Mechanical Behavior of High-Strength Bolted Joints Fabricated from Fire-Resistant Steel at Elevated Temperatures

,


Introduction
Steel has been used extensively in construction, followed by issues of fre resistance and elevated temperature resistance of steel structures [1]. In the case of fre, the mechanical properties of ordinary steel (such as strength and elastic modulus) will degrade sharply with the increase in temperature [2]. When the external temperature exceeds 600°C, ordinary steel will lose most of its strength and stifness. For ordinary steel members without fre protection, the temperature can usually rise to more than 600°C within 20 minutes of the fre event. Terefore, scholars have carried out research and production of high-strength and fre-resistant steel, some of which have been applied in practical engineering.
As the main connecting means of a steel structure, highstrength bolts play an important role in the seismic properties of a steel structure. Elevated temperature results in additional stress caused by the thermal expansion of components [3], which reduces the bearing capacity of highstrength bolts and thus easily leads to connection failure. In steel structures, a premature connection failure has a negative impact on the overall performance of the structure and even brings about structural damage and collapse. Consequently, the mechanical properties of high-strength bolts at high temperatures have attracted academic attention. Lawson [4] conducted fre resistance experiments on beamcolumn joints with high-strength bolts and found that the temperature in the bolt is much lower than that in the bottom fange of the beam when the joint is damaged. Chen and Wang [5] investigate how connection details afect structural survivability at high temperatures based on parametric simulations. Teodorou [6] obtained that the strength of high-strength grade 8.8 bolts began to decline rapidly after 300°C, and at 600°C, its ultimate strength had decreased to 35% of the ultimate strength under normal temperature. Król and Wachowski [7] found that the temperature level of the fre, fre duration, and fre-fghting method signifcantly afect the bolts' mechanical properties. Lange and González [8], Lou et al. [9], and Zhu et al. [10] studied the mechanical properties, slip coefcient, and pretension force of high-strength grade 10.9 bolts at high temperatures. Hanus et al. [11] investigated the mechanical behavior of high-strength grade 8.8 bolts under the situation of "nature fre," which means a temperature history including both heating and cooling phases. Kodur et al. [12], Hu et al. [13], Lange and Kawohl [14], and Moreno and Baddoo [15] studied the performance deterioration and failure mechanisms of high-strength bolts made of diferent materials (such as A325, A490, and stainless steel [16]) under fre. Ohlund et al. [17] discussed the physical properties of ultra-high-strength bolts at elevated temperatures from the perspective of microstructure. Using both steady-state and transient-state fre tests, Wang et al. [18] studied the shear behavior of lapped connections bolted by thread-fxed oneside bolts (TOB) at elevated temperatures. Fang et al. [19][20][21] investigated the robustness of steel-composite structures subject to localized fre and proposed a practical framework for robustness assessment. Santiago et al. [22] and Nadjai et al. [23] conducted natural fre tests of steel frames, and they studied the infuence of connection types on failure modes and the infuence of travelling fre on surrounding structures. Maraveas et al. [24] proposed an improved temperature-dependent constitutive model for steel to account for the local instabilities of slender plates in a fre situation.
