High Temperature Mechanical Properties of High Strength Structural Steels Q550, Q690 and Q890

Abstract

This paper reports the results of an experimental study on high temperature mechanical properties of high strength structural steel (HSSS) produced in accordance with Chinese materials standards. Steady-state tensile coupon tests were carried out on specimens made of China grade steels of Q550, Q690 and Q890. Nine elevated temperature levels up to 800 °C were considered. The elastic modulus, yield strength, ultimate strength and ultimate strain were derived from the measured stress-strain curves. A model was developed to predict the high temperature properties of these steels using an approach developed by the National Institute of Standards and Technology (NIST) and by calibrating the model parameters to the test data. The test results are compared to other tests on high strength steels reported in literature. The test results are also compared to predictions of high temperature properties from various building codes and other standards. The study found that steel grade has significant effect on the reduction factors. The difference between the reduction factors of elastic modulus for Q690 and Q550 was 30% at 600 °C. In this study, reduction factor is defined as the ratio of the high temperature property to the corresponding room temperature property. The study also found that the material models in current codes are not applicable to the investigated high strength steels.

1. Introduction

High strength structural steel (HSSS) usually refers to the steels with a nominal yield stress not less than 460 MPa [1]. HSSS can sometimes offer economic benefits compared to more conventional mild structural steel, and has been applied to several landmark buildings and major projects [2]. Although Q460 (the Chinese Grade of steel with nominal yield strength of 460 MPa [3]) to Q960 steels have been produced by the steel industry in China, the application of HSSS in China has to date been limited. Nonetheless, as the use of HSSS becomes more common in building and bridge construction in China and other locations, data is needed on the high temperature mechanical properties of these steels for structural-fire safety analysis.

The high temperature mechanical properties for steel are usually derived from tensile coupon tests which can be conducted in one of two ways, namely transient- and steady- state tests. In a transient-state test, the test specimen is first loaded to a target stress level and then exposed to an increasing temperature. Strain is measured as a function of temperature during the test. In a steady-state test, the test specimen is first heated to a specific temperature level and then loaded to failure. Strain is measured as a function of applied engineering stress, as is the case for a conventional room temperature tension test. There is no general agreement in the literature as to which testing technique is most appropriate for structural fire engineering applications, and results of tests using both methods are widely reported in the literature. One of the key issues discussed in the literature [4–6] in regard to the test methods is the role of creep in the final stress-strain curves. For transient state tests, the degree to which creep is represented in the final stress strain curves depends on the heating rate, with lower heating rates giving a greater amount of creep. For steady state tests, the degree to which creep is represented in the final stress strain curves depends on the loading rate, with lower load rates giving a greater amount of creep. For either test method, the effects of creep are included in the final stress strain curves in an approximate manner, and cannot be compared directly to real fire conditions. In real fire conditions, both temperature and stress levels change throughout a fire event, and can vary greatly depending on the fire scenario and structural framing. Further discussion comparing the two test methods is provided in [7]. Although both methods can be considered an approximate representation of the heating and stress conditions in an actual structure in an actual fire, both methods provide useful information on high temperature properties of steel for structural fire engineering design. For the test program reported herein, all tests were conducted using the steady state methodology. There were two reasons for this. The first reason was that the available test equipment permitted more accurate control of loading rates than of heating rates. The second reason was that an objective of this test program was to capture the entire stress strain curve, including the descending portion prior to fracture. It is not possible to capture the descending portion of the stress strain curve in a transient state test, since the stress is held constant for any given test.

