1. Introduction
Civil engineering structures are the most expensive investments in a country, therefore, special considerations are needed to preserve their structural integrity and longevity. They may be exposed to various conditions during their service life, which deteriorate their mechanical properties and structural performance. Amongst all, fire is one of the most threatening elements which render safety problems. The compressive strength of normal concrete would fall significantly when exposed to temperatures up to 400 °C [
1]. The dramatic decline in the strength caused by elevated temperatures (i.e., above 800 °C) and spalling damage would result in the loss of load bearing capacity of structural elements [
2,
3], which is mainly due to the chemical decomposition of cement components during exposure to elevated temperatures [
4]. Material selection and acquiring knowledge about the high-temperature characteristics of new concrete technology, such as self-consolidating or self-compacting concrete (SCC), high strength concrete (HSC), and heavyweight concrete (HWC), are the fundamentals for minimizing the high-temperature related problems and maintaining the structural integrality and sustainability of concrete structures at fire conditions.
By substituting normal aggregates with denser aggregates, such as barite, hematite, magnetite, iron shot, etc., within cement paste, heavyweight concrete will be produced. In general, the specific gravity threshold for heavyweight aggregates is 3000 kg/m
3. Concrete with a specific gravity of 2600 kg/m
3 is categorized as HWC [
5]. The electromagnetic shielding effectiveness of HWC makes it an appropriate structural material for nuclear power stations where insulation against radioactive radiation is the foremost concern. However, the sophisticated performance at elevated temperature is of paramount importance in such installations as well. According to Gencel [
6], there are two sources of heat generation in nuclear industries, including the internally produced heat owing to the gamma rays and the transferred heat from the hot parts of the reactors. In a research conducted by Horszczaruk et al. [
7], the effect of the addition of nanosilica on the mechanical performance of HWC containing quartz, barite, and magnetite aggregates exposed to elevated temperatures were explored. Of these, magnetite-based HWC showed the highest thermal stability, and hence it was concluded that the type of the heavyweight aggregate would determine the thermal resistance of the HWC. Moreover, they revealed that the addition of nanosilica up to 3% would enhance the thermal resistance of the cement mortar containing heavyweight aggregates, especially at temperatures lower than 200 °C [
7]. Ling and Poon [
8] also revealed that the type of heavyweight aggregate plays a key role in the retention of the mechanical properties of the HWC at elevated temperatures. In barite-based HWC, the thermal conductivity of barite leads to thermal spalling [
7] and a dramatic decrease in density and strength of the concrete at 300 °C [
8]. In another research, Gencel [
6] reported that hematite-based HWC has better compressive strength than plain concrete at elevated temperatures and by increasing the hematite content, reduction in the compressive strength would be lower since the porous and rough surface of hematite aggregate will enhance the bond formation in the interfacial transition zone (ITZ) and therefore improve the mechanical properties of the concrete. Therefore, the type, physiochemical characteristics, and structure of incorporated heavyweight aggregates play important roles in the fire performance of HWC.
SCC is another new-born technology in the concrete industry which has represented outstanding advantages over ordinary concrete since its development in 1986, including better flowability, high resistance to segregation or bleeding, filling capacity, and high strength [
9]. For the fabrication of structures with complex shapes or highly congested reinforcement, the superior characteristics of SCC make it an appropriate structural material, since it can flow under its own weight without segregation and easily reach to remote corners [
10]. For applications in which high durability is required, HSC is the superior structural material owing to its praiseworthy mechanical properties, including high compressive strength, high stiffness, and high tensile strength [
11]. More recently, high-strength SCCs have been the subject of various research [
12,
13]. A sufficient flowability and resistance to segregation would be provided by means of superplasticizer and viscosity modifying agent while water/binder and fine/coarse aggregate ratios, and maximum aggregate size are other critical influencing parameters [
14]. However, despite the growing applications of SCC, HSC, and high-strength SCC, their fire performance and long-term durability as exposed to elevated temperatures still need further research. Spalling and loss of mechanical properties are the major problems associated with high-temperature conditions [
15,
16,
17].
