1 Introduction

A recent point of interest to researchers has been on the performance of UHPC. UHPC is a cementitious composite material developed in the past few decades characterized by its high compressive strength, which is typically >150 MPa [8]. The high compressive strength of UHPC is a result of a specifically designed mixture which includes low water content, optimized granular packing, little to no coarse aggregates, high binder content, and high fine aggregate content. UHPC mixtures typically contain internal fibre reinforcement to help improve tensile strength and ensure non-brittle behaviour. Mixtures containing internal fibres are referred to as ultra-high-performance fibre-reinforced concretes (UHPFRC).

The definition of UHPC has developed over the years. Until the end of the 1960s, concrete with a compressive strength of between 35 and 45 MPa was considered high-strength concrete (HSC) [12]. By the middle of the 1980s, high-strength concrete with compressive strengths of 80 MPa was able to be produced. Additionally, in the 1980s, a distinction was made between high-strength concrete and high-performance concrete. High-strength concrete is defined solely based on its compressive strength values, while high-performance concrete is defined based on specific performance criteria such as its durability, high strength, and high workability. By the end of the twentieth century, high-strength concrete was achieving strengths of up to 150 MPa, which is more in-line with the present definition.

Despite having several advantages over conventional concrete, UHPC may experience severe explosive spalling when exposed to high temperatures [17]. If not designed correctly, the dense microstructure of UHPC can prevent water vapour pressures from freely dissipating, resulting in explosive spalling under elevated temperatures. When properly proportioned, the addition of synthetic fibres, such as PP fibres, was found to be efficient in improving the fire performance of UHPC [37]. This paper presents a critical review of previous studies and investigations reported in the literature about the performance of UHPC exposed to high temperatures. Gaps in knowledge about fire performance of UHPC are highlighted and future needs in research are identified.

2 Background

2.1 Effects of Fibre Type

Gurusideswar et al. [10] found that under quasi-static loading, the addition of micro-steel fibres into the UHPC matrix improved the tensile strength by 310.5%, compared to non-fibre specimens. The addition of fibres also increased the energy absorption capacity. Fibres kept the structural integrity of the matrix, which lead to multiple cracks and no complete fragmentation. The addition of PP fibres was found to reduce spalling by increasing permeability through three main mechanisms, including, the formation of capillary pores from melting of fibres, creation of transition zones opened to diffusion and the additional micro pores gained by structure aeration when fibres are mixed [7]. The positive effects on fire resistance of PP fibres were also reported by [11] who found that at 150 °C PP fibres began to melt and above 200 °C volatilize, providing additional capillary pores. The additional capillary pore allowed water pressure to dissipate out of the UHPC matrix, reducing internal tensile stresses. Ozawa and Morimoto [29] investigated the effects of PP, jute, or water-soluble polyvinyl alcohol (WSPVA) fibres on mitigating vapour pressures in high-performance concrete (HPC) at elevated temperatures. The PP, jute and WSPVA fibre specimens did not undergo explosive spalling. Both the jute and WSPVA fibres developed pressure-induced tangential spaces (PITS) at the fibre-concrete interface, which transported water vapour from inside to outside the concrete, thereby reducing the propensity for spalling.

When investigating UHPC specimens that contained PP fibres in the mix under exposure of temperatures up to 800 °C, Xiong and Liew [36] found that no spalling occurred. The residual elastic modulus, corresponding to secant modulus between the stress equal to 40% of peak stress and the stress corresponding to a strain of 5 × 10–5, of the UHPC mixtures with PP fibres reduced as the heating rate increased. After exposure to 200 °C, there was an increase in residual compressive strength but not in elastic modulus, which was attributed to shrinkage by water removal. Specimens only containing steel fibre reinforcement were also tested, and it was concluded that steel fibres cannot effectively prevent explosive spalling on their own. This may occur due to the fact that steel fibres do not melt at lower temperatures, and therefore do not provide an escape mechanism for pore-pressure to dissipate [4]. This conclusion was also reached by Nazri et al. [25] who found that at 400 and 600 °C, UHPC specimens with either 2% or 2.5% by volume of steel fibres underwent severe spalling, resulting in total disintegration. The results showed that UHPFRC specimens exposed to elevated temperatures cannot contain only steel fibres but must have a combination of steel and PP fibres.