There are several previous studies on high temperature mechanical properties of HSSS. Chen et al. [8] reported steady- and transient- state tests on Australia steel BISPLATE80 (nominal yield strength of 690 MPa). The tests were carried out in accordance with the ASTM E21-97 [9] and AS 2991 [10]. Heating rate in the transient test was not reported. Temperatures ranged from 100 to 1000 °C. The study found that the material models given in the AISC [11], ASCE [12], AS 4100 [13], and EC3 [14] generally overestimated the yield strengths at elevated temperatures. The elastic modulus predicted by AISC [11], ASCE [12], AS 4100 [13], and EC3 [14] were generally higher than measured for transient-state tests and less than measured for steady-state tests. Qutinen [5] reported transient- and steady-state tests on European steel S460M (460 MPa). The test specimens were made in accordance with DIN EN10002-5 [15], the heating rate was 10 °C/min, and the temperatures ranged from 100 to 900 °C. The study found that the high temperature material models given in the EC3 [14] were inapplicable to the investigated steels. Schneider and Lange [16] reported transient-state tests on European steel S460 (460 MPa). The test specimens were made in accordance with EN10002-5 [15], the heating rates ranged from 3 to 30 °C/min, and the temperatures ranged from 20 to 800 °C. The study found that the high temperature material models given in the EC3 [14] significantly overestimated the yield strength of the considered steels. Qiang et al. [17–19] reported steady- and transient- state tests on European steel S460N, S690 and S960 and found that AISC [11], ASCE [12], AS 4100 [13], and EC3 [14] also overestimated the yield strength at elevated temperatures. Heidarpour et al. [20] reported steady-state tests on Australia steel with nominal proof stress of 1200 to 1300 MPa and found that AISC [11] and AS 4100 [13] were inapplicable to the investigated steels. Wang et al. [21] reported steady-state tests on China made steel Q460. The tests were carried out in accordance with GB/T4338-2006 [22] and the temperatures ranged from 20 to 800 °C. The study found that the AISC [11], AS 4100 [13], EC3 [14] and CECS200 [23] were inapplicable to the investigated steel. Researches on high temperature mechanical properties of other HSSS may be found in Ref. [24].

Supported by a grant from the National Natural Science Foundation of China, the authors from Tongji University pursed a test program aimed at providing data and material models for China produced HSSS for structural fire safety design. The whole project includes development of stress-strain models and explicit creep models. This paper reports the experimental study and development of stress-strain models for China produced steels of grades Q550, Q690 and Q890.

2. Experimental study

2.1. Test machine and instruments

As shown in Fig. 1, the high temperature tensile coupon tests were conducted using an elevated-temperature electronic material testing machine using an AC servo and servo driver. The load capacity of the machine is 300 kN, and the crosshead displacement rate can be controlled from 0.001 to 250 mm/min. The heating device was a GW900 high temperature furnace that can heat the specimen up to 1100 °C. The minimum resolution of temperature measurement was 0.1 °C. The furnace consists of upper, middle and lower heating zones, and each zone was equipped with a type K wire thermocouple. Heating elements were turned on or off independently based on the reading of the thermocouples to maintain uniform heating for the coupon. A type K thermocouple was tied to the middle of the specimen with asbestos cloth to measure the temperature of the specimen. For the thermocouple measurement of the temperature of the specimen, the error was ± 2 °C. A high temperature contact-type extensometer with a precision of 0.001 mm was used to obtain the strain of the specimens.

Fig. 1.

Test machine and instruments for steady-state test. (a) Test machine; (b) Furnace; (c) Extensometer; (d) Specimen with thermocouple.

2.2. Test procedure

The tests were carried out in accordance with GB/T4338-2006 [22]. Fig. 2 shows the shape and dimensions of the test specimens. The test specimens were cut from quenched and tempered Q550, Q690, Q890 structural steel plates with a nominal thickness of 20 mm. The chemical composition of steel plates is shown in Table 1.

Fig. 2.

Dimensions of specimen for steady-state tests (unit in mm).

Table 1.

Chemical composition of 20 mm high strength steel plates (0.001%).

	C	Si	Mn	P	S	Cr	Al	Mo	Ti	Cu	Nb	Ni	V
Q550	80	200	1700	11	3	20	32	10	18	10	33	20	8
Q690	170	31	100	10	2		29					16	12
Q890	160	330	1250	11	2	270	41	560	17	20	21	110	48

In the steady-state tests, the specimen was heated to a target temperature level with a heating rate of 10 °C/min. The target temperature level was maintained for 15 mins, and the specimen was then loaded until failure. A total of nine elevated temperature levels were considered, which were 200 °C, 300 °C, 400 °C, 450 °C, 500 °C, 550 °C, 600 °C, 700 °C and 800 °C. For each temperature level, 2 to 3 specimens were tested (Table 2). Testing was performed under strain-control and the strain rate in the steady-state tests was 0.003/min. Room temperature tests were also conducted for reference.

Table 2.

