آنالیز آماری و بهینه سازی بتن خود متراکم (SCC) با استفاده از متاکائولین (MK)

ABSTRACT Optimizing the durability and service life of self-consolidating concrete containing metakaolin using statistical analysis

The production of self-consolidating concrete (SCC) is normally achieved by: (a) increasing the quantity of fines in the mixture, which could be done by incorporating one or more supplementary cementitious materials (SCM’s) [1–3]; (b) adding high range water reducer admixtures (HRWRA), and if necessary, viscosity modifying admixtures (VMA) [4–5]; and/or (c) decreasing the coarse aggregate content in the mixture [6–7].

Different SCM’s have been successfully used in the production of SCC, such as fly ash, ground granulated blast furnace slag, volcanic ash, cement kiln dust, rice husk ash, and silica fume [2]. Metakaolin (MK) is another type of SCM that is considered relatively new; it has been widely used for the production of high strength and high performance concrete over the past two decades [8]. In recent years, MK was introduced for the production of SCC. The behavior of MK in SCC mixtures was found to be similar to that in normal concrete mixtures, which showed an enhancement of the overall mechanical and durability performance [5].

Unfortunately, the production of SCC usually warrants a high cost, which is attributed to the high cement contents, high percentage of SCM’s, and/or high doses of HRWRA. For this reason, optimizing SCC mixtures is vitally required to minimize the cost while maintaining the best fresh and hardened properties of the mixture. Statistical design of experiments is a useful tool that can be used to optimize the mixture components of SCC. Additionally, prediction models can be developed to evaluate the response at different levels of the governing factors. By using such models, the numerical optimization can then be performed to minimize or maximize the response [9–10]. Statistical design of experiments has been widely used in the mixture design and optimization for both normal concrete and SCC specially those containing SCM’s [9–10]. Alternative soft computing methods are also applied to develop prediction models for different concrete properties. Examples of these methods are: artificial neural networks [11], genetic algorithm [12], and adaptive neuro-fuzzy inferencing systems [13]. However, most of these models require previous knowledge database about the input and output parameters along with significant mathematical work for the development of prediction relationships [13]. To this date, researchers apply the statistical design of experiments technique to analyze, model, study interaction between parameters, design, and optimize the behavior of SCC mixtures [14]. Durability and service life of concrete structures are considered to be the most important properties that need to be accounted for when optimizing SCC mixtures. Reinforced concrete structures, especially those subjected to high percentages of chlorides, should be given a special design consideration to extend their service life by reducing the corrosion of their embedded reinforcing steel [15–17]. The corrosion occurs when the chlorides from deicing salts, groundwater, or seawater penetrate the concrete cover and reach the reinforcing steel. Once the percentage of the chloride around the steel bar exceeds the threshold needed for corrosion initiation, the corrosion starts and rapidly propagates through the entire steel bar, leading to a mass loss and delamination of the concrete cover [17].

Chloride permeability is a significant property of concrete representing its service life. The assessment of the chloride permeability in concrete is usually performed using one of the following standard tests [18–19]: rapid chloride penetration test (RCPT) (ASTM C1202) and/or chloride bulk diffusion test (ASTM C1556). Most of the models used for predicting the corrosion initiation time account for the resistance of concrete to chlorides [16,20]. Concrete with low permeability and dense microstructure is proved to have more resistance to chloride ingress and longer time for corrosion initiation [17,21].

Recently, different methods were developed for predicting the service life of concrete structures. These methods usually monitor the service life of concrete structures in two periods, including initiation and propagation periods. The initiation period is the period in which the chlorides penetrate the concrete cover until they reach a threshold level (enough to initiate corrosion) at the rebar surface. The propagation period starts after the initiation period and ends when significant damage occurs in the structure resulting in cracks and rebar mass loss. The time of the initiation period is calculated using a simplified Fickian diffusion approach assuming that the chloride diffusion is the dominant mechanism of rebar corrosion. The time of the propagation period, however, depends on the definition of ‘‘significant damage.’’ This level of damage in general varies depending on the requirements of the owner and the nature of the structure.

The main objective of this study was to develop an optimum SCC mixture incorporating metakaolin using a statistical design of experiments approach. The total binder content, percentage of MK, water-to-binder ratio, and curing conditions were varied to obtain the best mixture in terms of chloride permeability (RCPT and chloride diffusion), and extended service life. The statistical design of experiments approach was also used to present the most significant factors affecting each response variable in the mixture and to develop some prediction models for each test result. The study also presented a relationship between the different methods of assessing the chloride permeability in concrete for predicting the chloride diffusion of SCC mixtures containing metakaolin.