Post-fire performance of steel reinforced concrete columns with different steel sections

Abstract

This study investigates the post-fire performance of steel reinforced concrete composite (SRC) columns used in building construction using experimental compression tests. Four conventional steel sections as I-shaped, Cross, Box, and Plus were considered as the steel core. For fire loading, all fabricated columns were subjected to five target temperatures, including 25°, 250°, 500°, 700°, and 900°C, and then the cooling phase of the columns was done under natural conditions. Then using pressure jacks, the compressive behavior including strength and axial deformation of each of them was measured, and force-deformation curves were plotted and investigated for each SRC column. The results of these tests showed that the compressive strength and the elasticity modulus of the columns decrease at higher temperatures. Also, the effect of the steel core was examined on the compressive strength. Among the tested sections, SRC columns with the Box-steel core showed a more recovery in its compressive strength and elasticity modulus, and hence, its performance was better than SRC columns with the other steel cores. Moreover, SRC columns with the Plus-steel core indicated the weakest compressive strength and elasticity modulus. Finally, some equations were proposed for the prediction of the compressive strength and elasticity modulus of SRC columns at different temperatures by applying gene expression programming (GEP modeling) to the results.

Keywords

Steel reinforced concrete columns compressive strength post-fire performance axial deformation gene expression programming experimental study

Introduction

In building construction, the steel reinforced concrete column (SRC) has been widely used in medium and high-rise buildings during recent years. Steel reinforced concrete column is a composite, consisting of steel cores (or steel sections), reinforcing bars, and concrete. This type of composite element is able to combine the advantages of both steel and concrete, and the structural performance will be improved accordingly. Since the concrete has a much lower heat conduction coefficient, effectively retards and reduces the rise of temperature in the steel sections. The peripheral concrete acts as a protective layer for the steel section under fire. Hence, it has a high fire resistance compared with the element without the concrete layer. A typical SRC column with different parts is shown in Figure 1.

Figure 1.

Different parts of a typical steel reinforced concrete column.

Due to the high construction speed of composite columns, their use in buildings, especially in high-rise buildings, has led researchers to study various aspects of its performance in recent years (Morino et al., 1984; Guo et al., 2009; Zhou et al., 2015). One of the constant concerns in buildings, especially in high-rise buildings, is the conflagration because it is difficult to access fire and extinguish it in tall structures, especially in the dense urban fabric.

So far various research studies have been carried out on steel and concrete structures and their performance (Ketabdari et al., 2019; Lou et al, 2016, 2018; Li et al., 2019; Pachideh and Gholhaki, 2020; Saedi Daryan and Yahyai, 2009a, 2009b, Saedi Daryan and Ketabdari, 2019;). In recent years, due to the importance of SRC columns in construction, research studies have been conducted on the performance of these columns at high temperatures. In this respect, various SRC columns have been studied, among which can be referred to composite columns with I-section, Box-section, Cross-section, and Plus-section steel cores. In 2006, a method for heat transfer analysis of SRC columns with I-shaped cores was proposed by (Wang and Tan, 2006). In this method, using Eurocode 3 and a one-dimensional heat transfer model, temperature responses were obtained at each point of the steel profile via equations. Yu et al. (2007) tested the temperature distribution and ultimate strength of SRC columns under the fire, in which 12 columns were tested at different temperatures, and finally by numerical modeling and generalization of the results, some equations for predicting the ultimate strength of SRC columns at different temperatures were provided. Huang et al. (2008) investigated the effect of dimensions of sections and load levels on the strength of SRC columns with I-steel cores at elevated temperatures. In this study, the dimensions of 250*250, 300*300, 350*350, 400*400 (mm²), and the load levels between 0.2 and 0.5 were used. The results of this study show that columns with smaller sections have no fire resistance and increased column sections increase the fire resistance. Mao and Kodur (2011) have investigated fire resistance of SRC columns under three-side and four-side standard heating. In this test, the steel core of some columns was eccentrically loaded to account for its impact on column performance. The results of these tests show that the eccentricity and the load ratio have a great impact on the fire resistance of these columns. In another research, Young and Ellobody (2011) presented a finite element model investigating the behavior of axially restrained I-shaped core columns at elevated temperatures. Han et al. (2014) experimentally and numerically investigated steel reinforced concrete composite columns under fire conditions in which four SRC columns with H-shaped and Cross-shaped sections were tested, and a finite element model has been applied to complement the results and finally the strength of these columns after being exposed to fire conditions has been evaluated. In another study, Han et al. (2016) experimentally and numerically investigated the post-fire steel reinforced concrete composite columns. In this study, flexural and compression failures were observed in the specimens. Also, peak load values of the calculated columns after exposure to high temperatures are about 26% lower than those at ambient temperature. Temperature distribution, load–axial deformation relation, and load distribution during the loading and temperature rise phase are among the other achievements of this research. The post-fire seismic analysis of the I-shaped SRC columns was conducted by Guangyong and Dongming (2017). In this experimental research, eight SRC columns were tested in which the length of the plastic hinge zone increases with increasing temperature. The results of this test indicate the appropriate energy absorption of these columns at high temperatures.