Maluk et al. [21] determined that the inclusion of PP fibres clearly reduces the propensity for heat-induced spalling. Mixes with short fibre length had a greater tendency to spall than those with longer PP fibres. This can be explained by the more effective bridging effects of longer fibres, which helps to reduce spalling by preventing cracks from growing in diameter. Compared to fibrillated PP fibres, monofilament or multifilament PP fibre types reduced heat-induced spalling due to their smaller cross-section, as smaller cross-sections tend to mitigate spalling more effectively than larger cross-sections.

Zhang et al. [37] investigated the mechanisms by which PP fibres mitigate explosive spalling in UHPC. It was observed that there was a significant increase in gas permeability, of two orders of magnitude, before the melting point of PP fibres was reached, which suggested that the melting of the PP fibres and creation of capillary pores is not critical in the mitigation of spalling. Additional tests showed that the creation of microcracks, which formed a network of interconnected cracks, was the main factor in increasing gas permeability. Microcracks formed due to a mismatch in thermal expansion coefficients between the matrix and the fibres. This suggests that it is not necessary to include PP fibres specifically to mitigate explosive spalling, but rather include materials that exhibit high thermal expansion coefficient mismatch with the cement matrix. This finding was collaborated by Li et al. [17], who found that the hot permeability of UHPC suddenly increased within the temperature range of 105 to 150 °C, in the presence of PP fibres, well below the melting point of PP fibres of 166.5 °C. This led the authors to the conclusion that it is in fact the creation of interconnected microcracks due to the large thermal expansion and mismatch between the PP fibre and the concrete matrix.

2.2 Effects of Fibre Volume

An early study by Richard and Cheyrezy [34] on Reactive Powder Concrete (RPC) found that the addition of steel fibres at a ratio between 1.5 and 3% by volume improved ductility. The economic optimum was found to be at 2% steel fibres by volume. Orgass and Klug [27] observed that the higher the fibre volume fraction, the more the specimen size influenced the flexural strength. Up until a short steel fibre volume content of 2%, a post-fracture ductile behaviour was observed in the specimens.

Heinz et al. [11] found that a UHPC mixture containing a combination of 3.5% by volume of steel fibres and 0.66% by volume of PP fibres demonstrated superior fire resistance. Specimens with this mixture did not undergo any spalling throughout a fire test duration of 90 min. The fire test followed the standardized ETK temperature curve, reaching 1020 °C after 90 min. When utilizing different fibre types, but a similar test setup, Ozawa and Morimoto [29] found that adding 1.03 kg/m3 of 12 mm long jute fibres and 1.95 kg/m3 of 4 mm long WSPVA fibres were both effective in mitigating spalling due to elevated temperatures.

Xiong and Liew [36] found that PP fibre dosage of only 0.1% by volume was effective in preventing spalling up to 800 °C. As the PP fibre dosage increased, the strength reduction was similar regardless of different heating rates. These results are important in that they demonstrated that fire resistance can be provided in low doses of PP fibres without compromising the high strength of UHPC. Maluk et al. (2017) also investigated how the volume of PP fibres can reduce the propensity for HPC to explosively spall. Higher doses of PP fibres were generally more effective in mitigating spalling, although some low doses of specific PP fibres, those with small cross-sections, were equally effective in reducing spalling. Again, this lessens the dependency on high doses of PP fibres to mitigate spalling. This was helpful to reduce spalling without significant compromise in the compressive strength of UHPC.

When UHPC with either 2 or 2.5% steel fibres by volume was tested for fire resistance, Nazri et al. [25] found that the higher percentages of steel fibres produced greater the 28-day strengths. However, the steel fibres were not an effective mechanism to mitigate spalling, and it was concluded that a synthetic fibre must be used in addition to the steel fibres.

Li et al. [17] found that fibre length and dosage were the critical factors to improving the permeability of UHPC post-exposure to elevated temperatures. The authors found that the permeability was positively correlated to the aspect ratio and dosage of fibres. This is an important finding in that it indicates that designers should focus on creating UHPC mixtures with both a high aspect ratio and fibre dosage, instead of focusing on the geometry of the fibres.