Temperature levels and number of test specimens.

	20 (°C)	200 (°C)	300 (°C)	400 (°C)	450 (°C)	500 (°C)	550 (°C)	600 (°C)	700 (°C)	800 (°C)
Q550	3	2	2	2	2	2	2	2	2	2
Q690	3	3	3	3	2	3	2	3	3	3
Q890	3	2	2	2	2	2	2	2	2	2

3. Experimental results

3.1. Visual observations

Figure 3 shows the coupons after completion of testing. The coupons exhibit different colors depending on the test temperature. The relation between the color and surface appearance of the steel with maximum temperature exposure can be useful of post-fire investigations of structures using high strength steels. Also, as seen in these photos, the length of the necked region of the coupon increases at higher temperatures. Fig. 4 shows close-up views of the fracture surfaces of the coupons after testing. The appearance of the fracture surfaces also changes with temperature. This suggests that the micromechanical features affecting fracture also change with temperature.

Fig. 3.

Failure models of the specimens at various temperatures

Fig. 4.

Fracture shapes of the specimens at various temperatures. (a) 20 °C; (b) 200 °C; (c) 300 °C; (d) 400 °C; (e) 450 °C; (f) 500 °C; (g) 550 °C; (h) 600 °C; (i) 700 °C; (g) 800 °C

3.2. Mechanical properties

Figure 5 shows the measured strain-stress curves for different steels at various temperatures. At high temperatures, the stress-strain curves exhibit no yield plateau although Q550 steel exhibited no yield plateau at room temperature. Furthermore, as temperature increases, the linear portion of the stress-strain curve shortens, the strain corresponding to ultimate strength decreases, and the descending stage of the curve shows a more gentle slope. The strain at fracture tends to decrease with increasing temperature, up to approximately 400 to 500 °C. For higher temperatures, the strain at fracture increases significantly compared to room temperature.

Fig. 5.

Stress-strain curves for Q550, Q690 and Q890 for temperatures ranging from 20 °C to 800 °C.

There is no general agreement on the definition of yield strength of steel at high temperature and various definitions have been reported in the literature. As shown in Fig. 6, the 0.2% offset yield strength (f_0.2,T) is the intersection point of the stress-strain curve and the proportional line offset by 0.2% strain. The yield strengths of f_0.5,T, f_1.0,T, f_1.5,T and f_2.0,T are those corresponding to stress at total strain levels of 0.5%, 1.0%, 1.5%, and 2.0%. ECCS [25] specifies that above 400 °C, the yield strength is taken as f_0.5,T; and below 400 °C, the yield strength is corresponding to an interpolation between f_0.2,T and f_0.5,T. BS5950 [26] specifies that the yield strength at high temperature can be taken as f_0.5,T, f_1.5,T or f_2.0,T, depending on the type of structural members. EC3 [14] specifies that the yield strength at high temperature is f_2.0,T. CECS200 [23] and Ref. [4] recommend the use of f_1.0,T. Table 3 to Table 5 give the elastic modulus, yield strengths, ultimate strength and ultimate strain for Q550, Q690 and Q890, respectively. The ultimate tensile strength (f_u,T) and ultimate strain (ε_u,T) were also given in the tables.

Fig. 6.

Definition of nominal yield strengths. f_0.2–0.2% offset strength; f_0.5–0.5% proof strength; f_1.0–1.0% proof strength; f_1.5–1.5% proof strength; f_2.0–2.0% proof strength; and f_u–ultimate strength. The initial tangent line is also shown, the slope of which is the elastic modulus E=tan α.

Table 3.

Elastic modulus, yield strengths, ultimate strength and ultimate strain for Q550.

T (°C)	ET (MPa)	f0.2,T (MPa)	f0.5,T (MPa)	f1.0,T (MPa)	f1.5,T (MPa)	f2.0,T (MPa)	fu,T (MPa)	εu,T (%)
20	214459	651	651	692	706	715	747	17.36
200	211190	631	631	720	747	760	778	14.74
300	206166	587	592	675	704	719	763	17.65
400	195955	537	545	623	642	651	656	16.45
450	178243	503	509	573	586	586	588	16.76
500	177007	430	441	471	472	469	473	20.37
550	147701	367	373	382	377	369	383	33.37
600	127236	282	283	271	258	248	285	44.25
700	87430	93	95	91	88	86	96	87.06
800	32782	27	28	29	29	28	29	172.67

Table 5.

