RESULTS AND DISCUSSION

Table 2 shows the data obtained from the two rheometers. It should be noted that the impeller rotation maximum speed used in the IBB test is shown in the table. In some cases (mixes ID 293), the rotational maximum speed was reduced because the torque generated by the resistance of the concrete was too high to be measured, i.e., the impeller will not rotate at a higher speed.

Comparison of the yield stresses from the two rheometers, IBB and BTRHEOM, indicated no correlation (Table 2). Lack of correlation may be due to the range of yield stresses measured, which was in the vicinity of zero and was sometimes negative. This situation is expected since all the concretes tested were highly flowable and therefore should have very small yield stresses. The negative values are due to the method used to calculate the yield stress. Whether the Bingham or the HB equation is used, yield stress is estimated from an extrapolation of the shear rate versus shear stress curve to zero shear rate. The negative values are attributed to the error in the extrapolation process and have no real physical meaning. It could be inferred that another equation other then Bingham equation should be used.

The comparison of the viscosity, measured by the two rheometers, shows a good correlation. Figure 2 shows the plot of the viscosity as measured by the two rheometers. For this comparison the concrete mixture ID 293 was not considered because the IBB impeller maximum rotational speed was different from all the other concrete mixtures. The correlation is relatively good (R² = 84 %) and can be approximated by:

B  = 35 +  16  i

( 1 )

where _Bis the viscosity measured with the BTRHEOM and _i is the viscosity measured with the IBB

This is an acceptable correlation considering the wide range of viscosity covered. This correlation is nevertheless preliminary due to the limited number of data points and the lack of variation in the properties of the materials used, i.e., one type of cement and one type of aggregates. It should be pointed out that only one other attempt to compare two rheometers is known [9]. The results were negative in the previous attempt, i.e., no correlation was found. For this reason, ACI Committee 236A is planning to conduct a series of tests to compare all the existing concrete rheometers. These tests are tentatively scheduled for the summer of 2000.

In the rest of this paper, only the viscosity obtained from the BTRHEOM is used to compare the rheometer results to other tests.

Figure 2: Comparison between the two rheometers: IBB and BTRHEOM

The results obtained for the slump and the slump spread are shown in Table 1. In this test program, the average slump for all mixes was 273 mm + 13 mm (one standard deviation), a variation of only about + 5 %. The average spread for all mixes was 655 mm + 39 mm (one standard deviation). Again, the spread can be considered constant for all the mixes with a variation of 6 %. This is not surprising because the HRWR dosage was adjusted to obtain a slump spread of at least 610 mm. Therefore, on the basis of slump and slump spread, we could conclude that all these concretes have the same workability. However, the results of the U-flow and V-flow and the rheometers clearly show that these concretes do not behave the same in the filling capacity (U-flow height) or in the ease of placement (V-flow time) or the viscosity. The scatter of data in these tests was quite large. The average filling height (Table 1) is 123 mm + 88 mm (one standard deviation). This corresponds to a variation of 71 %. For the V-flow time (Table 1) the average is 50 s +  53 s or a variation of 108 %. The plastic viscosity as measured by the rheometers varies by a factor of 2.4 over the range of the mixes for either rheometer. In conclusion, we can affirm that the slump and the slump spread, by themselves, are not the correct tests for measuring the workability of these types of concretes, because they do not predict the concrete behavior during placement. Hayakawa et al. also reached this conclusion [10]. They showed that for the same slump spread a wide range of filling abilities can be obtained. Therefore, no correlation between the V-flow or U-flow test and the slump or slump spread can be obtained.

The U-flow and the V-flow results were examined to determine a better definition of SCC. If the U-flow filling height criterion (>70 % maximum fill height) is used to detect SCC, there are only three concretes that are SCC in this series, namely the concrete with the ID: 286, 289, and 295. The U-flow values above 200 mm are shown in bold in Table 1. Figure 3 shows the comparison between the plastic viscosity and the U-flow and the V-flow tests. For clarity, the V-flow values of the three SCC mixtures are marked with a cross. The three SCC mixtures have, by definition, a U-flow value above 200 mm (black line). For these concrete mixtures, it seems that the V-flow value needs to be lower than 20 s for a concrete to be SCC. From these few data points, we cannot say that this V-flow value is applicable to all concretes. The data does not provide a correlation between the viscosity and the other two tests. Therefore, the viscosity alone cannot uniquely determine if a concrete is SCC or not. It is also important to note that mixture ID # 286 showed a visual indication of segregation. Therefore, it should be emphasized that U-flow alone may not be an universal single test indicator for SCC and that the filling capacity should be used with care.

Figure 3: U-flow and V-flow time compared with viscosity measured with the BTRHEOM.

According to Beaupré [6], a better method to evaluate concrete with a specified flow property is to plot the yield stress versus the viscosity. Concrete mixtures, determined to have the desired property, define an area in the plot called a "workability box." Figure 4 shows a plot of the viscosity versus the yield stress as measured with the BTRHEOM for this study. The points, marked with a cross in Figure 4, are SCC. A box can be traced to limit an area around these points that does not include other mixtures. The "workability box" defines the range of viscosity and yield stress needed for a SCC. If these results were trial batches, the drawing of the box would allow the operator to determine whether a mixture is SCC based on the rheometer results. As was mentioned above, the yield stresses measured with the two rheometers do not correlate. Nevertheless, a box can be traced on an equivalent graph plotted using the results from the IBB rheometer. As some of the yield stresses are negative in our study, further trials would be necessary to use the "workability box" from Figure 4. It should be reminded that the yield stresses are negative due to the extrapolation determined by the Bingham equation. This interpretation of the results was given just an indication that to define SCC there is a need for more than one rheological parameter, not as universal definition of SCC.