2.3 Additional Effects

2.3.1 Effects of Loading and Heating Rate

A study by Ali et al. [2] tested reinforced high-strength concrete (HSC) under various loading and heating rates. The authors found that the loading rate did not influence the propensity for explosive spalling, regardless of the heating regime. However, low heating rates minimized the risk of explosive spalling. A high heating rate reduced the column’s fire resistance. Liang et al. [19] also showed that the explosive spalling of UHPC specimens was dependent on the heating rate. This finding led to the suggestion that when testing UHPC under fire exposure, the heating rate should be limited when temperatures fall within the range of 300 to 500 °C, as between 320 and 380 °C too great of a heating rate will result in explosive spalling. Banerji et al. [5] found that a higher applied load level decreased the extent of spalling, as the high load increased the amount of cracking which released more pore pressure. For example, at an applied load of 40% of ultimate capacity, the extent of spalling in the beam was 13.3%. However, at a higher applied load of 60% of ultimate capacity, the spalling extent in the beam was reduced to 7.5%.

[37] found that as the heating rate was increased, the elastic modulus of the UHPC specimen with PP fibres reduced. As [2] found that a high heating rate reduced UHPC’s fire resistance, logically it follows that elastic modulus would be reduced. Elastic modulus is a measure of a substance’s resistance to non-permanent deformations, meaning that the higher the elastic modulus, the more an object will return to its initial state. As the elastic modulus decreases, the substance will be more vulnerable to permanent deformations. The structural performance of an element made from UHPC in fire will also be less desirable as the substance will be increasingly subjected to plastic behaviour and permanent deformations.

Liang [20] tested UHPC specimens with and without PP fibres under heating rates of either 1, 4 or 8 °C/min. Under all heating rates, specimens experienced explosive spalling, however under an increased heating rate, explosive spalling started and ended earlier, spalling starting and ending temperature increased and smoke escaping temperature range increased. Additionally, the debris of specimens under higher heating rates were smaller than that of those with lower heating rates, indicating that increased explosive spalling had occurred. Therefore, while explosive spalling cannot be completely prevented under any heating rate, it can be minimized by reducing the heating rate.

Pyo et al. [30] found that with low loading rates, the addition of steel fibres in UHPC resulted in the slowing of crack propagation. In contrast, at high loading rates, the number of fibres did not influence the rate of crack propagation. Although the use of fibres does not significantly help at high loading rates, their use can still be recommended as they significantly improve their performance under low loading rates. Gurusideswar et al. [10] observed that fibre reinforced UHPC specimens were less sensitive to loading rate changes than UHPC without fibre reinforcement. At higher loading rates, there was a noticeable enhancement in pullout resistance of steel fibres, bond strength and stiffness in UHPC specimens.

2.3.2 Effects of Strain Rate

Pyo et al. [31] observed that at room temperature, with increases in crack initiation strain rate, the crack speed asymptotically increased. However, visual observations showed that no changes in crack surfaces with increasing crack speeds. Simply, this means that while cracks will occur faster at high strain rates, their extent will be no worse than at lower strain rates. Liang et al. [19] performed Split Hopkins Pressure Bar (SHPB) on UHPC specimens and found that specimens exhibited the highest and lowest strain rates at 200 °C and 600 °C, respectively. Under combined loading tests, it was seen that with higher strain rates, the peak stress also increased resulting in a broader stress–strain curve. The broader stress–strain curve is an indication of greater energy dissipation, which is a beneficial property to the resistance of impact. This is also an indication that with increased strain rate under elevated temperatures, UHPC strength is increased. Additionally, it was observed that with increased strain rate under elevated temperatures, the modulus of elasticity decreased.

Mishra and Singh [23] revealed that concrete’s behaviour under differing strain rates depends on the loading conditions. For example, concrete is most sensitive to strain rate when loaded in tension and least sensitive when loaded in compression. Relatively speaking, it was seen that UHPC had high tensile resistance at high strain rates, attributed to the steel fibres. Gurusideswar et al. [10] examined the tensile strength and failure of five different UHPC mixes under a wide range of strain rates. Results showed that the dynamic tensile strength increased with an increased strain rate. It was found that under a high strain rate, the post-cracking strength of UHPC specimens was more sensitive than the initial peak strength.

2.3.3 Effects of Specimen Size and Geometry

Orgass and Klug [27] observed that the size of the specimen can influence how effective fibres are in UHPFRC. For example, the flexural strength of smaller prisms was found to be higher than that of larger beams, as fibre orientation tends to be stronger in smaller specimens. The fibre orientation effect was also shown in that smaller specimens had higher ductility than larger specimens. Additionally, it was observed that the higher the fibre content was, the more the specimen size influenced the flexural strength. Research by Reineck and Greiner [33] showed a similar effect as there was a noticeable decrease in flexural strength with the increased depth of specimens.

