EXPERIMENTAL PLAN

In an on-going experimental plan carried out at the National Institute of Standards and Technology (NIST), an attempt is made to assess the rheological behavior of all feasible concretes within a fixed set of components. In a first series of tests, 33 concretes or mortars without admixtures were generated by systematic variations of cement content, coarse/fine aggregate ratio and water content. In a similarly designed series, the mixtures were highly "superplasticized", with a high-range water reducer agent (HRWRA). Finally, two more series completed the program, one with intermediate amounts of HRWRA, and one with various additions of silica fume. A total of 78 mixes was studied. The concretes were produced with a gravel having a maximum size of 12 mm, a coarse sand, a single-size fine sand from Ottawa (Illinois), and an ordinary portland cement (ASTM Type I/II). The HRWRA used is of the naphthalene type (having a solid content of 40 %). Some of the mixtures are detailed in Table 1.

Within the three series considered herein, the mutual proportions of dry materials remain constant. Only the water concentration was changed from one mixture to another. The mixtures had water-cement ratios in the range 0.55-0.60 by mass ('Normal-Strength Concretes'), 0.36-0.38 ('High-Performance Concretes'), and 0.26-0.27 ('Self-Compacting Concretes'), respectively.

A slump test (ASTM C143 [9]) and a BTRHEOM rheometer test were performed on every mix. All slump values were higher than 80-100 mm, which ensured that the concretes were sufficiently fluid for testing in the rheometer. For the rheometer tests, a 15 s previbration was applied to the specimens (with a frequency of 40 Hz) ² , then the shear tests were carried out without vibration. Five measurements were taken for rotation speeds ranging between 0.2 and 0.8 rev/s. This corresponds to a strain gradient between 0.25 and 6 s-1. For two very thick mixtures (BHP8C & BHP8B), the maximum rotation speed had to be reduced, because the power of the motor was not high enough to reach the intended maximum speed. The results of the tests are given in Table 2 and Figure 2.

As a preliminary result, it appears that experimental relationships between torque and rotation speeds generally deviate from a straight line (see Fig. 2), even if the linear correlation is rather high (between 0.970 and 0.999). Moreover, if a linear regression is applied to the experimental points (following the usual procedure to deduce the Bingham parameters), a negative yield stress is sometimes found (in the case of the Self-Compacting mixtures).

Figure 2. - Relationship between torque and rotation speed in BRTRHEOM rheometer, for some concrete mixtures.

² The previbration ensures that the placement of the concrete sample is not influenced by the operator [5].