= 0
+ µ 
In the following section a brief review of recent developments in the characterization of the rheological properties of concrete is presented. Studies on the rheology of fresh concrete, from a materials science approach, are just beginning to appear. In fact, the concretes that are easier to describe by such an approach, i.e., concretes that are sufficiently fluid, have only been commercially used recently. Previously, it was considered adequate to use an engineering approach for characterizing rheological properties, consisting of subjecting a concrete sample to a more-or-less controlled loading and deriving an index (slump, flow time, compacting density, etc.) to classify the mixtures in terms of workability. This approach is limited as the classifications obtained by using different tests vary substantially [42].
Tattersall was one of the pioneers of concrete rheology when, in 1991, he proposed using an instrumented mixer to obtain a more complete characterization of the flow characteristics of fresh concrete [1]. He proposed describing the behavior of fresh concrete using the Bingham model in the following form:
| = 0
+ µ | (1) |
where is the shear
stress applied to the material (in Pa), is
the shear rate (also called the strain gradient) (in s-1 ),
0 is the yield stress
(in Pa) and µ is the plastic viscosity (in Pa· s).
The last two quantities characterize the flow properties of the material. However,
in Tattersall's apparatus, the velocity field is unknown and complex due to
the lack of symmetry. Therefore, he limited himself to an empirical
description of the material's behavior, by using the relationship between the
torque and the rotation speed of the mixing blades.
The same approach was pursued by Wallevik, who attempted to improve on Tattersall's apparatus by returning to the more classic geometry of a viscometer with coaxial cylinders (BML viscometer [2]). Unfortunately, this instrument's design did not consider the reasons that led to the shape of the blades in Tattersall's apparatus. Tattersall's blades were designed to overcome the natural tendency of concrete to segregate by creating a vertical movement through the tilt of the blades. The movement of the concrete in the BML viscometer is essentially horizontal, and places more than half of the total volume of concrete in "dead" zones, in which the concrete is not sheared during the test. Thus, segregation can occur through the migration of fine particles from the dead zones to the sheared zones, resulting in an underestimation of the rheological characteristics.
After the characteristics and shortcomings of these devices were analyzed [3], the BTRHEOM rheometer was developed [3,4, 8]. Designed on the basis of both scientific and practical requirements, this rheometer has the ability to provide fundamental rheological values, while being portable and usable on the job site.
Hu validated the measurements obtained with the BTRHEOM rheometer - as far as it is possible to do so- by performing finite element calculations and comparative tests with all existing rheometers [7]. It appears that the Bingham parameters are determined with an uncertainty that does not exceed 10%, which is satisfactory for a test of this kind. Once the basic characteristics of the apparatus were established, it remained to put it to use to increase the database of available knowledge on concrete rheology.
1.3 Potential benefits of this study
It was shown that the BTRHEOM rheometer is a valuable tool for monitoring concrete rheological properties [5]. The rheometer makes it possible to show the loss in workability over time for a single sample, and to provide a diagnosis of the causes of these changes, whether linked to absorption of water by the aggregates or to incompatibility between the cement and the HRWRA [6]. However, for the rheological approach to become an integrated part of concrete enginnering, two information gaps have to be bridged (see Figure 1).
Figure 1. Gaps, i.e., links missing, in the knowledge of concrete rheology to be filled in order to make rheology truly usable by the engineer
First, it is necessary to understand how the processing parameters are linked to the rheological behavior of the concrete. For example, it seems that concrete pumpability is controlled by plastic viscosity [7], [8]. Also, it would be interesting to determine if the stability of fresh concrete placed at an angle (or on a slope) is controlled by the yield stress [7]. One could establish requirements for concrete rheological characteristics that would make it possible to empty a concrete bucket within a given time based on finite element calculations of the same type as those used by Tanigawa et al. [9, 10].
Second, efforts must be made to link the rheology of the concrete to its composition. This is one of the objectives of the present study. When this objective is reached, the engineer will be able, at the mixture design stage, to optimize the mixture proportions taking into account placing methods and structure types.