*7.2. Cement Dosage and Type*

Concrete strength is among other functions a function of the water to cement ratio, so naturally, it influences residual strength. Very few papers tackle cement dosage, and even fewer address cement types. In [46], concrete samples with three different cement dosages and the same w/c ratio were tested. Normal Portland cement with the addition of fly ash was used; it can be classified as CEM II/B-V. Two different cooling regimes were used: inside the furnace and at room temperature. Tests were performed directly after cooling and after 30 days of re-curing. Analyzing the test results, one can conclude that the cement dosage is not influencing the residual strength, as all mixtures behave in a very similar way. In [24], the influence of the w/c ratio on the residual strength was tested. Two types of specimens with different w/c ratios were used (both using CEM I). The tests were performed after 7 days of re-curing. The results show that although the difference between the w/c ratios is substantial (17%), the influence on relative residual strength is negligible. In [50], three different w/c ratios: 0.35, 0.5, and 0.55 were tested. Specimens were heated to 600 ◦C, cooled down, and tested after various re-curing times. Results for w/c of 0.5 and 0.55 are almost identical, and for 0.35, the initial residual strength is much lower but the increment is similar for all w/c ratios.

Other w/c ratios (0.31 and 0.45) were examined in [42]. For all peak temperatures ranging from 200 to 800 ◦C, the w/c ratio has been shown to be insignificant when considering its influence on relative residual strength (Table 11). In [88], concrete with a w/c ratio of 0.22, 0.33, and 0.57 at temperatures up to 450 ◦C was tested. The results showed that the loss of initial strength was lowest for 0.22 (approximately 20%) and higher for 0.33 and 0.57 (approximately 30%). In [82], research on the w/c ratio was performed by exposing three different concrete mixes, with w/c ratios of 0.6, 0.42, and 0.27, to temperatures ranging from 150 to 900 ◦C. The heating rate was set at 3 ◦C/min, and the exposure time at the peak temperature was 1 h. The results show that the smaller w/c ratios perform slightly better and maintain more strength. A similar test was performed in [70]; concretes with w/c ratios of 0.5 and 0.7 were exposed to elevated temperatures (ranging from 200 to 1000 ◦C), and after 28 days of re-curing, residual strength was tested. The results showed that w/c influences residual properties in higher temperature registers, i.e., 600 ◦C and above. Higher w/c ratios resulted in lower residual strength. Both [89,90] present the influence of the w/c ratio on the residual strength of the concrete after exposure to 500 ◦C for 1 h and 4 h, respectively. Concrete mixes were prepared with normal and recycled aggregates. The results showed that a lower w/c is beneficial for residual strength, especially for recycled aggregate concrete.

In [59], concretes with different pozzolanic materials used as a partial replacement for Portland cement were tested. These were natural pozzolana and lignite fly ash. The conclusion was made that samples with pozzolanic additives are more sensitive to high temperatures, especially in the temperature magnitude of 200 to 400 ◦C. For 200 ◦C, OPC concrete registered a 25% reduction in strength, while in concretes with pozzolanic materials, this reduction ranged from 38 to 50%. Taking into account 400 ◦C, the disproportion was smaller: 50 to 65%. This behavior can be explained by the higher amount of calcium aluminates hydrate (loses part of its combined water at 105 ◦C), calcium aluminate sulfate hydrate (dehydrates at 150 ◦C), and amorphous tobermorite gel (dehydrates at 120 ◦C) in OPC–pozzolana and OPC–fly ash paste mixtures. The amount of strength gained in the

re-curing period is dependent on additives, where OPC concrete regains strength faster than concrete with pozzolanic additives.


**Table 11.** Relative residual compressive strength for different w/c ratios (0.58 and 0.68) and the difference between relative residual strengths of different w/c ratios according to [42].

\* Time maintained at peak temperature.

Four types of concrete with fly ash replacement for Portland cement were tested in [48]. The influence of the amount of fly ash directly after cooling was tested. Replacement ratios ranged from 10 to 40%. The results showed that there is no correlation between the residual strength directly after cooling and the quantity of fly ash. In [91], the influence of fly ash and metakaolin on the residual strength of HSC was tested. The results reveal that there is no large difference in residual strength directly after cooling for all mixes.