Given the above research on the SRC columns performance in fire, there are common limitations in this regard. The most important of these limitations include: (a) they often have considered the I-steel core and have not considered other types of steel cores; (b) the evaluation of the compressive strength of the columns has not been performed at very high temperatures (about 900°C); (c) limited number of tests has been performed in these studies; and (d) the cooling phase in the applied thermal regimes has not been considered in these research studies. For this purpose and to cover these limitations, this research experimentally investigates the compressive strength of SRC columns with different steel sections as the column core after exposing to different temperatures and cooling them. In this regard, four different I-section, Plus-section, Box-section, and Cross-section steel cores are considered as the core of concrete columns and each of these columns is exposed to five target temperatures of 25, 250, 500, 700, and 900° C, and the cooling is done naturally. In the following, the compressive strength of each specimen is measured experimentally, and the force-axial deformation curves for each specimen are presented. In the next phase, using the values obtained from the test results and applying the genetic algorithm, some equations are proposed for the prediction of the reduction factor of “the compressive strength and the elasticity modulus” of the SRC columns.

Experimental investigation

Specifications of the test specimens

To compare the impact of steel cores on the performance of SRC columns after fire exposure, four different steel sections including I- Section, Plus-Section, Box-Section, and Cross-Section are considered as cores of concrete columns. A view of the sections with column dimensions is shown in Figure 2. It is worth mentioning that the height of all columns is 550 mm. It should be noted that the dimensions shown in Figure 2 are nominal. The measured dimensions have about 2 mm of tolerance as shown in Table 1.

Figure 2.

Steel reinforced concrete columns with different steel sections (unit of measurement of dimensions in millimeter). (a) Column dimensions and reinforcement detailing for a typical column. (b) Parametric dimensions of column sections (used in Table 1).

Table 1.

Nominal and average measured values for the sets (unit of measurement of dimensions is millimeter).

Section	Nominal						Measured
Section	B	b	t	h	S	R	B	b	t	h	S	R
I-shaped	125	40	4	50	4	5	126.5	40.4	4	49.5	4	5
Plus	125	60	4	30	4	5	125.7	61	4	31.5	4	5
Box	125	30	4	30	4	5	126	31	4	30	4	5
Cross	125	30	3	50	3	5	124.5	29.7	4	49.5	3	5

In this research, the same and fixed mix design for the preparation of the required concrete was used for the fabrication of the specimens, as presented in Table 2.

Table 2.

Columns concrete mix design (Kg/m³).

Material	Gravel	Sand	Stone powder	Water	Cement	Micro silica
Content(kg)	666	1087	100	189.5	340	17

The SRC columns with the four steel cores of I-Shaped, Cross, Box, and Plus shapes are considered. Then each of them, including four sets, is exposed to five target temperatures of 25, 250, 500, 700, and 900°C. Given the importance of the issue, specimens were designed according to AISC360-16- standard, as explained in Compressive Strength in AISC Code. Figure 3 illustrates the different stages of the fabrication of the specimens. The number of fabricated specimens is presented in Table 3.

Figure 3.

Specimen’s fabricating. (a) Rebars and stirrups, (b) four fabricated steel sections, (c) molding, and (d) some fabricated specimens.

Table 3.

Number of fabricated specimens.