Various researchers reported explosive spalling of SCC in the range of 180–300 °C, leading to the loss of essential mechanical properties [
13,
18], which is mainly due to the microstructure of SCC. Compared to conventional concrete, SCC contains larger binder content accompanied by chemical components such as silica fume, fly ash, and ground granulated blast furnace slag (GGBFS), which is almost 200 kg/m
3 more filler than conventional concrete [
19]. This will result in a lower permeability and consequently more pressure built-up within SCC increasing the risk of spalling. Pathak and Siddique [
20] explored the effect of the addition of class F fly ash on the mechanical properties of SCC at elevated temperatures in the range of 20–300 °C and observed considerable mass loss in the range of 200–300 °C which was accompanied by a dramatic decline in splitting tensile strength due to the departure of bound water. Bakhtiyari et al. [
15] studied the fire performance of SCC containing limestone and quartz powder in the range of 150–1000 °C and reported that the temperature range of 480–650 °C is the most critical range for spalling in SCC and normal concrete, and SCC is more susceptible to spalling than normal concrete. However, SCC showed a better retention of mechanical properties at an elevated temperature compared to normal concrete. Li et al. [
12] explored the fire performance of HSC at 200, 400, 600, 800, and 1000 °C by means of an oil furnace. They observed a declining trend in the compressive and splitting tensile strengths, and spalling, as yellow, off-white and red straws appeared, starting at 200 °C. Their observations were not in line with other research, especially before 400 °C, due to the different heating profile and the utilization of an electric furnace by other researchers which is not capable of revealing the real situations of fire due to its very low heating rate. Kodur and Sultan [
11] pioneered the thermal properties of HSC at elevated temperatures in the range of 0–1000 °C and reported that carbonate aggregates would enhance the fire performance of HSC, and hence concluded, that the type of aggregate can affect the fire performance of HSC significantly.
The main objective of this study is to evaluate the fire performance of magnetite-based heavyweight SCC (HWSCC) and heavyweight HSC (HWHSC), as well as develop numerical models capable of predicting their performance as exposed to elevated temperatures. For that, by substituting magnetite aggregate in the cement paste, two mix series were prepared, of which one represented the characteristics of HWSCC while other one reveled HWHSC characteristics. Thereafter, the prepared mixes were subjected to the fresh- and hardened-state properties to analyze the workability and mechanical properties of HWSCC and HWHSC at the ambient and elevated temperatures including 20, 100, 300, 600, and 900 °C. Ultimately, empirical relationships for compressive strength and modulus of elasticity of HWSCC were proposed and validated by experimental data, which facilitate the assessment of the performance of HWSCC at fire conditions.
2. Experiment
2.1. Material
The General Portland (GP) and grade 1 fly ash complying with the requirements of AS3582.1 [
21] were used in this experimental study. Ground granulated blast furnace slag (GGFBS) is another supplementary cementitious material used in this study conforming to AS3582.2 [
22]. Very fine silica fume was utilized to provide a dense and impermeable concrete in accordance with AS3582.3 [
23].
Table 1 summarizes the chemical composition and properties of cement, fly ash, GGFBS, and silica fume.
Ten mm and <4 mm natural crushed aggregate were used as coarse and fine aggregate, in accordance with AS 1141 (2011), for both SCC and HSC control mixtures. Fine AFS 45-50 silica sand was also used as fine aggregate. Natural magnetite with a density of 3300 kg/m
3 was utilized with five various sizes including 0.5–1, 1–2, 2–4, 4–6, and 6–10 mm. The former three small particle sizes replaced the <4 mm natural crushed aggregate, as fine aggregate in proportion, whereas the latter two big sizes substituted the <10 mm, as coarse aggregate in the synchronous ratio. All the aggregates utilized in the preparation of concrete mix designs have a diameter of <10 mm, therefore micro-concretes were produced. The properties of natural aggregates and sand are presented in
Table 2. The particle distribution of magnetite heavyweight aggregate and its chemical composition are reported in
Table 3 and
Table 4 respectively.