The role of peak temperature and fly ash dosage on the residual strength of lightweight concrete was examined in [92]. The level of importance determined by the Anova method was extremely favorable to the peak temperature, showing that the fly ash quantity impact was minimal. In [93], fly ash dosage did not influence the self-compacting concrete residual strength for peak temperatures up to 300 ◦C. In [94], research on the influence of finely ground pumice and silica fume on the residual properties of concrete was carried out. Specimens with different dosages of FGP (finely ground pumice) and SF (silica fume) were exposed to high temperatures ranging from 400 to 800 ◦C and then tested. The results indicate (Figure 15) that FGP additions are beneficial for residual strength, while SF slightly reduces residual strength. A similar conclusion was reached regarding SF in [95].

**Figure 15.** Relative residual compressive strength as a function of the peak temperature for different cement replacements according to [94].

In [69], the residual strength of the samples with pozzolanic cement replacement was tested. Three types of binder were chosen: the replacement of ME (natural pozzolan), PFA (high calcium fly ash), and MFA (low calcium fly ash), and ordinary Portland cement (OPC) was proposed to be 10 and 30% high. The difference in residual strength induced by the replacement ratio of Portland cement was minimal, so only types of replacement binder were considered. The samples were heated to temperatures of 100 to 750 ◦C, and the heating rate was set at 2.5 ◦C/min. After exposure to a peak temperature of 2 h, the samples were naturally cooled inside of the furnace. The residual compressive strength was investigated, and the conclusion was reached that the type of pozzolanic replacement is important only in the lower temperature range (100–400 ◦C) (Figure 16).

**Figure 16.** Relative residual compressive strength as a function of the peak temperature for different cement replacements (ME—natural pozzolan, PFA—high calcium fly ash, MFA—low calcium fly ash, OPC—ordinary Portland cement) according to [69].

Another cement replacement, ground granulated blast furnace slag (GGBFS), was investigated in [96]. Mixes were made with the replacement ratios of 10, 30, and 50%. After exposure to elevated temperature (ranging from 150 to 700 ◦C) and natural cooling in the furnace, compressive tests were performed. The results presented in Figure 17 demonstrate that for low peak temperatures (below 400 ◦C), replacement of residual strength is insignificant. For higher temperature registers, mixes with the GGBSF replacement resulted in a lower residual strength. The higher replacement ratio resulted in lower mechanical properties.

**Figure 17.** Relative residual compressive strength as a function of the peak temperature for different GGBFS replacement ratios (0, 10%, 30%, and 50%) according to [96].

The available test data prove that the w/c ratio does not have a big influence on the residual strength; only in cases of drastically different ratios, the initial strength differs by a noticeable margin. The rate at which the concrete regains strength is similar for all tested w/c ratios. The cement type (binder additives) is crucial; pozzolans or slag [97] plays an important role in the strength of the cement paste after a fire. More research is needed to fully explain the influence it has on residual strength.

#### *7.3. Additives and Fibers*

Many additives can improve concrete properties. Macro-additives, such as polypropylene or steel fibers [98] and micro-additives, such as reactive powder [99] or palm oil fuel ash [55], are only a few examples. In [100], the influence of superplasticizer, hardening accelerator, setting retarders, and air entertainers was found to be minimal (only air entertainers showed a noticeable decrease in residual properties). A similar conclusion can be reached by analyzing the paper [101]. Polypropylene (PP) fibers are said to explicitly improve concrete strength at elevated temperatures. PP (polypropylene) fibers melt and create channels that help release the internal water pressure that was built due to the increase in temperature [102]. Without a doubt, it increases the strength of hot concrete, but its influence on residual strength is less pronounced [103,104]. In [23,29], the residual strength of HSC was tested with and without PP fibers. The tests were performed after cooling, and the same conclusion was reached: PP fibers increase residual strength by a small margin (Figures 18 and 19). In [105,106], the influence of different dosages of PP fibers on the residual strength of HSC was tested. The results showed that the differences between various dosages are limited (Figures 18 and 19). The marginal influence of PP and steel fibers on the residual strength of NSC was also reached in [107].

**Figure 18.** Relative residual compressive strength as a function of peak temperature—with and without PP fibers by [29]—Xiao, [23]—Toumi, [105]—Behnood, and [106]—Eidan.