Type of steel section	I-shape	Cross	Box	Plus
Number of fabricated specimens in any set	5	5	5	5

One of the points that were taken into consideration during the design was the uniformity of material consumption amount (representing the cost of construction) and the nearly identical axial pressure capacity of the column. Table 4 shows the amount of steel, rebar, and concrete consumed per meter of these specimens.

Table 4.

Amount of materials per unit length of the specimen.

Material	I-shape	Cross	Box	Plus
Concrete (cm³)	14,791	14,801	14,831	14,831
Rebar (cm³)	314	314	314	314
Plate (cm³)	520	510	480	480

In order to present the specifications of the steel and reinforcement bars used in this study, tensile tests on steel plates and rebars were performed, and stress–strain curves of which are presented in Figure 4(a) and (b).

Figure 4.

Tensile test results. (a) Steel plates and (b) rebar.

Based on the results, a summary of the specifications of steel plates and rebars is presented in Table 5.

Table 5.

Steel plates and rebar properties.

Material	Elasticity modulus (MPa)	Yield strength (MPa)	Ultimate strength (MPa)
Steel plate	205,000	325	433
Steel rebar	205,000	405	600

Heating–cooling regime and loading patterns

An electric furnace with manual power adjustment is used to heat the columns (Figure 5). Figure 6 shows a picture of the furnace. Two K-type thermocouples on the specimen and furnace recorded specimen and furnace temperatures. As shown theoretically in Figure 6, the columns were heated from ambient temperature to 4 target temperatures (T_u) of 250, 400, 500, 700, and 900 at a rate of 10° per minute. Once the specimen was heated to the target temperature, the temperature was kept constant for 60 min to prepare a stable heating condition for the specimens and to ensure a uniform temperature distribution in the columns. Then the furnace is switched off and the specimen cooling phase is started. The specimens were cooled naturally at an average rate of about 2–10° C to reach room temperature. Figure 7 presents the actual temperature–time curve in which the heating and cooling phases for the target temperatures of 700 and 900° C are measured for each SRC composite column. Upon completion of the cooling phase, the compressive strength of each column was measured by placing it between the pressure jack. In Figure 8, the test setup and an example of a column being placed under a pressure jack are shown. Once the specimens have been placed, the load is applied and continues until failure.

Figure 5.

Electrical furnace.

Figure 6.

Theoretical heating–cooling cycle diagram.

Figure 7.

Measured curve of the heating–cooling cycle diagram.

Figure 8.

Column under the pressure jack and test setup (a) test setup and (b) laboratory test.

Results and discussion

Compressive strength result

According to the descriptions given in the previous sections, this section first studies the appearance of the specimens and the occurrence of explosive spalling during heating and then describes the compressive strength tests on the specimens. The explosive spalling phenomenon occurs in concrete columns exposed to fire due to various reasons such as pore pressure, thermal stress, or both, and factors such as the type of materials and additives, the dimensions of the sections, and the moisture content have a significant impact on its occurrence. No explosive spalling was observed in the heated specimens, and only scaling was observed in the corner spalling of some specimens. Among the reasons, we can refer to control of the specimen’s moisture content in accordance with the recommendation (Eurocode 2, 2005), use of appropriate additives in the mix design, selection of the appropriate dimensions of the sections, control of the heating rate of the specimens, and failure to apply axial load during the heating. Therefore, the avoidance of spalling in the specimens has resulted in the proper conditions for performing compressive strength tests and comparing the compressive strength as well as the axial deformation between different specimens.

Figure 9 shows a comparison of the compressive strength of SRC columns with different steel cores at different temperatures. According to the results, the values of the compressive strength of the columns at the ambient temperature are almost similar, and this similarity is also true up to 250° C. As the temperature increases, the compressive strength of the SRC columns with different steel cores follows a decreasing trend, and the impacts of such temperature rise on all columns strongly increase at temperatures above 700°C. The reason is the evaporation and cracking with respect to the non-homogeneous and distinctive expansion and contraction of concrete. Among the specimens, the column with the Box-shaped core showed better performance at temperatures above 700° C due to the confinement of the concrete core.

Figure 9.

Comparison of compressive strength of steel reinforced concrete columns with steel cores at different temperatures.

Figure 10 also presents force–axial deformation curves of the columns and their performance at different temperatures. The axial deformation of the columns with different cores had a similar performance in such a manner that the strain corresponding to the ultimate strength increases by the temperature rise, but the final strain of the columns did not change so much by temperature rise.