Fehling et al. [9] observed that compared to traditional reinforced concrete (RC), the geometry of UHPC specimens had less of an influence on compressive strength. However, in contrast, Naeimi and Moustafa [24] showed that 50.8 × 101.6 mm cylinders had a 10% higher compressive strength than 76.2 × 152.4 mm cylinders. The discrepancies in results point to the need for further research on the influence of specimen geometry on the mechanical properties of UHPC.

2.3.4 Effects of Aggregate

In the development of reactive powder concrete (RPC), Richard and Cheyrezy [34] found that the elimination of coarse aggregates and optimization of granular mixtures created a homogenous and dense cementitious matrix. RPC exhibited superior mechanical performance at room temperature, specifically, the compressive strength was higher compared to traditional RC. The improved mechanical properties were attributed to the elimination of coarse aggregates from the concrete matrix.

Arioz [3] investigated how different aggregate types influenced concrete’s fire resistance. The concrete mixtures contained either crushed limestone or river gravel as aggregate. Results showed that the aggregate type significantly influences concrete’s resistance to fire, as specimens containing river gravel underwent larger strength reductions under exposure to elevated temperatures. These tests were performed on traditional RC, and show the advantage of eliminating coarse aggregate from the UHPC matrix in regards to fire performance.

Liang et al. [18] developed various UHPC mixtures with the primary goal being to create one with high fire resistance. The authors used either steel slag or quartz sand as fine aggregate. The specimens were reinforced with PP fibres, steel fibres, or a combination of both. It was found that the use of steel slag as fine aggregate, as well as a hybrid mixture of steel and PP fibres, resulted in excellent fire resistance properties compared to the mixture with no fibres. Specifically, after exposure to 1000 °C, specimens with the hybrid mix retained 69% of their room-temperature compressive strength.

A study by Li et al. [17] investigated how the aggregate size influences the behaviour of UHPC at elevated temperatures. The authors found that the use of larger aggregates helped to increase the permeability of the UHPC mixture, which prevented the build-up of internal pore pressure, which in turn prevented the occurrence of explosive spalling. These factors improved the fire resistance of the UHPC specimens. The results show that when formulating a UHPC mixture, designers should consider utilizing slightly larger aggregates, with maximum aggregate sizes up to 5 mm.

UHPC is a highly cementitious material, raising concerns about how it contributes to climate-change [28]. Researchers have been investigating how to improve the ecological properties of UHPC by partially replacing cement with different forms of aggregate. Abdulkareem et al. [1] substituted 30% of cement in a UHPC mixture with Blast Furnace Slag (BFS). It was found that the substitution improved the packing density, workability, accelerated setting time, and promoted cement hydration which by extension increased early-age compressive strength. It also decreased superplasticizer content by two times, which improved the ecological properties of the UHPC mix. However, the effects of an ecological UHPC mixture on fire resistance have not been thoroughly investigated. It is recommended that further research is performed on how the ecological UHPC mixtures perform under elevated temperatures.

3 Research Methodology

Reviewing available research is the first step in a long-term plan to provide further research into UHPC DWI panels exposed to fire. The literature review revealed that considerable research has been performed on the mechanical properties of UHPC after exposure to high temperatures. However, there has been little focus placed on mechanical properties, specifically compressive and tensile strengths, at elevated temperatures. Additionally, the effect of adding polyvinyl alcohol (PVA) fibres on UHPC’s physical, mechanical, and thermal properties has not been thoroughly investigated. Finally, only preliminary research has been performed on UHPC DWI wall panels. The initial promising work must be expanded to inform future code provisions and safely recommend its use.

The first phase of research will focus on material testing of the UHPC planned to be implemented into the DWI panels. Three main objectives will be the focus for the initial research phase of material testing: (1) to assess the potential for small-scale UHPFRC specimens to spall, (2) to quantify the strength at and after exposure to elevated temperatures of small-scale cylinder and prism UHPFRC specimens, and (3) to quantify the thermal properties at high temperatures of small-scale UHPFRC specimens. Table 1 provides an outline of the testing matrix proposed for the material testing. For each type of testing, the mixture type in terms of the percentage of fibres and the number of specimens required for each mixture type is given.