PP fibers improve residual compressive strength directly after cooling but to a very limited degree. It should not be taken into account when assessing the residual strength of concrete. More tests need to be performed on the influence of PP fibers on the re-curing rate. Channels made by melted PP fibers are impossible to repair, as they do not regenerate. The pore structure of concrete without PP can be restored to a value similar to the initial one, while concrete with PP cannot regain its previous state. In [108], it was suggested that microchannels created in place of melted PP fibers have a positive effect on water re-curing of concrete, as they accelerate the water diffusion rate but negatively impact the residual strength of air re-cured concrete. The influence of steel and PP fibers on NSC and residual properties of HSC was studied in [109], and it was found that steel fibers have a minimal effect on NSC and change the spalling temperature to a higher level. The use of PP fibers increased the spalling resistance for all samples, but a negative effect on residual mechanical properties was noticed. In [110] the influence of steel fibers on residual strength at very high temperatures (900–1200 ◦C) was analyzed, and a minimal influence

was observed. Concrete with glass and steel fibers tests were performed in [111]. Up to 30%, higher compressive strength was noted for steel fiber concrete (for glass fiber, up to 20%). This increase was especially visible for the 300–500 ◦C temperature range. In [112], PP fibers in concrete mixes were tested, and the results showed that thermal behavior and stability are not influenced by type and dosage. Similar research should be conducted on residual mechanical properties.

**Figure 19.** The difference in relative residual compressive strength with and without PP fibers by [29]—Xiao, Toumi [23], Behnood [105], and Eidan [106].

#### **8. Porosity**

Porosity, pore size, and pore distribution are the primary factors influencing the strength, durability, and permeability of concrete. Research already done in this area is collected in Table 12. High-temperature exposure increases the porosity and coarsening of the pore structure of concrete [58]. The dehydration process that occurs in the C-S-H (calcium–silicate–hydrate) gel decreases its volume and subsequently increases its porosity. Although up to 200 ◦C, a slight expansion of the cement paste is observed, above this temperature, rapid shrinkage occurs. This phenomenon greatly influences the evolution of porosity. In [60], porosity was measured using the mercury intrusion porosimeter (MIP). Compared to the preheating values, porosity directly after cooling was two times higher for 600 ◦C and three times higher for 800 ◦C. However, post-fire re-curing significantly reduced the porosity by the rehydration of particles that filled capillaries. Lower porosity results in a dense microstructure and better mechanical properties. When comparing the results presented in [60] (Figure 20), lower initial porosity leads to slightly higher relative residual strength. If two concretes with different initial porosities are compared with relation not to initial but after cooling strength and porosity, then re-curing gains are very similar (Table 13).


**Table 12.** Summary of the research on porosity.


**Figure 20.** Relative porosity as a function of re-curing time with different re-curing methods (air re-curing and water re-curing) and different peak temperatures *θ* = 600 and 800 ◦C according to [60].

In [27], an MIT (mercury intrusion porosimetry test) test was used to measure changes in connected porosity (open pore network). Tests were performed after cooling. The results compared to previous research are consistent as presented in Table 14. The relative porosity increases at a rate similar to that observed in [60], and the relative residual strength at the corresponding temperature is similar.

**Table 14.** Relative residual compressive strength and relative porosity as a function of the peak temperature according to [27].


An interesting relation was observed in [113], comparing porosity at relative peak temperature, between tests performed directly after cooling (a) and 2 months of water curing (b). Porosity in the temperature range of up to 400 ◦C is constant. In the range from 400 to 800 ◦C, porosity is increasing for the sample that was tested directly after cooling, but the porosity level of the water-cured sample is still constant. From 800 to 1000 ◦C, (a) is slowly increasing, while (b) noted rapid growth, and at temperatures above 1000 ◦C, the type of curing is irrelevant (Figure 21).

**Figure 21.** Total porosity as a function of the peak temperature and curing according to [113].

Porosity is rapidly growing, with temperature increasing above 400 ◦C and having a great impact on the strength of concrete. The calculated correlation factors between porosity after exposure and residual strength shows a significant negative porosity influence (Table 15). This relationship can be nonlinear, as the Pearson and Spearman coefficients are not very high. It was noted that the porosity growth is higher for HSC [114]. Nevertheless, damages caused by dehydration can be repaired by rehydration of the concrete. With an appropriate curing method applied to concrete exposed to peak temperatures not exceeding 800 ◦C, porosity can be regenerated to levels before heating. Taking these facts into account, the initial porosity of concrete plays a marginal role in the relative residual strength. The relative growth of residual strength (with relation to strength directly after cooling) of concretes with different initial porosities is very similar.

