ارزیابی آسان از شاخص دوام برای پیش بینی عمر خدمات و یا کنترل کیفیت بتن با حجم بالایی از مواد مکمل سیمانی

This paper investigates whether durability indicators (DIs), more specifically transport properties, can be assessed by simple methods, e.g. direct experimental methods or indirect methods based on analytical formulas, for every type of concrete. First the results of electrical resistivity and apparent chloride diffusion coefficient obtained by direct measurement on a broad range of materials, particularly on high-volume supplementary cementitious materials (SCM) mixtures, are discussed. Then, various methods, in particular methods based on these last parameters, are compared for the assessment of effective chloride diffusion coefficient and “intrinsic” liquid water permeability, including for the latter a sophisticated method based on numerical inverse analysis. The good agreement observed between the various methods points out that simple methods can allow DI assessment with sufficient accuracy. Moreover, the available values of electrical resistivity, effective/apparent chloride diffusion coefficients and “intrinsic” liquid water permeability can be included in a database. Throughout the paper, the specificities of high-volume SCM mixtures are highlighted.

In recent years, sustainable development has become a major concern, in particular in the construction field [1] and [2]. In this context, a global approach is needed, in order to meet technical, economical, environmental and societal requirements in an optimized way for the whole life cycle of a concrete structure [3] and [4]. The task of designers, engineers and infrastructure owners is therefore now more complex. In particular, they need to combine improved durability and environmentally friendly materials and structures. From a (concrete) material point of view, they are thus interested, on the one hand in relevant parameters, which can characterize durability, and on the other hand in the use of wastes, by-products or recycled materials, which are, at least at present time, regarded as zero-CO2 emission constituents. As a consequence and in order to make this task easier, there is an increasing demand to include in current concrete or design standards (e.g. EN 206-1 [5]) the advanced concepts of durability and service life (SL) prediction, including performance-based and/or probabilistic approaches (see e.g. [6]), in particular with respect to the prevention of steel corrosion in reinforced concrete (RC) structures. In this context, a general approach based on so-called durability indicators (DIs), which are key material properties with regard to durability (e.g. porosity, permeability or diffusion coefficient), has been developed [3], [7], [8] and [9]. Note that other similar concepts can be found in the literature [10], [11], [12] and [13]. A system of classes of “potential” durability with respect to reinforcement corrosion has been proposed for each DI. These five classes – very low (VL), low (L), medium (M), high (H) and very high (VH) “potential” durability – can be used for example for mixture comparison or quality control. The evaluation of the “potential” durability of a given RC will consist of comparing the values of the measured DIs to the threshold values of the associated classes. Another purpose of this approach is to design concrete mixtures capable of protecting structures against degradation for given target SL and environmental conditions, using performance-based criteria (specifications) related to the DIs. Furthermore, a multi-level modelling concept has been developed for SL prediction [8] and [14]. It can be applied to predict the SL of a new structure at the design stage or the “residual” lifetime of an existing and possibly deteriorated structure. Since concrete composition, which is often lacking for existing structures, is not needed, this approach can be very easily applied to them, in view of monitoring, diagnosis, maintenance and support to serviceability extension or repair decisions. Note that this is the same DI set, which is involved in the “potential” durability classes, in the specifications and as input data for the models of SL prediction [7], [9], [14] and [15]. Hence, it is important to: • investigate whether DIs can be assessed by simple and rapid methods, in order to easily include them in a more general set of (sustainability) indicators and in harmonized standard methodologies, • focus on “green” materials, which incorporate local aggregates or high volumes of supplementary cementitious materials (SCMs) such as fly ash (FA) or ground granulated blast furnace slag (GGBS). With regard to the second item, it is well known that a complete characterization of SCMs can be very complex [16]. In addition, some properties of one FA for example cannot be generalized to all FAs, since the properties of a given type of SCM can be very variable (e.g. effects of the specific surface area, alkalinity and glass content). However, it is possible to benefit from the numerous researches carried out worldwide over a long period in particular in North America on these materials (see e.g. [17], [18], [19] and [20]). For example the effect of pozzolanic constituents on pore structure and thus on transport properties has been clearly described in the literature. It is not only a filler effect, but also the result of chemical reactions. The presence of fine particles can also induce acceleration of hydration reactions of the cement (potential nucleation sites for Ca(OH)2 and C–S–H precipitation) and therefore induce earlier densification of the microstructure [21]. These effects are more significant with silica fume (SF), which has a 100-time smaller size than FA. Ca(OH)2 crystals are consumed by the pozzolanic reaction, while finely divided C–S–H hydrate gel is formed, thus yielding a denser microstructure. In addition, when the FA content increases, fibril-type C–S–H are progressively replaced by foil-type C–S–H, which are more efficient to fill capillary pores [22]. Moreover, additional C–S–H (or other gel-type hydrates) form mainly far from the initial cement grains covered by pseudoform C–S–H [23]. These additional C–S–H thus create solid “islands”, between partially reacted grains or pre-existing hydrate clusters, which increase the pore network tortuosity. However, this physical effect induced by chemical reactions will be efficient only once the pozzolanic reaction has significantly progressed, which means in the case of FA long after hydration (or pozzolanic reaction with SF [21]) has started (e.g. several months [3] and [24]). In addition, this effect depends on the initially available Ca(OH)2 amount. Yet, Ca(OH)2 is formed by portland cement hydration and can be affected by early-age drying or carbonation [3] and [24]. Further, SCMs are known to change the concentration and the mobility of the ions in the pore solution (e.g. as a result of modifications of the electrical double-layer at the solid–liquid interface [23], [25], [26] and [27]). For example, the presence of SF induces a significant decrease in alkali and hydroxyl ion concentrations in the pore solution [28] and [29], and according to [30] when 30% of the cement is replaced by FA the hydroxyl concentration is also reduced (see also [23]). This paper will focus on the assessment of DIs, more specifically of the transport properties effective chloride diffusion coefficient and “intrinsic” liquid water permeability, by various methods, on a broad range of materials including mortars or concretes with FA, as well as CEM-III concretes. The range of SCM contents has been selected to be relevant from a practical point of view or to be that commonly used in concretes. The purpose is to validate the use of simple and rapid methods (e.g. direct experimental methods or indirect methods based on analytical formulas) and to check their applicability particularly to high-volume SCM mixtures. Comparison with a more theoretical and sophisticated method based on numerical inverse analysis will be carried out in the case of the “intrinsic” liquid water permeability. Another aim is to contribute to the constitution of a database, at least for the concretes most likely to be used in practice, for apparent/effective chloride diffusion coefficients, “intrinsic” liquid water permeability and electrical resistivity. This will provide in particular an estimate of the values expected for DIs and their corresponding classes, within the framework of the associated performance-based approach. This aspect is very important for future recommendations/standardisation on this topic. Moreover, the specificities of high-volume SCM mixtures will be pointed out.