Te study on the mechanical properties of ordinary highstrength bolts under elevated temperatures has been extended from simple to complex stress states, and some highperformance (such as fre resistant) steels are beginning to get the attention of some scholars. Fire-resistant steel, by adding specifc alloying elements (such as Cr, Mo, and Nb), can meet strength requirements at high temperatures (usually refers to 600°C) for a certain period of time (usually refers to 1∼3 hours), thereby increasing the fre resistance ability and safety of buildings. In addition, fre-resistant steel can be used without spraying a freproof coating or surface treatment. Compared with ordinary steel that require spraying a freproof coating to achieve fre resistance, freresistant steel can avoid environmental issues caused by the coating peeling under long-term application. Sakumoto et al. [25] studied the high-temperature mechanical properties of high-strength fre-resistant bolts and the shear strength of friction joints using fre-resistant bolts. Ban et al. [26] compared the high-temperature mechanical properties of conventional high-strength grade 10.9 bolts and high-performance bolts produced by the Shougang Group Research Institute of Technology. Tey proposed constitutive models and prediction equations for describing the material properties of high-performance bolts at diferent temperatures. Li et al. [27,28] conducted the material test of 20 MnTiB high-strength grade 10.9 bolts from room temperature to 700°C. According to the testing results, they obtained calculation models used for reduction coefcients of yield strength and the initial elastic modulus of ordinary high-strength bolts and fre-resistant high-strength bolts at high temperatures. Meng et al. [29] investigated the shear performance of bolted joints fabricated with BFRW10 highstrength fre-resistant bolts. Tey found that the failure mode of bolted joints changes from compressive failure on the bolt hole to shear failure of the bolt screw with increasing temperature.
Current studies have brought signifcant advances in fnding the resistance to high temperatures of joints and connections fabricated with high-strength bolts. However, there is still a lack of research and relevant design methods on the mechanical properties of joints fabricated by freresistant steel and fre-resistant high-strength bolts under elevated temperatures or fre. Given the existing issues and lack of experimental data, this paper experimentally studies the mechanical performance of fre-resistant and highstrength bolted joints under the conditions of the room and elevated temperatures (20°C∼700°C). Calculated models under diferent reliability degrees are proposed to predict the strength degradation of bolted joints (including fre-resistant and ordinary steel specimens) under elevated temperatures, which can provide a reference for the application of freresistant high-strength bolts and fre-resistant steel in practical engineering. Figure 1 shows details of the specimen, which is the joint connected by high-strength bolts. Te steel plate of the joint is made of fre-resistant steel Q460 produced by Shougang Group. Te fre-resistant steel Q460 is manufactured through thermo-mechanical controlled processing (TMCP). Compared with ordinary steel processing, there are three points to be noted. (1) Alloying elements: it is necessary to properly add alloying elements (such as Cr and Mo), which are benefcial to the fre resistance of steel and strictly control the content of impurity elements (such as P and S). Te chemical composition of the fre-resistant steel Q460 is shown in Table 1. (2) Rolling temperature: the beginning and fnal rolling temperatures of the fre-resistant steel Q460 in the recrystallized zone are 1150°C and 1120°C, respectively, while the beginning and fnal rolling temperatures in the unrecrystallized zone are 950°C and 850°C, respectively. (3) Cooling condition: an automatic control system, accelerated cooling control (ACC), is used to cool the fre-resistant steel Q460. Te beginning and ending cooling temperatures are 780°C and 450°C, respectively, and the cooling rate is 20°C/s. After the heat preservation treatment at 600°C, Q460 is air-cooled.

Specimen Design.
Te steel plate has only one type, which is fre resistant. Table 2 shows the mechanical behavior of a fre-resistant steel plate. Te high-strength grade 10.9 bolts (M20) are used in the joints, which have two types, ordinary and fre-resistant, respectively. Table 3 shows the mechanical behavior of the fre-resistant and high-strength grade 10.9 bolt. Treatments of the friction surface between the steel plate and high-strength bolt have two types, which are impeller blasting and sprayed hard quartz sand. Both treatments are commonly used in practical engineering. Te mechanical properties of high-strength bolted joints at elevated temperatures are investigated. Considering the maximum temperature (600°C∼800°C) [23,30] of locally directly heated members of a steel structure in a natural fre and the optimal temperature interval of the testing machine, fve temperatures are chosen, which are 20°C (room temperature), 200°C, 400°C, 600°C, and 700°C. Table 4 shows the experimental condition and parameters of each specimen. Te experiment with each conditional parameter is repeated twice, and the total quantity of specimens is 40 (including 8 specimens at room temperature and 32 specimens at elevated temperatures).