**Table 15.** Correlation factors between porosity after exposure and relative residual strength, based on data presented by various authors.


#### **9. Age of Concrete at Exposure**

Fire can happen in a building regardless of its age. Both very old buildings and new buildings (or even still under construction) can experience exposure to fire situations. The behavior of concrete at high temperatures will be different in the mentioned cases. Moisture and the amount of concrete that was already hydrated influence both hot and residual strength but also the re-curing rate. In [115,116], concrete that was exposed to an elevated temperature at an early age (from 1 to 28 days after casting) was tested. Residual strength tests showed that young concrete had a better recovery rate (with the exception of 1-day-old concrete, as it did not have enough strength to withstand high temperature, and damage during the heating period was considerable. In [117], the strain-hardening cementitious composite was tested, and a similar conclusion was reached.

#### **10. Load Level at Exposure**

The level of load at exposure is an important issue that needs to be addressed, as every building is constantly subjected to loads. Strength at high temperature is positively affected by load level, as it reduces the speed of decrease of strength [118–120]. In [121], the influence of preload on residual strength was analyzed. There were three preload levels (0, 20, and 40% of the ultimate load at room temperature), and after exposure to high temperature and cooling, compressive tests were performed. The results showed that preload results in a higher residual strength (for 20% of preload increase, it is approximately 15%). This can be attributed to the restriction of thermal expansion by acting on the load, thus minimizing concrete damage. In [122], free expansion deformation of unstressed specimens exposed to high temperature at different heating rates was investigated. The linear expansion rate (LER) measured at high temperature was a linear function of temperature and did not depend on the heating rate. Restraining expansion and thus minimalizing internal cracking can greatly benefit residual strength. Thus, compressive stress plays a positive role.

#### **11. Heat Accumulation Factor**

The hot and residual strengths of concrete depend on the dehydration of the cement gel. Dehydration is mainly related to high exposure to heat, both the peak temperature and the exposure time. The factor that evaluates high-temperature exposure is the heat accumulation factor. It is defined as the area under the temperature–time curve. This idea was proposed in [46], and various studies [60,123–125] show that cement paste decomposition begins when the temperature exceeds 400 ◦C. Thus, the heat accumulation factor influencing the strength of the concrete should take into account only temperatures above

400 ◦C. Exposure to 200 ◦C will damage the concrete to an incomparably smaller degree than short exposure to 400 ◦C.

The heat accumulation factor can be calculated in two ways (Figure 22), as originally proposed in [46] by the use of Equation (1)  *H*<sup>400</sup> and by the method proposed by the authors (*H*400*r*):

$$H^{400} = \int\_{tr}^{ts} T(t)dt\tag{1}$$

.

$$H^{400r} = \int\_{tr}^{ts} (T(t) - 400) \, dt. \tag{2}$$

**Figure 22.** Diagram of the calculation of the heat accumulation factor, *H*<sup>400</sup> and *H*400*<sup>r</sup>*

A summary of the existing data is presented in Figures 23–30. Heating and cooling rates were treated as constant to simplify the calculation. It is visible that *H*<sup>400</sup> follows a certain decreasing line (Figures 23 and 25), which was easy to predict, but the scatter is big, and there is no easy way to generalize the results. *The H*400*<sup>r</sup>* is less scattered, but it is still hard to derive a solid function (Figures 24 and 26). A comparison between *H*<sup>400</sup> and *H*400*<sup>r</sup>* is presented in Figures 27 and 28 for NSC and Figures 29 and 30 for HSC. The three correlation factors presented in Table 16 demonstrate the superiority of the modified heat accumulation factor (Equation (2)). The values of the factors suggest a nonlinear relationship.


**Table 16.** Correlation factors between heat accumulation factors and relative residual strength based on data presented by various authors.

