Evaluation on steel cyclic response to strength reducing heat treatment for seismic design application

Construction and Building Materials 140 (2017) 468–484

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Evaluation on steel cyclic response to strength reducing heat treatment for seismic design application Yujie Yu b, Xiaoxiang Wang b,c, Zhihua Chen a,b,⇑ a

State Key Laboratory of Hydraulic Engineering Simulation and Safety, Tianjin University, Tianjin, China Department of Civil Engineering, Tianjin University, Tianjin, China c Department of Civil Engineering and Engineering Mechanics, Columbia University, New York, NY, USA b

h i g h l i g h t s
Effect of an ultra-high temperature heating slow cooling heat treating process were studied on four steel grades.
The strength reduction extent from heat treatment ranked decreasingly as Q345, Q390, AG235 and Q235.
The heated Q235 and AG235 presented considerate strain hardening and cyclic hardening effect.
The strength reduction effect of heated Q345 and Q390 steel can be maintained under cyclic loading.
Chaboche model parameters were calibrated for all tested specimens.

a r t i c l e

i n f o

Article history: Received 15 January 2017 Received in revised form 15 February 2017 Accepted 16 February 2017

Keywords: Strength reducing heat treatment Cyclic hardening Strain hardening Seismic design application Chaboche model

a b s t r a c t Using fundamental material cycling loading tests, this paper examined the application potential of a new strength reducing heat treatment on seismic design. This heat treatment method aims to reduce mechanical strength from commonly used structural steel grades, and create locally weakened regions that can participate into energy dissipation. Focusing on stiffness, strain/cyclic hardening, and energy dissipation abilities, heat treatment influence on monotonic behavior, cyclic characteristics, and residual loading performances were all studied. Results showed that this slow-cooling heat treatment can effectively reduce yield and ultimate strength, with the strength reduction extents ranked decreasingly as Q345, Q390, AG235, and Q235. Heated Q235 and AG235 specimens showed considerate strain hardening and cyclic hardening effects. On the other hand, heated Q345 and Q390 steel specimens exhibited apparent reduction on both yield and ultimate strengths, and this reduction effect was retained during different cyclic loading protocols. Hardening characteristics, material elastic stiffness, and energy dissipation ratio were all related to loading histories. Finally, to explore the application of this heat-treatment method on seismic design, Chaboche’s cyclic constitutive model parameters were calibrated for future analyses on the structural level. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Structures are subjected to huge amounts of energy during earthquakes. In modern seismic designs, energy dissipation regions, members, or dampers are often incorporated into structures for energy dissipation. The main goal of this design is to protect main structural components from severe damage; therefore special designs are needed to induce inelastic deformation or energy dissipation onto desired unimportant locations or compo-

⇑ Corresponding author at: State Key Laboratory of Hydraulic Engineering Simulation and Safety, Tianjin University, Tianjin, China. E-mail address: [email protected] (Z. Chen). http://dx.doi.org/10.1016/j.conbuildmat.2017.02.088 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

nents [1]. This goal is generally attained through creating the locally weakened regions that can enter plastic state first before large deformation happens at key structural components. And this artificial weakening process is often ensured through sectionreduction measures like dog bone connections or perforated designs [2–4]. With development of steel-making techniques in recent decades, low yield strength steels have been widely used in seismic design for energy dissipation because of their significant mechanical properties, including, low yield stress, high ductility, and excellent energy dissipation capacity [5]. These steels have been widely applied in earthquake resistant, passive control devices, such as, shear dampers, buckling restrained braces (BRBs), and steel plate shear walls (SPSWs) [6,7].

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

In recent years, a novel heat treating process has been proposed to modify the mechanical strengths from existing steel, which can also achieve the locally weakened effect for energy dissipation [8]. The strength reducing heat treatment mainly includes two steps: Heating the steel to extremely high temperature (600 °C– 1200 °C), and then cooling it within a controlled and low rate. With a series of austenite crystal transformations, overall steel strengths could be reduced, creating a regional soft area that can act as ductile fuses without section reduction and corresponding stiffness variations. The strength reduction effect of the heating method has been validated through a series of material tests on low alloy A992 steel under American standards [9], and on generally used Q235B, Q345B, Q390B, and Q235 angle steel types under Chinese standards [10,11]. Previous material studies on heat-treated steels were mainly through uniaxial material tension tests. However, for most polycrystalline metals, kinematic hardening and isotropic hardening are generally observed under cyclic loading. Many experiments have been performed on cyclic response of different structural steel types under different inelastic strain ranges [12–14]. And results displayed quite different cyclic strength developing patterns when compared with monotonic behavior. Especially for low yield point steel [5,15,16], cyclic strain hardening is dramatic and plays a critical role in mechanical response. Given that the heating technique is applied to produce energy dissipation regions or dampers, during an earthquake, treated steel parts need to absorb energy through elasto-plastic deformation, especially under cyclic loading. In order to accurately evaluate the effectiveness and application potential of the heating process for seismic energy dissipation design, thoroughly understanding the developing rule of stress-strain relations for heat treated steels under both monotonic and cyclic loads is critical. An experimental study on Q235B, Q345B, Q390B, and Q235 angle steel types was initiated to comprehensively investigate the influencing pattern of strength reducing heat treatment on mechanical property variations, and cyclic strength developing characteristics of structural steels. Both monotonic and cyclic tests were performed and the obtained results were studied and compared with focus on the hardening responses, cumulative damage degradation properties, loading history differences, steel types, and heating process Based on strength reduction extent and hardening performance, primary evaluation on the application potential of