Confguration and Measurements during
Testing. All 40 specimens were tested in the MTS high-temperature testing machine (see Figure 2), which has a loading capacity of 1000 kN in the Key Laboratory of Tongji University. Te steady-state test of elevated temperature is carried out. Te temperature is controlled by the electric-heating furnace, which is supported by intelligent temperature control equipment. Te electric-heating furnace can be controlled according to a given temperature curve. Te highest temperature can reach 1000°C, and the maximum heating rate is 30°C/min. Te testing machine is equipped with an extensometer to continuously measure the displacement of the specimen.

Experimental Procedure.
According to the current code [31,32] of fre resistance tests, the experimental procedure is as follows: (1) Installation of high-strength bolts: frst, for the convenience, high-strength bolts of the specimen are initially screwed with an ordinary wrench. Ten, we tighten it with a torque wrench to complete the installation. Te pretightening force of a highstrength bolt is controlled and recorded by the axialforce detector. Te pretightening force of a highstrength grade 10.9 bolt (M20) should be 155 kN.
(2) Te assembled specimen is installed in the electricheating furnace of the testing machine, and the center of the specimen is completely aligned with the center of the fxed position (see Figure 2) of the testing machine. Te specimen is heated to the specifed temperature and kept at a constant temperature for 60 min.
(3) After the surface temperature of the specimen is stabilized, the specimen is loaded, and the loading force and displacement data are collected and recorded in real time. (4) Te loading process is ended when the loading force is reduced to less than 85% of the ultimate bearing capacity.  Note. D is the diameter of the bending indenter and a is the thickness of the fre-resistant steel. Advances in Materials Science and Engineering 3 (5) When the temperature in the furnace falls below 200°C, we open the furnace to cool down. We replace the specimen when the surface temperature is less than 50°C. Figure 3 shows the failure mode of ordinary high-strength bolt specimens. Under room   Advances in Materials Science and Engineering temperature, the bolt in the specimen almost has no shear deformation (see Figure 3(a)), but the bolt hole of the freresistant steel plate has signifcant deformation, which shows that the fnal failure mode is the compressive failure of the bolt hole wall. Under the condition of 200°C∼600°C, both the bolt hole and screw have large plastic deformation under the external loading (see Figure 3(b)). Te steel plate and bolt share the shear force, and the shear deformation degree of the screw is smaller than that of the bolt hole. At 700°C, there is almost no deformation of the bolt hole, but the shear deformation of the screw is signifcant (see Figure 3(c)). Tus, the fnal failure mode of the specimen is a double shear failure of the bolt screw. For the joint connected by ordinary high-strength bolts and a fre-resistant steel plate, the higher the temperature is, the more likely the failure is to occur on the bolt. Figure 4 shows the failure mode of fre-resistant highstrength bolt specimens. Under the condition of 20°C∼700°C, all specimens have the same failure mode, which is the compressive failure of the bolt-hole wall. Te bolt hole experiences signifcant plastic deformation, while the deformation of the screw is not signifcant. Figure 5 shows the relationship between loading force and displacement of specimens under diferent experimental conditions. According to this fgure, there are four stages during the whole loading process (shown in Figure 5(a)). (1) Stage 1 is the initial elastic stage. In this stage, the loading force of the specimen nearly increases linearly with the increase in displacement. Tere is no contact between the bolt screw and bolt hole wall, and thus, the bearing capacity of the specimen mainly depends on the friction between the bolt and the steel plate. (2) Stage 2 has slip characteristics. At this stage, relative slip displacement occurs between the bolt and steel plate. Tis is the reason that the loading force-displacement curve appears to have a signifcant platform segment or descending segment. (3) Stage 3 corresponds to the extrusion-strengthening stage. At this stage, the bolt is in close contact with the bolt hole wall. Te hole wall is under pressure, the bolt screw is under shear, and the loading force increases again with the increase in displacement. As can be seen from the fgure, under the same deformation condition, strength and stifness (slope of the curve) of the specimen decrease with the increase of temperature, which indicates that the "strengthening" efect tends to weaken. (4) Stage 4 corresponds to the period from the ultimate strength (F N ) to the failure of the specimen. At this stage, the slope of the loading force-displacement curve begins to decline, and both material strength and specimen deformation reach the limit state.