Thus, nonlinear exponential fitting was made. The R<sup>2</sup> factor (not deciding for nonlinear regression) was again higher in the *H*400*<sup>r</sup>* variation. However, the data of NSC and HSC behave similarly. With this in mind, for future reference, NSC and HSC can be treated identically, and separation is unnecessary.

**Figure 23.** *H*<sup>400</sup> coefficient for NSC by different authors.

**Figure 24.** *H*400*<sup>r</sup>* coefficient for NSC by different authors.

**Figure 25.** *H*<sup>400</sup> coefficient for HSC by different authors.

**Figure 26.** *H*400*<sup>r</sup>* coefficient for HSC by different authors.

**Figure 27.** *H*<sup>400</sup> coefficient for NSC—summary and fitting.

**Figure 28.** *H*400*<sup>r</sup>* coefficient for NSC—summary and fitting.

**Figure 29.** *H*<sup>400</sup> coefficient for HSC—summary and fitting.

**Figure 30.** *H*400*<sup>r</sup>* coefficient for HSC—summary and fitting.

#### **12. Discussion**

Although experimental research on the residual strength of concrete is extensive, the results appear to be incomplete. The main factors influencing residual strength (peak temperature, heating time, cooling regime, post-heating re-curing, load level at exposure) are already identified. Other factors, but with less pronounced effects (heating rate, type of aggregate, cement type, and dosage) have also been examined. However, some factors do not contribute to the strength, such as common additives, or the influence is not straightforward, such as porosity. However, with this in mind, experimental research can be limited to the temperature range 300–700 ◦C because the residual strength for temperatures up to 300 ◦C and more than 750 ◦C can be obtained only on the basis of the peak temperature. Moreover, the limit of two hours of high-temperature exposure can be used, as the main strength loss occurs within this period, and later, the impact is minimal.

Both essential factors influencing the residual strength (i.e., peak temperature and heating time) exhibit a negative linear influence on the residual strength, while for the other factors, the influence is not so straightforward, as the correlation factors are smaller. Thus, nonlinear functions would properly govern these relationships. Furthermore, the influence of the peak temperature on the residual strength is not as direct as on the hot concrete strength. Only the strength directly after cooling is highly dependent on the peak temperature reached. A similar conclusion is drawn for the cooling regime; e.g., for water cooling, its negative impact is apparent in the early stage of recurring but for the longer re-curing time, the impact diminishes. The influence of rapid cooling is also limited to the external layers, and it is not deciding for whole structure load capacity.

What should be stressed is the ability of concrete to regain its strength due to the rehydration of cement paste. Therefore, re-curing time and type are crucial for assessing the residual strength of concrete. From this perspective, it is not surprising that the influence of the w/c ratio can be omitted. However, there are reports where a lower w/c ratio results in a higher residual strength. This problem may be posed by cement type, as there is not enough comparative research focusing on the type of cement and its influence on residual strength. There exists also an indirect relationship between rehydration and porosity that requires further research. Porosity is notably higher after high-temperature exposure, and reversal is possible if the temperature was not greater than 800 ◦C. Moreover, lower porosity results in higher residual strength directly after cooling, but after re-curing, this difference is equalized.

The time after cooling and its type also regulate the residual strength. Directly after water cooling, a lower residual strength is obtained. Nevertheless, the re-curing significantly reduces the influence of cooling type. However, additional consideration of these phenomena is necessary for the external part of an element that is severely damaged. The internal part, which is crucial from a residual strength point of view, is immune to damages caused by different types of cooling. The interaction between layers of heated and cooled concrete should also be studied. The transient temperature field and associated strain and stress can contribute to material damage and strength reduction. Moreover, the possible volume changes due to chemical reactions and thermal expansion would also be considered.

There is also a need for more research on the influence of PP fibers on regaining residual strength. Changes in concrete structure left by PP fibers are evident, and their impact needs to be assessed. More research is also needed to determine the relation between preload level and residual strength, as it can be important in practice.

Since the ability of concrete to regain its strength due to rehydration determines the residual strength, only temperatures above 400 ◦C should be taken into account, because the decomposition of cement paste starts at this temperature. The modified heat accumulation factor *H*400*<sup>r</sup>* gives more coherent results than the unmodified *H*<sup>400</sup> as demonstrated by correlation factors, and therefore, it can be used to assess the residual strength of concrete.
