The purpose of this research was to determine if the rheological properties of cement pastes containing mineral admixtures could be used to predict the rheological properties of the equivalent concrete mixtures. In cement paste, several mineral admixtures were examined and it was determined that mixtures with ultra-fine fly ash (UFFA) represented the best rheological improvement while silica fume (SF) represented the worst. It was shown that the replacement of cement by SF results in an increase in the water demand and HRWR dosage to maintain the rheological properties of the control. In contrast, the replacement of cement by UFFA resulted in a reduction of the water demand and HRWR dosage to maintain the same rheological properties of the control. The other mineral admixtures, MK, CFA, FA and FFA, gave results in between the SF and UFFA and therefore were not as extensively studied as SF and UFFA.
The concrete tests, due to the amount of materials and labor needed, were limited to SF and UFFA and also to selected dosages of HRWR and mineral admixtures. The results obtained showed that the replacement of cement by UFFA up to 12 % by mass of cement reduces the HRWR dosage and w/c needed to obtain the same slump as in the control. Additional concrete tests were not possible due to the time and labor needed.
Another goal of this research was to determine the optimum replacement level of cement (dosage) by UFFA as defined by the highest reduction in yield stress or plastic viscosity (higher slump). The dosage of UFFA also depended on the w/c and the HRWR dosage. Using the cement paste measurements it could be concluded that the optimum dosage of UFFA is 12 % for replacement of cement by mass (Figure 7). This result needed to be validated with the concrete data. Unfortunately, the concrete measurements were not done at a constant dosage of HRWR as were the cement paste. Therefore, it is harder to determine if the best dosage of UFFA (Figure 9) should be attributed to the increased HRWR dosage (compared to 16 %) or entirely to the UFFA optimum dosage. These concrete results further show the value of the cement paste tests: it is easier and less labor intense to prepare extra mixes to examine the influence of various factors.
It can be concluded that cement paste measurements of rheological properties to screen the dosage and type of mineral admixture to be used in concrete is a promising approach, provided the tests are conducted following the methodology developed at NIST. This implies that the cement paste should be mixed using a blender and not an Hobart mixer and that the measurements of yield stress and viscosity should be performed using a fluid rheometer.
Less sophisticated tests, such as those using the mini-slump and the Marsh cone, cannot be relied upon. The mini-slump test results correlate in certain cases with the yield stress but there is a wide scatter of the data. However, no correlation was observed between the time of flow in the Marsh cone versus the plastic viscosity.
Further testing is needed to confirm that the cement paste rheological measurements can be used as a standard test to predict concrete rheological parameters of a wide variety of compositions.
Figure 9: Influence of partial replacement of cement by UFFA dosage on concrete slump. These are unique tests, therefore no error can be estimated.