Different methods for assessing, on the one hand the effective chloride diffusion coefficient Deff, and on the other hand the “intrinsic” liquid water permeability kl, of saturated materials have been compared in this paper on a set of water-cured materials ranging from low-grade to high-performance concretes. The very good agreement observed in each case points out the validity and the reliability of the proposed methods. More precisely, Deff can be assessed from resistivity measurement, ss migration test, nss migration test and nss diffusion test (and numerical inverse analysis [48] and [65]). Likewise, kl can be assessed from the Katz–Thompson relationship, as well as from numerical inverse analysis from drying kinetics. The application of the “resistivity method” is a little more complicated for high-porosity plain OPC – CEM I concretes. Difficulties can also be exhibited in nss migration (or diffusion) tests with dense or high-volume GGBS concretes. Nevertheless, simple and rapid test methods can allow assessment of durability indicators with sufficient accuracy in the general case. An advanced statistically-based investigation of the precision of the lab tests for a broad range of materials remains to be done, as well as direct measurement of kl, in order to finalize the evaluation of the methods. The specificities of the behaviour of SCM-mixtures have been pointed out. As expected, good durability-related properties have been pointed out for mature concretes made with CEM III/A (62% GGBS) or with FA (20–35%). The transport properties (apparent and effective chloride diffusion coefficients, as well as “intrinsic” liquid water permeability) of such normal-strength concretes can be close to that of HPCs with SF, even if the porosity is not very low. Nevertheless, the results have also enhanced the more marked and determining ageing effect associated with these materials (in particular with FA), and this is emphasized in the case of kl. As a matter of fact, significantly better properties are recorded at later ages, after water curing, within the range investigated here, and hence very different DI values can be measured, depending on the age. For example with the mortars tested here (10–30% FA), even if the FA beneficial effect is exhibited on the apparent chloride diffusion coefficient from 90 days, FA are revealed as beneficial only at 180 days with regard to kl. This confirms the usefulness and the complementary nature of these two DIs to assess the “potential” durability related to reinforcement corrosion. Furthermore, no improvement has been recorded in the apparent and effective chloride diffusion coefficients for the 39%-FA concrete tested here even at 180 days. In field conditions, the beneficial effect of FA or GGBS can be compromised (as a result of early-age drying, early exposure to aggressive species, carbonation, skin effect, …) and will markedly depend on the curing and environmental conditions, since such SCM-materials (in particular FA-materials) are very sensitive to these conditions. This has been illustrated for example within the framework of the BHP 2000 French National Project [33] in a study carried out both in the lab and in various outdoor environments on some of the concretes investigated here. Therefore, it is particularly important to investigate the behaviour both in lab and field conditions in the case of concretes with high volumes of SCM. First it is of importance to perform lab tests at 90 days and even at 180 days, in order to, on the one hand characterize the bulk concrete properties, and on the other hand avoid errors and variability associated with lab tests due to ageing effects. In addition, it is necessary to account for early-age behaviour, which will characterize the “covercrete” of the structure. The methods and the corresponding results on various types of materials presented in this paper can be useful to assist engineers in selecting a reliable methodology to assess DIs and carry out durability assessment or prediction. In particular, the data displayed in the paper can provide helpful references when testing new materials or searching to meet durability criteria. The values can also be used as input data in predictive models for similar mixtures as those tested in this study, thus allowing one to avoid long lab tests before implementation of the model. A performance-based approach using a set of relevant and complementary DIs, along with the simple methods discussed here to assess these DIs, can constitute a useful tool to address the durability of new materials, to quantify the actual benefit of SCMs on medium-term and long-term durability, and more generally to perform concrete mix-design for a predefined durability. This will help in the development of new (high-tech) and green solutions within the framework of sustainable development.