To determine the validity of the methodology adopted, the following tests were performed:
To determine that the controlled mixer was necessary and that no other cheaper method could be used, some tests were made using an adjustable voltage controller to set the blade speed. This simple system does not allow a feedback loop to keep the speed constant if the load from the cement paste changes from mix to mix. Six batches were preparesd of an identical cement paste design and tested them according to the procedure described above. Three specimens were mixed using the controlled voltage mixer and three using the controlled speed mixer. The cement paste had a w/c of 0.4 and no admixtures were used. It is clear from Figure 3 that the controlled speed mixer gives a better reproducibility of the data.
To determine the influence of the gap and of the dosage of HRWRA on the yield stress and viscosity, six batches of cement paste at a constant w/c of 0.36 and three dosages of a melamine based HRWRA were prepared. The dosages were 0, 1%, and 2% by mass of the cement. The cement was of type I and its complete characteristics are reported in Ref. [5]. Two gaps were selected: 0.16 mm and 0.26 mm. A full factorial design using the dosages and the gaps was used as described above. All measurements were performed twice using a new batch of cement paste each time. The yield stress and the plastic viscosity were calculated according to a Bingham model. Figure 4 and Figure 5 show, respectively, the yield stress and the plastic viscosity. The error bars in the figures represent one standard deviation.
The yield stress shows, as expected, a dependence on the dosage of the HRWRA. With the higher dosages giving lower yield stresses. The plastic viscosity is also reduced by the addition of the HRWRA but a higher dosage does not seem to significantly reduce it. These statements need to be verified further because the standard deviation is relatively high in some of the data.
Figure 3: Torque vs. shear rate using two methods for controlling the speed of the blades during mixing. A) Adjustable voltage controlled mixer with no feedback; B) Controlled speed mixer with feedback. The error in the measurement of the torque is about 10 %.
Figure 4: Yield stress vs. gap
for various dosages of HRWRA. The error bars represent one
.
Figure 5: Plastic Viscosity vs. gap for various
dosages of HRWRA. The
error bars represent one
.
The influence of the gap showed an interesting result. The yield stress increases with a decreasing gap size. This was expected because HRWRA is designed to reduce the yield stress. In concrete the yield stress is the most commonly measured value because it is related to the slump cone test. On the other hand, the viscosity shows either the opposite influence or no influence at all. The explanation of the strange behavior of the viscosity needs to be further investigated especially because the error (showed in Figure 5 by the bars) is very large for the measurement done at 0.16 mm gap with no HRWRA. The non-influence of the gap on the viscosities for the cement paste containing HRWRA is an interesting discovery. It could imply that if the yield stress is overcome then the paste is not influenced by the proximity of the aggregates, represented here by the gap.