Degradation of Deformation
Capacity. Te ultimate displacement Δ N is defned as the displacement when the specimen reaches ultimate strength (see Figure 6(a)), which represents the ductility and deformation capacity of the specimen to some extent. Figure 6 shows the relationship between ultimate displacement and temperature. Diferent treatments of friction surfaces (specifcally, impeller blasting and sprayed hard quartz sand) almost have no obvious efect on the Δ N . For both ordinary bolt specimens and fre-resistant bolt specimens, Δ N generally shows a trend of decreasing with the increase of temperature. In case of the temperature from 20°C to 400°C, Δ N of ordinary bolt specimens and that of fre-resistant bolt specimens are similar and do not show obvious regularity. When the temperature is between 400°C and 700°C, the decreasing rate of Δ N of ordinary bolt specimens is faster than that of freresistant bolt specimens. As a result, Δ N of fre-resistant bolt specimens is even 1.5 times as much as high of ordinary bolt specimens at 700°C. Tis result is related to the failure mode of specimens (see Section 3.1). Te ordinary bolt specimens show the double shear failure of a bolt screw at elevated temperatures (such as 600°C and 700°C), and the fre-resistant steel plate cannot contribute its plastic deformation ability. Te specimen deformation mainly comes from the very limited shear deformation of high-strength bolts, and thus, Δ N is signifcantly small. Table 5 shows the experimental results of specimen bearing capacity. Te ultimate strength (F N ) and slip load (F S ) of specimens correspond to the beginning points of Stage 2 and Stage 4 (see Section 3.2 and Figure 5(a)), respectively. To facilitate comparison, F N and F S are standardized. Taking the ultimate strength of specimens at room temperature (F N,20 ) as a reference, the reduction factor of the tensile strength (K N ) is calculated, where K N � F N,T /F N, 20 . F N,T is the ultimate strength of specimens at elevated temperatures, whose treatment of the friction surface and type of high-strength bolts are the same as the corresponding specimen at room temperature. Similarly, the reduction factor of the slip strength (K S ) is calculated, where K S � F S,T /F S, 20 . F S,T and F S,20 are the slip loads of specimens at elevated temperatures and room temperature, respectively. Figure 7 shows the relationship of K N and temperature. For the specimens of both ordinary and fre-resistant bolts, the reduction factor of tensile strength (K N ) decreases with the increase in temperature. In addition, diferent treatments of the friction surface almost have no efect on the K N . For specimens of ordinary bolts, when the temperature is less than 400°C, K N decreases slowly with the increase in temperature. When the temperature reaches 400°C, K N decreases to 0.9 times the original value (20°C). As the temperature exceeds 400°C, K N decreases rapidly. At 600°C, K N decreases to below 0.4, and K N is only 0.2 at 700°C, indicating that the specimen has lost most of its bearing capacity. For specimens of fre-resistant bolts, K N of them is always larger than that of ordinary bolt specimens. Similar to specimens of ordinary bolts, K N of fre-resistant bolts specimens decreases slowly from 20°C to 400°C, and there is little diference between K N of ordinary bolt specimens and K N of fre-resistant bolt specimens. When the temperature is between 400°C and 700°C, K N of fre-resistant bolt specimens decreases relatively rapidly, and the rate of decline is smaller than that of ordinary bolt specimens. For K N of fre-resistant Advances in Materials Science and Engineering 5    Advances in Materials Science and Engineering bolt specimens, it is about 0.7 at 600°C and 0.4 at 700°C, which is far greater than that of ordinary bolt specimens. Te abovementioned results remind us that when the design temperature is greater than 400°C, there is a great diference between the properties of ordinary and fre-resistant bolted joints. Terefore, in this case, the performance matching of connecting members should be considered, such as the connecting bolts of fre-resistant steel plates, which should also be made of fre-resistant steel. Figure 8 shows the relationship between K S and temperature. For both ordinary bolt specimens and fre-resistant bolt specimens, the reduction factor of slip strength (K S ) shows a trend of increasing frst (20°C∼400°C) and then decreasing (400°C∼700°C). When the temperature exceeds 400°C, K S decreases sharply, and the specimen has almost no slip load at 700°C. Diferent treatments of friction surfaces (specifcally, impeller blasting and sprayed hard quartz sand) have a visible infuence on K S , especially on fre-resistant bolt specimens. Under the same treatment of the friction surface, K S of fre-resistant bolt specimens is always larger than that of ordinary bolt specimens.

