Page 6 - February-2022

TECHNICAL PAPER

With a known solid volume of particulate material (V p ), the bulk 3. RESULTS AND DISCUSSION
volume (V b ) at the capillary state is the summation of water
volume at maximum power consumption and solid volume of 3.1 Output material combinations
particulate material. Therefore, packing density, α (the solid
volume fraction at maximum power consumption), is determined Calibration was achieved using preliminary experimental data
according to Equation 1. from the mixing energy test, and entailed the analysis of packing
density results of cement blends with limestone of varying
V p (1)
α = fineness (Figure 2). Each packing density result is the average
V b
of at least two mixing energy tests. Detailed considerations for
Initially, the packing density of plain cement and ground the calibration of the CIPM are given in . The chosen model
[6]
limestone was determined, and thereafter various combinations constants (K=12.2 for mixing energy test, C a =13.5, C b = 1.0, and
of cement with various ground limestones were tested. The d c = 10 µm) enabled minimization of the average prediction error
minimization of the difference between the predicted packing
density and the experimental packing density was used as to 1 % and a maximum prediction error of 2.8 %.
the basis for selecting model constants. Once calibrated, the Figure 2 reveals that KB2 and KB45 blends offer the most
integrated use of the packing models guided the choice of significant increase in packing density, and these were therefore
material quantities for the optimized concrete mixes. of interest for constructing optimized concrete mixes. The
composition of the powder phases of concrete mixes was
2.3 Materials and concrete mix design constructed to enable maximum packing density from the
CIPM. Remaining material quantities were selected to enable
2.3.1 Materials the best fit of the overall mixture grading curve to the MAAC

2
CEM II A-L 52.5 N (CEM II) with Blaine fineness of 2 700 cm /g for the given constituents. Figure 3 portrays the PSD for Phase
was used as a reference cement and blended with two limestone 1 mixes with minimum particle size D min = 0.42 µm, and Figure
fillers of varying fineness. These were Kulubrite 2 (KB2) and 4 portrays Phase 1 mixes with D min = 0.36 µm. Figure 5 portrays
Kulubrite 45 (KB45) with relative finenesses of 13 100 cm /g and Phase 2 mixes with D min equal to 0.36 µm. D max was 9.5 mm for all
2
600 cm /g, respectively. The use of materials finer (KB2) and phases (Phase 1 and 2 mixes). The ideal MAAC grading curve is
2
coarser (KB45) than cement reportedly has the most significant presented for comparison in each instance.
potential for increasing powder packing density and binder
efficiency . Additionally, two more limestone fillers were used 3.1.1 Phase 1 mixes
[9]
2
in the calibration procedure, namely, KB5 (7 100 cm /g) and
KB10 (4 800 cm /g). Fine aggregates were a Philippi dune The reference mixture (portrayed as Mix 1-1 in all figures) was
2
sand and a granite crusher sand, and coarse aggregates were designed according to the locally applied C & CI ‘Method
nominal 9.5 mm crushed granite. The only additive used was of Mix Design’ [12] , and although the method is not aimed at
MasterGlenium ACE 456, a poly-carboxylate superplasticizer. achieving maximum packing density or at matching an ideal
grading curve, the overall constituent grading curve reasonably
2.3.2 Concrete mix design matches the MAAC grading curve, represented by a R-squared

Two concrete mix design phases were used. Water content statistic of 0.98. Incorporating parameters of bulk density and
was held constant for Phase 1 mixes, which consisted of a fineness modulus when determining aggregate quantities
reference mix with 100 % CEM II, and five mixes with increasing imply the filling of void space between coarse aggregates with
limestone replacement, with a water/powder (w/p) ratio of 0.5. fine aggregates, responsible for a high R-squared statistic. The
Phase 2 mixes had reduced water content and reduced w/p and value of the R-squared statistic increased to 0.99 for Mixes 1-2,
were designed with constant powder-paste (material <125 µm 1-3 and 1-5, inferring increased packing density. Despite small
+ water) volume. Phase 2 consisted of a mix with a CEM II/ discrepancies in fine and coarse aggregate quantities between
limestone blend of 45/55 vol. % and a mix with a CEM II/fly ash/ these mixtures, their PSDs tended to the same function. As
limestone blend of 45/20/35 vol. % respectively. Table 1 gives limestone content increased, the mixture PSD tended more
the mix designs and mix ratios. It also includes approximate closely to the ideal MAAC, also inferring increased packing
values for the CO 2eq of the various mixes (in kg/m ). density. Incorporating KB2, Mixes 1-4 had a slightly improved
3
packing density relative to the reference mix (R-squared of
2.4 Concrete tests
0.99 relative to 0.98), but Mix 1-6 had an equivalent R-squared
Slump [10] and compressive strength [11] were tested for all mixes, statistic to the reference mix. This implied the overall mix
and additionally, durability index and accelerated shrinkage tests (powders through coarse aggregates) of Mix 1-6 did not have an
were conducted for Phase 1 mixes. However, this paper only improved packing density, despite the powder phases having
reports results for slump and compressive strength tests. increased packing density relative to plain CEM II.

THE INDIAN CONCRETE JOURNAL | FEBRUARY 2022 9