Elastic modulus, yield strengths, ultimate strength and ultimate strain for Q890.

T (°C)	ET (MPa)	f0.2,T (MPa)	f0.5,T (MPa)	f1.0,T (MPa)	f1.5,T (MPa)	f2.0,T (MPa)	fu,T (MPa)	εu,T (%)
20	204976	1003	1002	1010	1021	1027	1057	16.64
200	204400	893	855	938	965	981	1025	13.87
300	201032	847	803	904	933	948	990	16.97
400	183367	798	753	849	875	889	906	16.93
450	175592	761	712	810	831	840	845	18.68
500	174246	698	668	738	750	752	754	22.17
550	150692	593	563	633	638	638	640	26.91
600	132509	441	435	472	473	467	475	27.14
700	60189	104	110	117	118	119	120	70.80
800	23209	37	39	43	44	44	45	162.02

4. Constitutive models

4.1. NIST steel stress-strain model

The National Institute of Standards and Technology (NIST), as part of its report on the collapse of the World Trade Center [27], characterized many grades of steel recovered from the buildings to provide stress-strain models to analyze the impact, fires and resulting collapse. Based on the material test data collected in the investigation and additional data found in literature, a constitutive model, namely the NIST steel stress-strain model or NIST model in this paper, was developed to predict the elevated temperature behavior of structural steels [28]. The general NIST steel stress-strain model accounts for temperature dependence of elastic modulus, yield strength, strain-hardening behavior, and strain rate sensitivity. The mathematical form of the NIST steel stress-strain model is assumed to be applicable to other specialty steels, such as high strength bolt steels, quenched- and-tempered steels, and fire-resistant steels. A NIST technical report [29] contains supporting data, details of regression analyses and approaches, and estimates of the uncertainties in the fitted parameters. In this study, the NIST steel stress-strain model, without considering the strain rate effects, was used in curve fitting of the test data, which has the expression of [30]

$$\sigma =\epsilon E_T(\epsilon \le \frac{f_{y,T}}{E_T})$$ (1a)

$$\sigma =f_{y,T}+(k_3-k_4{f}_{y,20})\exp [-{(\frac{T}{k_2})}^{k_1}]{(\epsilon -\frac{f_{y,T}}{E_T})}^n(\epsilon >\frac{f_{y,T}}{E_T})$$ (1b)

with

$$\frac{E_T}{E_{20}}=exp[-\frac{1}{2}{(\frac{T-20}{e_3})}^{e_1}-\frac{1}{2}{(\frac{T-20}{e_4})}^{e_2}]$$ (2)

and

$$\frac{f_{y,T}}{f_{y,20}}=r_5+(1-r_5)exp[-\frac{1}{2}{(\frac{T-20}{r_3})}^{r_1}-\frac{1}{2}{(\frac{T-20}{r_4})}^{r_2}]$$ (3)

where e₁ to e₄ are parameters, determined by regression analysis of the test data on high temperature elastic modulus; r₁ to r₅ are parameters, determined by regression analysis of the test data on high temperature yield strength; and k₁, k₂, k₃, k₄ and n are parameters considering the strain hardening behavior, determined by regression analysis of the measured stressstrain curves. The yield strength in the NIST model, f_y,T is based on the 0.2% offset definition (f_0.2).