Figure 10.

Axial–force deformation diagram.

Compressive Strength in AISC Code

Furthermore, given the specifications of the materials used to fabricate the SRC columns in this study, the design compressive strength values of the axial members can be obtained by applying the equations of the AISC360-10 standard and using equation (1)

\begin{array}{l} I f \frac{P_{n 0}}{P_{e}} \leq 2.25 \to P_{n} = P_{n 0} [{0.658}^{P_{n 0} / P_{e}}] \\ Else \to P_{n} = 0.877 P_{e} \end{array}

(1)

Where

P_{e} = π^{2} \frac{E I_{eff}}{{(K L)}^{2}}

(2)

P_{n 0} = F_{y} A_{s} + F_{y s r} A_{s r} + 0.85 f_{c}^{'} A_{c}

(3)

E I_{f f} = (E_{s} I_{s 33}) + (0.5 E_{s} I_{s r}) + (C_{1} E_{c} I_{c})

(4)

C_{1} = \min (0.25 + 3 [\frac{A_{s} + A_{s r}}{A_{g}}], 0.7)

(5)

E_{c} = 0.43 {(W_{c})}^{1.5} \sqrt{f_{c}^{'}}

(6)

Where P_e, P_n0, A_g, A_s, A_sr, E_c, E_s, F_y, F_ysr, I_c, I_s, I_sr, K, E_s, L, f_c, and W_c are taken as buckling limit state, nominal compressive strength, gross area of composite member, area of the steel section, area of continuous reinforcing bars, modulus of elasticity of concrete, modulus of elasticity of steel, the yield stress of steel section, the yield stress of the longitudinal rebars, the moment of inertia of the concrete section, moment of inertia of the steel section, moment of inertia of the longitudinal bar, effective length factor of the axial member, laterally unbraced length of the axial member, and specified compressive strength of concrete and weight of concrete, respectively.

Comparison of the standard values (P_n0) and laboratory values (F_c) of the compressive strength of SRC columns at ambient temperature are presented in Table 6. The proximity of the table values indicates the accuracy of the laboratory conditions and performance of the tests. The purpose of calculating the compressive strength of the specimens according to the AISC is to compare the values obtained from the test with those obtained from the code formulas. This comparison was made to validate the test setup, test method, and test equipment at ambient temperature.

Table 6.

Compressive strength values of the steel reinforced concrete columns obtained through the regulation at ambient temperature.

Compression strength	I-shape	Cross	Box	Plus
P_n0 (kN)	586.2	584.4	577.7	577.5
F_c (kN)	635	625	600	627

Failure modes

The reduction factors for the compressive strength and the modulus elasticity of the various columns calculated using equations (7) and (8) are presented in Figure 11. As it is shown, with increasing temperature, the Plus-shaped section loses about 70% of its compressive strength, which has the worst performance among the sections. The Cross-shaped section with a 68% reduction in strength and the I-shaped section with a 65% reduction in compressive strength had a performance similar to the Cross-shaped sections. But the performance of the Box-shaped section is different from other sections, and by recovering about 45% of the compressive strength at 900° C, it shows a performance much better than other sections at high temperature. Investigation of the reduction factors for modulus of elasticity shows that as the temperature increases, the reduction slope of modulus of elasticity is much greater than that of the compressive strength, so that at 700° C more than half of the elasticity modulus of all four columns is lost, and at temperatures above 700° C, this decrease occurs more strongly.

R_{f_{c, SRC}} = \frac{f_{c, p f}^{'}}{f_{c}^{'}}

(7)

R_{E_{c, SRC}} = \frac{E_{c, p f}}{E_{c}}

(8)

Figure 11.

Reduction factor–temperature curve for steel reinforced concrete columns (a) compressive strength and (b) elasticity modulus.

One of the important factors in analyzing the behavior of composite columns during a fire is to determine the behavioral modes of the specimens. One of the failure modes in the columns is the failure due to buckling. This failure mode occurs in the slender columns, so if the slenderness ratios of the columns are high, this failure mode will be observed. In this study, the columns had a low slenderness ratio and their failure mode at various temperatures has not occurred as buckling mode, and it is often observed due to crushing of concrete and yield of steel core (compressive failure mode). Equations (9) and Table 7 show the slenderness ratios of the specimens. As it is seen, all specimens are fat and therefore their main failure mode is the compressive failure mode. Figure 12 shows pictures of each SRC column at different temperatures after failure. As expected, the columns reached their maximum capacity according to the compressive failure mode

λ = \frac{k l}{r}

(9)

Where K, L, and r are taken as the effective length factoring, the unsupported length of the column, and the radius of gyration of the cross-section, respectively.