The superplasticizer admixture (SP) was used that satisfies Type SN chemical admixture. It is designed to improve the flow properties of concrete by lowering the viscosity and yield stress of fresh concrete. High-range water reducer agent (HRWRA) satisfies type HWR. To reduce the difference between disparate content and maintain the homogeneity of the concrete, viscosity modifying admixture (VMA) was added to the mixes which represent SCC characteristics.
2.2. Mixing Design
Eight concrete mix designs were prepared, in which cement, aggregate, water, and admixtures were the main constituents. A parallel control trial was also adopted. Reducing the water/cementitious materials (w/cm) ratio by means of water-reducing admixtures and superplasticizer, and utilization of high-strength micro-cement containing GGBFS, fly ash, and in particular silica fume, were the main approaches to obtain HSC and SCC. Meanwhile, magnetite heavyweight aggregate was incorporated to produce HWSCC and HWHSC.
Figure 1 shows the concrete mixer used in this study.
Based on the absolute volume approach developed by the American Concrete Institute, HWSCC and HWHSC mix designs were prepared [
24,
25], and then categorized as series 1 and series 2. The specimens in series 1 were designed to comprise SCC with different heavyweight aggregate replacement ratios. This category contained four different proportions of magnetite aggregate including 0, 50, 75, and 100% labeled SCC, HWSCC50, HWSCC75, and HWSCC100 respectively, of which SCC was the control mix design. The mixture proportions of series 1 are shown in
Table 5. The prepared specimens in series 2 represented HSC with various heavyweight aggregate replacement ratios, similar to series 1 named HSC, HWHSC50, HWHSC75, and HWHSC100. The mixture proportions of series 2 are shown in
Table 6.
The binder compositions of control mixtures of both series 1 and 2 were 51.45% GP cement, 25.7% fly ash, 17.2% GGBFS and 5.7% silica fume. Both control samples were prepared based on 583 kg/m3 binder content. The w/cm ratio and binder/aggregate ratio were 0.45 and 0.25 for series 1 and 0.3 and 0.29 for series 2, respectively.
2.3. Fresh-State Properties of SCC
The fresh-state characteristics of SCCs (series 1), including the flowability and passing ability were evaluated based on the European Guidelines for SCCs, including the slump test, T
500mm time, and J-ring flow test [
26]. For HSCs (series 2), only the slump test was conducted due to the negligible difference between the fresh properties of HSC and conventional normal concrete.
Figure 2 shows specimen during the slump and J-ring tests.
2.4. Hardened-State Properties
To measure the compressive strength and strain gauge test, each mix design required twenty 100 × 200 mm cylinders for five different temperatures (20, 100, 300, 600, and 900 °C) of which, three cylindrical samples were used for the compressive strength test and one sample for the strain gauge test at each temperature. All hardened-property measurements were carried out 28 days after casting. To do this, after demolding after 24 hours, all samples were cured in a curing room with a controlled humidity and temperature for 28 days. For series 1, both the compressive strength and strain gauge test were carried out by a BLH (Baldwin Lima–Hamilton) pressure testing machine. Compared with the compressive strength test, one more vertical strain gauge was attached to the strain gauge test which was parallel to the direction of the applied load. For series 2, the upper force limit of the BLH was 600 kN which was insufficient to reach the breaking stress of HSC. Hence, an ALSMER pressure testing machine was used to complete both tests. Harden densities were evaluated at room temperature while mass loss and spalling were assessed at 100, 300, 600, and 900 °C.