4.2. Curve fitting

Figure 7 shows the reduction factors of elastic modulus and yield strength for Q690 steel. The reduction factor is defined as the ratio of the high temperature property to the corresponding room temperature property. The plots are based on the data in Table 4 for Q690 steel. The dash blue line and the solid red line are results from curve fitting using the NIST model and using a polynominal function, respectively. It shows that both the NIST model and the polynominal function fit well with the test data. At very high temperatures (above 700 °C), the NIST model gives a slightly better prediction than the polynominal function. Table 6 gives the values of the curve fitting parameters for the NIST model (e₁ to e₄ for elastic modulus and r₁ to r₅ for f_0.2) for different steels. The values for parameters r₁ to r₅ for 1.0% proof strength (f_1.0) are also provided in the table for reference purpose. It should be noted that using the values of curve fitting parameters given in the NIST technical report [29] did not provide a good prediction of the reduction factors. Using the parameters for reduction factors given in Table 6 and properly choosing parameters for strain hardening behavior (k₁ to k₄ and n), the NIST model could adequately predict the stress-strain behavior of each steel grade. Table 7 to Table 9 give the values for curve fitting parameters (based on f_0.2) for strain hardening behavior for different steels. The values for curve fitting parameters based on f_1.0 are also provided for reference purpose. Fig. 8 shows a comparison between the measured stress-strain behavior for Q690 steel and that predicted by the NIST model after calibrating the model parameters. The calibrated NIST model generally matches the experimental curves well, except for the nonlinearity in the experimental curves at low strains. This nonlinearity, missed by the NIST model, may be significant when modeling buckling phenomena at high temperatures.

Fig. 7.

Elevated temperature reduction factors for Q690 steel for modulus of elasticity and for yield strength (f_0.2, f_0.5, f_1.0, f_1.5 and f_2.0).

Table 4.

Elastic modulus, yield strengths, ultimate strength and ultimate strain for Q690.

T (°C)	ET (MPa)	f0.2,T (MPa)	f0.5,T (MPa)	f1.0,T (MPa)	f1.5,T (MPa)	f2.0,T (MPa)	fu,T (MPa)	εu,T (%)
20	196779	771	773	770	771	772	795	19.84
200	201351	706	700	729	740	748	773	17.89
300	196923	690	681	723	736	745	759	17.61
400	189423	641	634	676	687	692	694	18.43
450	178032	607	599	637	643	642	644	17.46
500	174203	543	542	563	560	552	564	19.41
550	171271	446	451	461	452	442	463	25.02
600	152784	319	329	335	326	318	337	27.39
700	79979	93	98	98	97	95	100	70.37
800	22606	19	21	22	23	24	24	132.66

Table 6.

Values of the parameters in NIST model for curve fitting (elastic modulus and 0.2% offset strength). Values in bracket are for curve fitting of f_1.0.

	Q550	Q690	Q890	All
e1	1.421	8.308	2.076	7.026
e2	3.803	2.684	6.442	2.723
e3	782 °C	656 °C	1044 °C	657 °C
e4	580 °C	1035 °C	620 °C	810 °C
r1	6.239 (4.963)	6.245 (5.953)	8.031 (7.734)	6.700 (5.98)
r2	1.815 (6.866)	0.461 (0.3328)	0.758 (0.9486)	0.945 (1.168)
r3	576 °C (512 °C)	547 °C (541 °C)	575 °C (574 °C)	563 °C (546 °C)
r4	729 °C (1910 °C)	7921°C (184,000 °C)	1207 °C (1359 °C)	1256 °C (1932 °C)
r5	0	0	0	0

Table 7.

Values of the parameters in NIST model for curve fitting (strain-hardening) for Q550 steel (based on f_0.2). Values in bracket are for curve fitting based on f_1.0.

T (°C)	k1	k2 (°C)	k3 (MPa)	k4	n
20	1.012 (−2.872)	480 (27.58)	713.1 (600)	0.3651 (−4.4)	0.4668 (−0.1574)
200	0.796 (0.5916)	450 (6444)	650 (595.4)	−0.6595 (0.6636)	0.4277 (−0.5787)
300	2.257 (1.658)	450 (3276)	650 (−29.71)	−0.7929 (−0.2153)	0.4769 (−1.015)
400	1.065 (−0.5323)	449 (211.7)	590 (671.7)	−0.7087 (0.1876)	0.3548 (−0.3674)
450	0.7498 (−15.33)	536 (187.8)	508.7 (534.6)	−0.3373 (0.7581)	0.3326 (−3.016)
500	0.899 (−1.034)	410 (205.3)	651.3 (496.2)	−0.7456 (0.639)	0.4395 (−0.3612)
550	0.8919 (−0.373)	330 (43.42)	786.5 (25.47)	−0.1361 (0.8891)	0.5254 (−0.5642)
600	0.7524 (−1.224)	394 (176.7)	684 (119.6)	−0.1643 (1.407)	0.8311 (−0.6852)
700	80.49 (1.051)	1004 (336.5)	748.5 (183.9)	−76320 (9.896)	2.957 (−0.292)
800	20.75 (−1.621)	1500 (225.9)	802.1 (440.1)	−1391 (0.6394)	2.467 (−0.5559)

Table 9.