Table 7.

Slenderness ratios of composite columns.

Section type	Slenderness coefficient
I-shape	16.77
Box	16.52
Cross	16.67
Plus	16.57

Figure 12.

Composite columns with different steel cores after failure.

According to the observations, the stability of Box-core SRC columns is more than that of the other specimens. The reason can be investigated from three-point of view. First, confinement of the concrete inside the Box-shaped section increases the strength of the column and its stability. Second, since the plates of the Box-shaped section are restrained at both ends, local buckling does not occur therein, and consequently their strength increases. However, due to local buckling in section components of the other three sections, the column fails sooner. Last, since the Box-shaped section is closed, one-sided heating of its plates affects, and the impact of heating on it is less. But the core components of the other three specimens are exposed to heating on both sides.

Application of the metaheuristic algorithms for predicting

Background of metaheuristic algorithms’ for predicting the performance of SRC columns under fire

The ability of metaheuristic algorithms and artificial intelligence to establish relationships of different levels of complexity between different input parameters has led to use in structural fire engineering in recent years. One of the pioneers in the use of artificial intelligence and metaheuristic algorithms is Chan et al. (1998) who predicted the compressive strength of concrete under fire conditions using a neural network. In recent years, most of the relationships presented in structural fire engineering have been done using artificial neural networks among which can be referred to as semi-rigid flexible end-plates and flush end-plate joints at elevated temperatures (Al-Jabri and Al-Alawi, 2007), the behavior of steel frames under fire conditions (Hozjan et al., 2007), prediction of the flexural capacity of concrete slabs under fire conditions (Erdem, 2009), prediction of temperature rise in tubular steel trusses under fire conditions (Xu et al., 2013), the prediction of the behavior of the welded angle connections at high temperatures (Saedi Daryan and Yahyai, 2018), and the presentation of temperature-dependent material models for steel structures (Naser, 2018). A combination of genetic algorithm and neural network is also used to predict the flexural capacity of steel beams exposed to fire (Zhao, 2006) and to evaluate wood structures under fire conditions (Naser, 2019).

The ability of the genetic algorithm is to present high-precision relationships and nonuse of them in research studies on SRC columns. The lack of reliable and applicable relationships for the performance of these columns under fire conditions as well as the use of linear regressions and the presentation of low-precision relationships in the previous research studies have led to the use of one of the powerful subdivisions of the genetic algorithm called gene expression programming (GEP). The modeling method and results are provided below.

GEP modeling process

The genetic algorithm used by many researchers in recent years has several subdivisions, and one of the most important and powerful subdivisions is the gene expression programming. The summarized GEP modeling process is provided in the following steps:

Laboratory data gathering.

Generation of the initial answers using the data obtained and probabilistic rules.

Determination of objective function and fitness function. In this study, equations (10) and (11) are assumed as objective function and fitness function

Objective Function = (\frac{n_{L} + n_{V}}{n_{T}} \times \frac{MA E_{L} + RMS E_{L}}{1 + R_{V}}) + (\frac{2 n_{V}}{n_{T}} \times \frac{MA E_{V} + RMS E_{V}}{1 + R_{V}})

(10)

Fitness function = (\frac{1}{1 + RMSE}) \times 1000

(11)

Where, n_L, n_V, and n_T are the data count for learning, validation, and training, respectively

RMSE = \sqrt{\frac{\sum_{i = 1}^{n} (h_{i}^{2} - t_{i}^{2})}{n}}

(12)

MAE = \frac{\sum_{i = 1}^{n} | h_{i} - t_{i} |}{n}

(13)

R = \frac{\sum_{i = 1}^{n} (h_{i} - {\bar{h}}_{i}) (t_{i} - {\bar{t}}_{i})}{\sqrt{\sum_{i = 1}^{n} {(h_{i} - {\bar{h}}_{i})}^{2} \sum_{i = 1}^{n} {(t_{i} - {\bar{t}}_{i})}^{2}}}

(14)

where, hi and ti are the ith experimental data and the obtained output from the ith numerical modeling, hi and I are the average of the experimental data and the obtained output from the ith numerical modeling, and n is the specimen count.

