, regrettably, are not made in the
PCA book. Moreover, in practical batching, oven dry
aggregates are NOT made available.
In contrast, Mindess at al make their adjustment relative
to the SSD weights. The answer for making adjustments
based on OD weights and SSD weights should be the

same. However, the main issue is whether the moisture

correction is based on the quantity you weigh at the
batch station to put into the mix, or whether you base
the moisture correction on a prior number. Hard reality
says that you go by the facts of what you are actually
batching. Moreover, for scientific rigor, there is no
fear of having the base value of the ovendry weight or
base value of moisture correction to fluctuate, since the
main issue is the surface moisture present in batched
aggregate.
Again, Kosmatka (2003) writes correctly that if one wants a correct answer for the amount of
moisture to be added, one must consider how much
one is batching, consider the moisture content in the
batched amounts, and make corrections accordingly.
The rationale is that one must subtract the moisture
contained in the aggregates we are eventually
batching.
But then makes the mistake of further claiming that
This is exactly what the book calculation does.
Thus, the PCA book misinforms the users, because the
book does not do what PCA proposes. This oversight
has perpetuated for 75 years.
For the sake of expediency and convenience, many a
technique can be justified, for which basic ratio mix
designs have often come in handy. However, the author
does not believe that the emphasis of the profession
is toward rough-and-ready techniques just to make
things work, but that there is an emphasis on steadily
increasing exactness whenever possible. The elaborate
mix design method of ACI, and the enormous effort put
into its development, is proof that the profession seeks
accuracy and reliability.

Making revisions to design tables:

Suggestions

What is the remedy for this problem? If we want to retain

the integrity of arithmetic, the science of Archimedes,
and simple science, then it is incumbent upon us to
make some changes in the mix design charts of ACI.
Since the main parameter being affected through the
moisture correction calculations is the total amount of
water to be added, perhaps some changes should be
brought about to chart no. 5 of Table 1, Approximate
Mixing Water and Target Air Content Requirements for
Different Slumps and Nominal Maximum Sizes of Aggregate.
Since the correct moisture correction calculations will

SEPTEMBER 2010 The Indian Concrete Journal

57

Point of View
demand that less water be added, the easiest thing to
do if we want to retain the total set of the mix design
charts and techniques is to increase the recommended
water content in above chart.

(Note: Original calculations are all in lbs. Conversions

from lb to kg have generally been rounded down to the
whole numbers.)

How much should the water recommendations be

increased by? Considering that the water amounts in
the above chart are in round numbers in multiples of
five, and since the correction amount we are talking
about is small to the magnitude of four or five pounds,
depending on the conditions and circumstances -- it
may be reasonable, as an empirical recommendation,
to increase all water content values in above chart by
five pounds each.

1. ______Standard practice for selecting proportions for normal, heavy weight and mass
concrete (Reaproved 2009), ACI 211, 211.1-91, 1991-2009, American Concrete
Institute, Farmington Hills, MI, USA.

Further studies

It might be more effective to conduct a study to see if

there are more precise answers for the amount by which
the water quantities in above chart should be increased.
Such a study could consider various scenarios and
conditions, and different types of aggregates having
different moisture content and absorption. But, further
discussion on this is beyond the scope of this paper.

Conclusions

The observations and arguments reveal that the

calculations for moisture correction in the ACI and PCA
recommendations, are unfortunately mistaken, albeit by
a small amount. Using corrected calculations results in
stiff mixes, considerably more than designed for. Using
the erroneous calculations pose an ethical dilemma
to the conscientious engineer, since they compromise
standards in basic physics and mathematics.
The conflict is between continuing to use the ACI
and PCA-recommended mix design along with the
recommended tables and charts for mix design that are
based on a false premise, or to use the proper calculations
and revise the tables and charts. Said another way, shall
we continue as normal and bury our face in the sand and
insist we dont see anything wrong, or shall we accept
the error for the betterment and enhancement of our
profession, and thus take steps to improve our manuals
and guides. For the author, the only ethical choice is the
latter. The major and simplest improvement would
be made to the recommended water content of the ACI
211.1 chart titled Approximate Mixing Water and Target
Air Content Requirements for Different Slumps and Nominal
Maximum Sizes of Aggregate. Correspondingly, the
recommended method for moisture correction should
be rectified to make the moisture correction based on
the actual batched amount of aggregates.

58

The Indian Concrete Journal SEPTEMBER 2010

References

2. Kosmatka, S.H. and Kerkhoff, B., and Panarese, W.C., Design and Control of
Concrete Mixtures, 14th ed., Portland Cement Association, IL, 2002.
nd
3. Nawy, E. G., Fundamentals of High-Performance Concrete, 2 ed., John Wiley,
NY, 2001.

4. Waddell, J.J. and Dobrowolski, J.A., Concrete Construction Handbook, McGrawHill, Inc., NY, 1993.
5. Derucher, K.N. and Korfiatis, G.P., Materials for Civil & Highway Engineers,
4th ed., Prentice Hall, NJ, 1998.
nd
6. Mindess, S., Young, J.F., and Darwin, D., Concrete, 2 ed., Prentice Hall,
NJ, 2003.

7. ______Guide for selecting proportions for no-slump concrete, ACI 211.3R-97,

ACI Committee 211, 1997, American Concrete Institute, Farmington Hills,
MI, USA.
8. Kosmatka, S., E-mail communication with author, June 10, 2003.

Prof. Amarjit Singh PhD is professor of civil

engineering at the University of Hawaii at Manoa.
He earned his bachelors degree from the Indian
Institute of Technology, Delhi, and PhD from
Purdue University. He is a licensed engineer in
Canada, and a chartered engineer in U.K. He
served on sabbatical at the University of New
South Wales, Australia and Cornell University, New York. He
has worked in the construction industries of India, Nepal,
Kuwait, and Canada. He was Chair of the Hawaii Council
of Engineering Societies, and is currently the President of
the International Structural Engineering and Construction
Society. His research interests are in concrete construction,
cost engineering, future engineering systems, and asset
management.

What is your opinion?

Do you wish to share your
thoughts/views regarding the
prevalent construction practices
in the construction industry with
our readers?
If yes, The Indian Concrete Journal
gives a chance to the engineering
fraternity to express their views in
its columns.
These shall be reviewed by a
panel of experts. Your views could
be limited to about 2000 words supplemented with good
photographs and neat line drawings. Send them across by e-mail
to info@icjonline.com.