Values of the parameters in NIST model for curve fitting (strain-hardening) for Q890 steel (based on f_0.2). Values in bracket are for curve fitting based on f_1.0.

T (°C)	k1	k2 (°C)	k3 (MPa)	k4	n
20	0.4947 (−0.2019)	364 (25.78)	1034 (92.87)	0.3192 (−0.4995)	0.7659 (−0.2947)
200	1.034 (−0.3308)	449 (0.002181)	658.8 (−4.811)	−0.2729 (−0.1283)	0.4869 (−0.9432)
300	1.021 (−1.328)	457 (96.16)	642.8 (24.71)	−0.4093 (−0.1924)	0.433 (−0.4947)
400	0.7535 (−2.096)	450 (124.8)	650 (35.97)	−0.5159 (−0.08894)	0.4203 (−0.6469)
450	1.439 (−2.927)	559 (224.7)	825.4 (1.669)	−1.096 (−0.09737)	0.5633 (−0.5193)
500	1.206 (−20.61)	465 (193.9)	678.8 (3.463)	−0.3257 (0.007797)	0.4484 (−4.096)
550	0.8666 (−18.59)	439 (415.3)	578 (5.657)	−0.4995 (−0.02547)	0.4272 (−1.236)
600	0.6693 (−2.574)	456 (187)	800.3 (−0.8423)	−3.599 (1.622)	0.7693 (−1.151)
700	0.4561 (−0.9775)	396 (201.2)	840 (230)	−0.4962 (0.2263)	0.6926 (0.41)
800	0.7754 (0.1866)	442 (732.3)	793.7 (−2.268)	−0.2465 (−0.06013)	0.7339 (−0.2919)

Fig. 8.

Comparison of measured stress-strain curves with predictions from calibrated NIST model for Q690 steel.

Figure 9 shows the reduction factors for all three grades of steel tested in this program. Curve fitting of the dataset including all steels was also performed and the values of the curve-fitting parameters are reported in Table 6. As shown in Fig. 9a, the reduction factors of elastic modulus for the three steels were very similar at both low and very high temperatures. However, from 550 to 700 °C, somewhat larger differences among the three steels can be observed. For example, at 600 °C, the difference between the reduction factors of elastic modulus for Q690 and Q550 was 0.18 (30%). As shown in Figs. 9b–f, the reduction factors of yield strength for different steels were similar. At low temperatures, the reduction factors of yield strength for different steels showed greater variation. For example, at 200 °C, the difference of f_2.0,T/f_2.0,20 for Q550 and Q890 was 0.11 (11%). As a whole, the calibrated NIST model fit well with the dataset of all steels for predicting reduction factors for elastic modulus and yield strength with temperature.

Fig. 9.

Elevated temperature reduction factors for Q550, Q690 and Q890 steels for modulus of elasticity and for yield strength (f_0.2, f_0.5, f_1.0, f_1.5 and f_2.0).

4.3. Comparative study

Figure 10 compares the results of reduction factors for elastic modulus from this study with the results from other researchers and from codes. Test data considered include that reported by Chen et al. [8] (marked as “BISPLATE80”in Fig. 10), Qiang et al. [17,18](“S460N” and “S690”), Choi et al. [31] (“HSA800”), and Chiew et al. [32](“SQ690L”). The values of reduction factors predicted by Maraveas et al. [24] (Eq.1 in the paper) are also considered (“Ref.[24]”). The nominal yield strength for S690, SQ690L, BISPLATE80 and Q690 steels is 690 MPa, and the nominal yield strength for HSA800 is 650 MPa. Fig. 10a shows significant differences in reduction factors for elastic modulus among these different steels that have similar nominal yield strength. Fig. 10b shows that the various models for predicting elastic modulus reduction factors differ significantly from one another and from the test data in this program. This observation is consistent with previous studies on other high strength steels. The current models tend to underestimate the elastic modulus, which is also consistent with previous studies. Fig. 11 compares the results for reduction factors of yield strength (f_2.0) from this study with the results from other researchers and from codes. Fig. 11a again shows considerable variation in reduction factor for yield strength among high strength steels with similar nominal yield strength. Fig. 11b shows there is also variations among various models, with the calibrated NIST model showing the best prediction of the experimental data. Fig. 12 compares the results for reduction factors of yield strength (f_1.0) from this study with the results from other researchers and from codes.