3. Determining the response fitness rate

4. Implementing genetic operators on the response

To define these operators, the GEP modeling language, named the Karva language is provided in Figure 13(a). GEP is developed based on two essential elements: chromosomes and expression trees (ETs). The chromosomes contain one or more genes (Beheshti Aval et al., 2017; Ketabdari et al., 2020; Ketabdari et al., 2021). As observed in Figure 13(a), the genes are formed in the two sections of head and tail which include the mathematical equations, variables, and constants. The expression tree and its related mathematical equations are shown in Figure 13(a). The genetic operators consist of crossover and mutation, expressed in Figure 13(b) and (c), respectively.

5. Generating new respond.

6. Determining the fitness of the responses.

7. Repeating stages 5–7 until algorithm ending conditions are met.

8. Measuring the accuracy of the equations by comparing their error rates.

Figure 13.

Gene expression programming modeling details and operators (a) expression tree, (b) mutation, and (c) crossover.

Propose prediction equations

By modeling and adjusting the genetic parameters, according to the content of Table 8, the equations related to compressive strength and elasticity modulus reduction factors in the SRC column at different temperatures are presented in Table 9.

Table 8.

Adjusted parameters and their operators' rate in modeling.

Parameter	Adjustment
Chromosome	50-100-150
Gene	1-2-3-4
Head size	3-5-8-10
Tail size	4-6-9-11
Function’s connection	+
Mutation rate	0.1
Gene replacement rate	0.15
Single point convergence rate	0.1
Double point convergence rate	0.15
Uniform point convergence rate	0.1
Gene point convergence rate	0.15

Table 9.

Proposed equations.

Type of column	Equation
I-shape	$R_{E_{SRC}} = \sqrt[3]{6.93 - \sqrt[3]{T / 3.75}} - 0.65$	(15)
I-shape	$R_{f_{c, SRC}} = \sqrt[3]{\sqrt[3]{\sqrt[3]{3} + 4.477} - (1 / 4.06 \times \sqrt[3]{T})}$	(16)
Cross shape	$R_{E_{SRC}} = \sqrt[3]{\sqrt{T} - 0.028 T - 3.21} \times \sqrt[3]{22 / T}$	(17)
Cross shape	$R_{f_{c, SRC}} = \sqrt[3]{\sqrt[3]{- 50.5 / T + 7.4} - (\sqrt[3]{T + 31.4} / 5.14)}$	(18)
Box	$R_{E_{SRC}} = \sqrt[3]{9.54 - \sqrt[3]{T} / 7}$	(19)
Box	$R_{f_{c, SRC}} = {(\sqrt{T} - 8.51)}^{3} / - 20.6 T + 0.96$	(20)
Plus	$R_{E_{SRC}} = \sqrt{{(- 7.42 / T)}^{3} + 1 + 0.0008 T - 0.034}$	(21)
Plus	$R_{f_{c, SRC}} = \sqrt{1 - T / {10.06}^{3}}$	(22)

These equations have appropriate and acceptable accuracy. That attributed to GEP’s ability in establishing the correlation between (a) complex systems, (b) complete search among all response spaces, and (c) applying probability laws and not being trapped in local optimized zones during modeling. Table 10 is provided to compare the measured values with the proposed formulas in Table 9.

Table 10.

Comparison of the measured values with the proposed model.

Temperature, °C	Measured				Proposed method
Temperature, °C	I-shape	Plus-shape	Box-shape	Cross-shape	I-shape	Plus-shape	Box-shape	Cross-shape
25	65	64	61	64	69	64	64	63
250	54	57	53	58	56	55	54	56
500	39	47	48	47	46	46	43	46
700	35	37	46	39	37	37	34	36
900	23	21	26	20	22	22	26	22

As observed the functionality of specimens is almost similar, proposed equations are applicable in all SRC columns. To determine the accuracy of the proposed equations, both the statistical correlation coefficient and relative error are calculated, as shown in Table 11.

Table 11.

Correlation coefficient and relative error of the proposed equation.