Degradation of Bearing Capacity.
Considering the abovementioned analysis of the experimental results of bearing capacity degradation with temperature, in cases where the temperature is less than 400°C, fre-resistant and high-strength bolted joints can be designed according to the friction type connection. In cases where the temperature is greater than 400°C, fre-resistant and high-strength bolted joints are suggested not to consider the slip bearing capacity and to be designed according to the shear bearing type connection.

Calculation Model of Reduction
Factor. Te reduction factor of tensile strength plays an important role in evaluating the bearing capacity of components, joints, and structures under elevated temperature and fre conditions. Based on the relationship between K N and temperature in the experimental results, the calculated model of K N is proposed in linear form (equation (1)), which is convenient for application and structural design. Equations (2) and (3) are calculated models of K N for specimens of fre-resistant bolts and for specimens of ordinary bolts, respectively. Te parameters a and b in the calculated model are obtained by ftting. Te degrees of ftting of equation (2) are R � 0.842 (20°C ≤ T < 400°C) and R � 0.976 (400°C ≤ T < 700°C), respectively. Te degrees of ftting of equation (3) are R � 0.912 (20°C ≤ T < 400°C) and R � 0.994 (400°C ≤ T < 700°C), respectively. Figure 9(a) shows experimental and calculated reduction factors K N plotted with respect to temperatures.
Te degree of reliability of the calculated model (equations (2) and (3)) obtained by ftting the experimental data is 50%, and the probability of the predicted K N being larger or smaller than the actual value is almost the same, which is logically unsafe. Te large degree of reliability  indicates the great probability that the actual K N is greater than the calculated K N . In practical engineering, the calculated model of K N with larger reliability is usually required for safety. K N is a variable between 0 and 1, so it is assumed to satisfy the log-normal distribution. Te estimated value of b with the degree of reliability p j is expressed in equations (4) and (5). Table 6 shows the model parameters of K N under diferent degrees of reliability, which can be selected for the suitable reliability according to requirements. Figure 9(b) shows the calculated model of K N for fre-resistant highstrength bolt specimens under diferent reliability levels (including 50%, 85%, 95%, and 99%), and Figure 9(c) shows the calculated model of K N for ordinary high-strength bolt specimens under diferent reliability levels: for specimens of fire − resistant bolts, for specimens of or di nary bolts, where a and b are model parameters of the linear calculated model, b and b are estimated value and average value of b, respectively, μ pj is the standard normal deviator (where subscript pj represents j degrees of reliability), s is the standard deviation of experimental specimens, μ pj s is the model error of the linear calculated model, n is the total quantity of specimens, K Ni is the reduction factor of each specimen, and K N is the calculated reduction factor based on the model parameter b. Figure 10 shows the comparison between experimental and calculated reduction factors of tensile strength. Te solid line is the 1 : 1 line, two dash lines are bounds that correspond to the scatter band of 1.05, and two dot lines are bounds that correspond to the scatter band of 1.15. Most calculated reduction factors are located within a safe scatter band of 1.05. All calculated reduction factors are located within a scatter band of 1.15, which shows good accuracy of the proposed calculated model. Te proposed linear model has relatively great convenience for application and reliability design, and considering the limited experimental data, more work can be carried out to have a solid guideline. Figure 11 shows an illustrative design example. In a steel-braced frame, a beam is usually connected to the cantilever arm of a column by connecting plates and bolts, Advances in Materials Science and Engineering which is a typical beam-column connection. As shown in this fgure, the force conditions of connecting plates and bolts in the fange connection are the same as those in the bolted joint of this paper experiment. Using the proposed prediction model of K N , the design value of high temperature bearing capacity at the fange connection can be obtained for a given temperature, bolt type, and reliability. In addition, this calculated process and example can be applied to structures of other steel types (such as stainless steel) for fre design.