Fig. 10.

Comparison of reduction factors of elastic modulus predicted by others (a) and by codes (b) with the test data.

Fig. 11.

Comparison of reduction factors of yield strength (f_2.0) predicted by others (a) and by codes (b) with the test data.

Fig. 12.

Comparison of predicted and measured reduction factors of 1.0% proof strength.

While the results of this test program on HSSS showed a number of differences between the reducation factors derived from experiments with those predicted by other researchers and by building codes, it should not necessarily be concluded that elevated temperature reduction factors for HSSS are intrinsically different from those for normal strength steels. Even for steels of normal strength, there is a large variation in elevated temperature reduction factors [28]. The mechanical properties of steel at elevated temperatures, and the associated retention factors can be affect by many parameters, including not only the chemistry and microstructure of the steel, but also testing approach (transient-state vs steady-state), heating rates used in transient-state tests, loading-rates used in steady state, interpretation of mechanical properties (e.g. differing definitions of yield stress at elevated temperature), and other factors. Thus, while the assertion that retention factors for HSSS are significantly different than those for normal strength steels may or may not be accurate, the database on high temperature properties of HSSS is currently too small to draw definitive conclusions. Further work is needed on high temperatureproperties of HSSS to broaden the experimental database and to characterize variability in these properties.

5. Conclusions

This paper reported the results of an experimental program to investigate the high temperature mechanical properties of China grade high strength steels of Q550, Q690 and Q890. By conducting steady-state tensile coupon tests, the experimental stress-strain curves at elevated temperatures up to 800 °C were measured. By using the NIST stress-strain model and by calibrating the model parameters to the test data, good agreement was achieved between the test data and the model predictions. It should be noted that the number of tests in this study is limited, and the calibrated model only works for the investigated steels, and not necessarily for other steels. By comparative study with other test results reported in the literature, it was found that there were considerable variations among the high temperature mechanical properties (e.g. reduction factors) for high strength structural steels with similar nominal yield strength. Further, there were significant variations in the predicted reduction factors given in current Chinese, European, American, Australian codes. The predicted reduction factors differed considerably among the codes as well as between the codes and the experimental data. Further data on high temperature properties of high strength steels would be desirable to better characterize variability in these properties as well as to develop models that are most appropriate for these steels.

Table 8.

Values of the parameters in NIST model for curve fitting (strain-hardening) for Q690 steel (based on f_0.2). Values in bracket are for curve fitting based on f_1.0.

T (°C)	k1	k2 (°C)	k3 (MPa)	k4	n
20	5.815 (−0.7606)	2220 (100.7)	−6.878 (106.5)	−5.528 (−12.48)	1.617 (−0.1863)
200	0.9013 (−4.014)	392 (117.3)	644.3 (500.3)	−0.2537 (0.4788)	0.6055 (−0.6861)
300	1.016 (−4.54)	351 (31.62)	654.4 (500)	−0.03647 (0.6801)	0.4076 (−1.251)
400	2.502 (−35.17)	398 (260)	809.7 (164)	−1.027 (0.1913)	0.5473 (−3.267)
450	1.082 (−42.37)	466 (307.2)	724.6 (375.8)	0.2142 (0.4772)	0.4043 (−3.288)
500	0.7514 (−48.31)	480 (300.8)	568.8 (428.9)	−0.1765 (0.5543)	0.4794 (−3.189)
550	1.138 (0.02807)	465 (549.3)	551.3 (221.4)	0.04458 (7.591)	0.4656 (−0.4221)
600	0.3261 (0.6993)	442 (403.5)	641.3 (216.5)	−1.896 (8.637)	0.6607 (−0.3651)
700	0.6335 (1.462)	445 (412.6)	638.5 (260.1)	−0.1289 (8.62)	0.7312 (−0.3212)
800	0.2451 (−1.275)	450 (236.8)	650 (527.6)	−0.2648 (0.6731)	0.8594 (−0.5443)