Equation	Type of column	Relative error (%)	Correlation coefficient
GEP for compressive strength	I-shape	9	0.91
	Cross-shape	8	0.93
	Box-shape	7.5	0.93
	Plus-shape	10	0.90
GEP for elasticity modulus	I-shape	7	0.94
	Cross-shape	10	0.91
	Box-shape	10	0.90
	Plus-shape	9	0.92

Note: GEP: gene expression programming.

Conclusions

Given the importance of the compressive strength as one of the effective parameters in the design and analysis of structures, this study investigates the impact of the fire on the compressive strength of SRC columns. For this purpose, four sections including I-shaped, Cross-, Box-, and Plus-shaped sections were considered as the column core to investigate the impact of the steel core on the compressive strength of the SRC columns at elevated temperatures. After constructing the columns, five target temperatures were considered for each column. The columns were exposed to 4-sided heating in an electric furnace, and the cooling phase was done naturally, and the temperature of each specimen was lowered to the ambient temperature. Finally, the compressive strength and axial deformation of each specimen were measured. The summarized results of these experiments are presented as follows:

The compressive strength and the elasticity modulus of the SRC columns with different steel cores reduce with increasing temperature, and the rate of this reduction increases sharply at temperatures above 700° C.

By comparing the performance of the steel cores, the Box-shaped section with 45% recovery of the compressive strength and 25% recovery of the elasticity modulus at 900° C showed a better performance than the other sections. The reason can be originated from three-point of view. First, confinement of the concrete inside the Box-shaped section increases the strength of the column. Second, since the plates of the Box-shaped section are restrained at both ends, the local buckling does not occur, and consequently their strength increases. Last, since the Box-shaped section is closed, only the one side of its plates affects the heating, and the impact of heating on it is less. However, the core component of the other three specimens is exposed to heating on their both sides.

Plus-section SRC columns with 30% recovery of the compressive strength and 18% recovery of the elasticity modulus at 900° C also showed the weakest performance among the four steel cores.

The strains of the SRC columns with different steel sections at different temperatures have almost a similar performance to each other.

Some equations are proposed to predict the compressive strength and the elasticity modulus of the SRC columns at high temperatures using the obtained experimental results and applying the GEP modeling. The values of correlation coefficients and relative errors of the relationships indicate their applicability in structural fire engineering.

Footnotes

Acknowledgment

The research work presented in this study was done in the Laboratory of Structural and Earthquake of Shahid Beheshti University. The authors would like to express gratitude to the laboratory manager who prepared materials and equipment for the experimental test to proceed this study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Mohammad Safi

Amir Saedi Daryan

References

Al-Jabri

Al-Alawi

(2007) Predicting the behaviour of semi-rigid joints in fire using an artificial neural network. Steel Structures 7: 209–217.

ANSI

(2016) AISC 360-16, Specification for Structural Steel Buildings. Chicago: AISC.

Beheshti Aval

Ketabdari

Asil Gharebaghi

(2017) Estimating shear strength of short rectangular reinforced concrete columns using nonlinear regression and gene expression programming. Structures 12: 13–23.

Chan

Jin

Anson

, et al. (1998) Fire resistance of concrete: prediction using artificial neural networks. Magazine of Concrete Research 50(4): 353–358.

Erdem

(2009) Prediction of the moment capacity of reinforced concrete slabs in fire using artificial neural networks. Advances in Engineering Software 41: 270–276.

Guangyong

Dongming

(2017) Experimental research on the post-fire seismic performance of steel reinforced concrete columns. Procedia engineering 210: 456–463.

Guo

Zhang

Huang

, et al. (2009) Experimental study on hysteretic model of SRC columns [J]. Journal of Earthquake Engineering and Engineering Vibration 5.

Han

Tan

Song

(2014) Fire performance of steel reinforced concrete columns. Journal of Structural Engineering 141(4): 04014128.

Han

L-H

Zhou

Tan

Q-H

, et al. (2016) Performance of steel-reinforced concrete column after exposure to fire: FEA model and experiments. Journal of Structural Engineering 142(9): 04016055.

10.

Hozjan

Turk

Srpčič

(2007) Fire analysis of steel frames with the use of artificial neural networks. Journal of Constructional Steel Research 63: 1396–1403.

11.