2.5. High-Temperature Test
A furnace was used to simulate the fire ignition conditions. For each mixing, four samples were exposed to the fire under each target temperature. The rate of heating was 5 °C/min. Before exposing to high temperatures, the weight of all samples was measured in an ambient temperature (20 °C). After exposure to the target elevated temperatures, the weight of the samples was measured again. The change in weight before and after firing revealed the mass loss of the examined samples. Furthermore, a high temperature leads to a change in the color and occurrence of cracks owing to the weight loss, which are the visual inspections of spalling. To evaluate the residual compressive strength and strain gauge test after heating, all samples must be maintained against explosion after exposure to the target elevated temperatures. These samples were assessed by the same mechanical-property-test procedure at room temperature.
2.6. Numerical Modelling
To predict the behavior of concrete at fire conditions, establishing appropriate empirical relationships is essential, since elevated temperatures influence the mechanical properties of the concrete structures, and hence, these impacts should be taken into consideration by engineers to fulfill safety concerns related to the fire performance of concrete structures. More specifically, due to the typical environments where the HWC would be utilized, e.g., nuclear power plants, a deep understanding of the possible changes in the mechanical properties of HWC at elevated temperatures is of paramount importance to maintain the structural integrity of the HWC-based structures. Nevertheless, a limited number of research exploring the relationships for HWC at elevated temperatures is available. Therefore, numerical models for residual compressive strength and moduli of elasticity at high temperature are proposed and validated by experimental results. Moreover, these established relationships are compared with available models built by other researchers to assess their generality and workability for HWC at elevated temperatures.
4. Numerical Modelling
Developing numerical models to describe stress-strain curves for concrete at elevated temperatures is of great importance [
54,
55,
56,
57,
58,
59,
60,
61,
62] since, from one hand, it would not be an easy-achieving task to measure the mechanical properties in such high-temperature range, particularly above 300 °C, and from the other hand, HWC-based structures usually will expose to high temperatures due to the places where they would be installed. Herein, based on the experimental data and regression analysis, empirical relationships for residual compressive strength and modulus of elasticity of HWSCC containing magnetite aggregate at elevated temperatures have been proposed in the Equations (1) and (2). According to the results obtained for hardened-state density, only SCC containing 100% magnetite content can be considered as HWSCC, therefore, the regression analyses were conducted based on the experimental results for HWSCC10.
In which and are the residual compressive strength and modulus of elasticity, and and are the compressive strength and modulus of elasticity at ambient temperature, respectively.
Table 12 represents available relationships for compressive strength and modulus of elasticity of unconfined concrete in compression at ambient temperature which are applicable for HWC with density up to 4500 kg/m
3 [
63].
Figure 8 shows the obtained results from these relationships compared with the obtained results from the proposed relationships.
The only available relationship which can be used to predict the compressive strength of HWSCC containing magnetite at elevated temperature was stablished by Carreira and Chu [
64]. However, this model predicts the compressive strength as a function of strain. As can be seen in
Figure 8a, the proposed model by Carreira and Chu [
64] is only capable of predicting the decreasing trend in the compressive strength up to 600 °C, and it is not applicable for temperatures above 600 °C.
Regarding the modulus of elasticity, Carreira and Chu [
64] and Yang et al. [
63] models are based on the density of the specimen, since the modulus of elasticity increases by increasing the density, especially for values higher than 2500 kg/m
3 [
63]. However, as can be seen in
Figure 8b, these models are not able to predict the behavior of HWSCC at elevated temperatures since the
soared by losing mass, and subsequently density of HWSCC containing 100 magnetite aggregate up to 300 °C is reduced. Above 300 °C, both density and
plunged but since the change in the mass at elevated temperature is not as significant as the observed declining trend in the modulus of elasticity, hence the available models cannot predict the modulus of elasticity for HWSCC at elevated temperatures accurately. The proposed model, which only considered the change in the modulus of elasticity by increasing the temperature, reveals a better workability than the reference models. Nevertheless, considering both the density and effect of temperature on
results in a more accurate model for the modulus of elasticity of HWSCC at elevated temperatures, which needs further research.