Table 1 The material testing matrix
Full size table

Three different fibre combinations of PP and polyvinyl-alcohol (PVA) fibres will be used to reinforce the small-scale specimens. The UHPC mixtures’ fibre contents are shown in Table 1. Cylinders will be tested for compressive strength after seven and 28 days. The material tests will be completed at Queen’s University using the Instron SATEC machine (Fig. 1), which incorporates a high-temperature environmental chamber. Both compressive and tensile tests will be performed with the Instron SATEC. Cylinders of 75 mm diameter × 150 mm height will be used for compression and prisms of 700 mm length × 45 mm height × 45 mm width will be used for tension. The cylinder and prism dimensions were chosen based on the size capacity of the high-temperature environmental chamber.

Fig. 1
figure 1

Instron SATEC machine

Full size image

Tests at high temperatures will be performed under both steady-state and transient conditions, and a summary of the proposed testing matrix is shown in Table 1. Steady-state condition tests involve heating the specimens to a specified temperature and then loading until failure. Steady-state tests will be performed at temperatures of 25, 200, 300, 400, and 500 °C for compression specimens. For tension specimens, steady-state tests will be performed at temperatures of 25, 300, 500, and 800 °C. In total, 60 specimens will be tested in steady-state compression and 48 specimens will be tested in steady-state tension. Transient condition tests involve loading the specimens to a specified load-level and then increasing the temperature until failure. Transient tests will be completed at 30 and 50% of ultimate for compression and tension specimens. Twenty-four specimens will be tested in both transient compression and transient tension. Residual tests will also be performed for both compression and tension specimens. Residual tests will be completed after exposure to 200, 300, 400, and 500 °C for compression specimens. Similarly, after exposure to 300, 500 and 800 °C, residual tests will be performed on compression specimens. Twenty-four specimens will be tested in residual compression and 36 specimens will be tested in residual tension.

One high-temperature environmental chamber can reach a maximum temperature of 550 °C and the SATEC Instron can apply a maximum load of 600 kN to specimens. The environmental chamber will only be used for the planned compressive tests. One end of the specimens will be pulled while the other remains embedded in the jaws of the machine. Only the middle of samples will be heated, to ensure that failure does not occur at the ends of the samples. A second high-temperature environmental chamber can reach temperatures up to 1200 °C.

To characterize thermal properties, specifically thermal conductivity and specific heat, the HotDisk system at National Research Council (NRC) will be used. The HotDisk system is a non-destructive test method to determine thermal properties. It is an absolute method that does not require repeated calibrations with each use. Prism specimens with dimensions of 50 mm length × 50 mm height × 0.0125 mm width, will be used in HotDisk testing. Twenty-four specimens will be fabricated to test in the HotDisk system. The results from the HotDisk system in terms of thermal properties will be used to develop numerical models of UHPC DWI panels exposed to fire.

4 Conclusion

The literature review shows that UHPC is potentially more vulnerable to elevated temperatures because of obstruction in the dissipation of vapour pressure, that can potentially lead to physical damage through explosive spalling. Based on the review, the following conclusions can be derived:

  • Multiple researchers have shown that UHPC spalling under elevated temperatures can be mitigated by incorporating polypropylene fibres into the mix. The influence of other types of fibres (including steel fibres) on UHPC’s resistance to explosive spalling is still questionable, but a combination of steel and PP fibres was noticed to be promising in preventing the spalling of UHPC.

  • Fibre volume is an important factor that influences the efficiency of fibres in improving UHPC fire performance. Researchers investigated the use of 1–3.5% of fibre volume on the fire performance of UHPC. However, the optimum fibre volume is still debated amongst researchers.

  • The review reveals that other factors, such as loading and heating rate, strain rate, specimen size and geometry, and aggregate type and size influence the performance of UHPFRC exposed to elevated temperatures.

  • While a considerable amount of research has been conducted on the mechanical properties of UHPC after exposure to high temperatures, less attention has been directed toward the mechanical properties of UHPC during exposure to high temperatures, which better represents realistic fire scenarios. The review also shows that little focus has been placed on the structural behaviour of UHPC in fire.

To fill the research gap about the UHPC performance during and after exposure to elevated temperatures, work is in progress at Queen`s University, Kingston, Canada to assess the fire performance of DWI wall panels cast with UHPC under fire exposure. Additional compressive and tensile behaviour of concrete specimens reinforced with various amounts and combinations of PVA fibres and PP fibres will be tested at temperatures as high as 500 °C under steady-state and transient temperature heating regimes.