Huang

Z-F

Tan

K-H

Toh

W-S

, et al. (2008) Fire resistance of composite columns with embedded I-section steel - Effects of section size and load level. Journal of Constructional Steel Research 64(3): 312–325.

12.

Ketabdari

Saedi Daryan

Hassani

(2019) Predicting post-fire mechanical properties of grade 8.8 and 10.9 steel bolts. Journal of Constructional Steel Research 162: 105735.

13.

Ketabdari

Karimi

Rasouli

(2020) Shear strength prediction of short circular reinforced-concrete columns using soft computing methods. Advances in Structural Engineering 23(14): 3048–3061.

14.

Ketabdari

Saedi Daryan

Hassani

Safi

(2021). Fire response of steel connections using gene expression programming and finite element method. The Structural Design of Tall and Special Buildings 30(11): e1862.

15.

L-Z

Liu

J-T

, et al. (2019) Experimental study on seismic performance of post-fire reinforced concrete frames. Engineering Structures 179: 161–173.

16.

Lou

Wang

Jiang

, et al. (2018) Experimental and numerical study on thermal-structural behavior of steel portal frames in real fires. Fire Safety Journal 98: 48–62.

17.

Mao

Kodur

VKR

(2011) Fire resistance of concrete encased steel columns under 3- and 4-side standard heating. Journal of constructional steel Research 67(3): 270–280.

18.

Morino

Matsui

Watanabe

(1984) Strength of biaxially loaded SRC columns. In: Composite and mixed construction. Virginia, USA: ASCE, 185–194.

19.

Narayanan

Beeby

(2005) Designers' Guide to EN 1992-1-1 and EN 1992-12. Eurocode 2: Design of Concrete Structures: General Rules and Rules for Buildings and Structural Fire Design. London, UK: Thomas Telford, 230.

20.

Naser

(2018) Deriving temperature-dependent material models for structural steel through artificial intelligence. Construction and Building Materials 191: 56–68.

21.

Naser

(2019) Fire resistance evaluation through artificial intelligence-A case for timber structures. Fire Safety Journal 105: 1–12.

22.

Pachideh

Gholhaki

(2020) Assessment of post-heat behavior of cement mortar incorporating silica fume and granulated blast-furnace slag. Journal of Structural Fire Engineering 11: 221–246.

23.

Saedi Daryan

Bahrampoor

(2009) Behavior of Khorjini connections in fire. Fire safety journal 44: 659–664.

24.

Saedi Daryan

Ketabdari

(2019). Mechanical properties of steel bolts with different diameters after exposure to high temperatures. Journal of Materials in Civil Engineering 31(10): 04019221.

25.

Saedi Daryan

Yahyai

(2009a) Behavior of bolted top-seat angle connections in fire. Journal of Constructional Steel Research 65: 531–541.

26.

Saedi Daryan

Yahyai

(2009b) Behaviour of welded top-seat angle connections exposed to fire. Fire Safety Journal 44: 603–611.

27.

Saedi Daryan

Yahyai

(2018) Predicting the behavior of welded angle connections in fire using artificial neural network. Journal of Structural Fire Engineering 9(1): 28–52.

28.

Wang

Yuan

Huang

, et al. (2016) Experimental study on the fire behaviour of reinforced concrete slabs under combined uni-axial in-plane and out-of-plane loads. Engineering Structures 128: 316–332.

29.

Wang

Z-H

Tan

(2006) Residual area method for heat transfer analysis of concrete-encased I-sections in fire. Engineering Structures 28(3): 411–422.

30.

Zhao

Wang

, et al. (2013) Prediction of temperature of tubular truss under fire using artificial neural networks. Fire Safety Journal 56: 74–80.

31.

Young

Ellobody

(2011) Performance of axially restrained concrete encased steel composite columns at elevated temperatures. Engineering Structures 33(1): 245–254.

32.

J-T

Z-D

Xie

(2007) Nonlinear analysis of SRC columns subjected to fire. Fire Safety Journal 42(1): 1–10.

33.

Zhao

(2006) Steel columns under fire-a neural network based strength model. Advances in Engineering Software 37: 97–105.

34.

Zhou

Yan

Liu

(2015) Behavior of square tubed steel reinforced-concrete (SRC) columns under eccentric compression. Thin-Walled Structures 91: 129–138.