In addition, material reduction factors and experimental results are compared, which is shown in Figure 12   Establish the prediction model of strength reduction factor Determine the temperature, reliability, and bolt type, and obtain the corresponding K ′ N Multiply K ′ N with the design value of member strength at room temperature, and obtain that at elevated temperature

Conclusion
Experiments on 40 specimens of high-strength and freresistant bolted joints are conducted. Te experimental results are used to investigate the efect of temperature, treatment of the friction surface, and steel type of a highstrength bolt on the mechanical performance of specimens. A calculation model is proposed to predict the reduction factor of tensile strength. Te main fndings and conclusions are summarized as follows: (1) Te bearing and deformation capacities of fre-resistant bolt specimens are generally better than those of ordinary bolt specimens at elevated temperatures, especially when the temperature exceeds 400°C. As the temperature increases, the failure mode of ordinary bolt specimens changes from a compressive failure of the bolt hole wall (20°C∼400°C) to a double shear failure of the bolt screw (>600°C). Te freresistant bolt specimens have the same failure mode (which is the compressive failure of the bolt hole wall) under the condition of 20°C∼700°C. (2) Te efect of diferent treatments of the friction surface (specifcally, impeller blasting and sprayed hard quartz sand) on the ultimate displacement and ultimate strength of specimens can be negligible. However, the treatments of the friction surface have a relatively signifcant efect on the slip load of specimens. (3) Te fre-resistant and high-strength bolted joints can be designed according to the friction type of connection when the temperature is between 20°C∼400°C. In case of the temperature greater than 400°C (400°C∼700°C), fre-resistant and highstrength bolted joints are suggested to be designed according to the shear bearing type connection. (4) For both ordinary and fre-resistant bolt specimens, the reduction factor of tensile strength (K N ) decreases with the increase in temperature. K N decreases slowly when the temperature is less than 400°C, and in the case of 400°C, K N is 0.9∼0.95 times the value at 20°C. As the temperature exceeds 400°C, K N decreases rapidly. K N of ordinary and fre-resistant bolt specimens at 700°C are about 0.2 and 0.4, respectively, indicating that the specimen has lost most of its bearing capacity.
(5) Te calculated model of K N under diferent degrees of reliability is proposed to predict the tensile strength of bolted joints (including fre-resistant and ordinary high-strength bolted specimens) under elevated temperatures, which has the potential to be used for the fre-resistant design of connections and structures. At 20°C∼400°C, K N of fre-resistant and ordinary high-strength bolted specimens is similar, and K N of fre-resistant bolt specimens is larger than that of ordinary bolt specimens under the condition of 400°C∼700°C.
According to the results in this study, it is necessary to use fre-resistant bolts to connect fre-resistant steel plates. Te prediction models of the strength reduction factor for fre-resistant steel plates matched with diferent types of bolts are proposed for reference in structural fre design. Te abovementioned study is helpful in understanding and promoting the application of fre-resistant steel, especially in a country like China, where fre-resistant steel is not widely used. Given the better practical signifcance, we prepare to conduct an experimental study on the joint of ordinary steel plates and fre-resistant high-strength bolts to consider the infuence of the type of steel plate on the mechanical properties of bolted joints at elevated temperatures.

Data Availability
Te experimental data used to support the fndings of this study are currently under embargo while the research fndings are commercialized. Te data, 6 months after publication of this article, can be obtained from the corresponding author upon request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.