this heat treatment into seismic design can be worked out. Evaluation result can be used as reference for future applications. 2. Experimental program 2.1. Specimen details The primary objective of the experimental program is to investigate the effect of this strength reducing heat treatment and loading protocols on structure steels’ uniaxial and hysteresis behavior. Studied steel grades were hot rolled Q235, Q345, Q390, and angle steel Q235 under Chinese standards, with nominal yield stress of 235, 345, 390, and 235 MPa, respectively. As steel plates are widely used in engineering structures, specimens adopted plate type rather than cylindrical one. These plates were cut in rolling direction from the hot rolled plates and along the longitudinal direction of finished angles. Detailed information on tested coupons is given in Fig. 1. Considering generally used thickness selections and specimen identifiability during testing, different coupon thickness d were adopted for each steel type: dðQ 235Þ ¼ 16 mm;dðQ 345Þ ¼ 12 mm, dðQ 390Þ ¼ 14 mm, dðAG235Þ ¼ 10 mm. Due to manufacturing, real dimensions might slightly vary from the designed value. Therefore, before tensile tests, the real dimensions were measured as the average value of three measurements within gauge length. 2.2. Heating strategies The strength reducing heating process involves locally heating the existing steel to a high temperature (800 °C) followed by slowly cooling (cooling speed was less than 1.5 °C/min) to enable formulations of coarse grain and low strength pearlite-ferrite microstructure [8]. The heating process is given in Fig. 2(a). A serial of material tension tests have been performed on the influence of different heating parameters with the detailed data and comparisons given in Ref. [11]. Then, based on the results of previous study, the heating process with ultimate temperature 1000 °C, 20 min holding time, and 0.5 °C/min cooling speed to 500 °C, can lead to a satisfying strength reduction effect (the reduction extents for yield strength and ultimate strength were 27.7% versus 12.3% for Q235 steel, 37.5% versus 17.8% for AG235 steel, 40.2% versus 30.2% for Q345 steel, and 35.6% versus 20.8% for Q390 steel). Furthermore, strength reduction is mainly related to the ultimate

Rolling direction

(a) coupon cutting from the plate

469

(b) coupon cutting from the finished angle steel

(c) Dimensions (in mm) Fig. 1. Location and details of the cyclic test specimens.

470

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

T u=ultimate temperature T c=Ceased temperature

th=holding time V c=cooling rate

Controlled slow cooling

Heating

Air cooling

(a) Heat treating process

(b) Electric furnace Fig. 2. Heat treating information.

Table 1 Heat treating strategies in cyclic tests. Number

Peak temp (°C)

Holding time (min)

Stop temp (°C)

Cooling rate (C/min)

Total heat time (h)

R H1 H2

Room 1000 950

– 20 40

– 500 600

– 0.5 0.75

– 18.67 9.86

temperature. To shorten heating duration without greatly influencing strength reduction effect, slight modification on cooling rate Vc and ceased temperature Tc is acceptable. Therefore, in this cyclic test, two types of heating processes were adopted (Table 1). The heating processes were accomplished in an electric furnace (Fig. 2(b)), in which factors were able to settle and the whole heating process was well controlled. 2.3. Cyclic loading tests All cyclic tests were performed with universal tension and compression testing machine INSTRON Model 8803 (Fig. 3(a)). All tests were conducted at room temperature, using strain-control mode, and a triangular wave control signal with the help of extensometer (gauge length of 25 mm). Loading frequency was 0.01 Hz to avoid the effects of temperature increase and strength hardening caused by fast loading, ensuring that steel mechanical behaviors were kept within time-independent. Given the limitation of gauge length and the risk to the instrument during sudden buckling, the loading process would be terminated once obvious buckling took place. Three different loading protocols, namely Cyclic tension random, Cyclic ascend, and Cyclic alternate, were adopted in the tests. Cyclic tension random loading process had varied strain amplitudes at increased mean strains to simulate random earthquake excitations. Cyclic ascend had gradual increased cyclic loading amplitudes to study gradual strength developments. Cyclic alternate had alternate high strain and low strain amplitude combinations to explore strain history memory effect. To consider gradual saturation of isotopic hardening, the first two loading protocols had four cycles for each strain varying range or amplitudes, whereas the last loading protocol had five cycles for each amplitude to ensure enough saturation. Due to different specimen thicknesses, two sets of loading amplitudes were adopted. The amplitudes of large strain range cycles were reduced for Q345 and AG235 steel coupons due to a thinner thickness. Summary of the loading schemes is given in Table 2. Detailed loading processes and amplitudes are indicated in Fig. 3(b–f). 2.4. Monotonic results Before the cyclic tests, a series of monotonic tension tests was conducted to obtain the uniaxial strength conditions of the four

steel types and the strength reduction effect from the two heat treating processes. The monotonic tests were conducted at mechanical lab in Tianjin University. Stress-strain curves of untreated (R condition) and two-heat treated cases are given in Fig. 4. Critical mechanical index as yield strength, ultimate strength (the maximum strength), and ultimate strain (tensile strain corresponding to the maximum strength) are listed in Table 3. Monotonic data showed that the strength reducing heat treatment effectively reduced both yield and ultimate strengths of all four steel grades. With a large extension range extensometer, the whole loading period was captured. Results indicated that after the heat treatment, no reduction was presented on the elongation. Furthermore, the tensile strain corresponding to the ultimate strength also got increased after heat treatment (from eu/eu(R) ratio in Table 3), indicating delayed strength growth and better material ductility. This effect was especially prominent on Q345 steel, for which the ultimate strain increased to more than 1.5 times of the original magnitude, and the overall ductility improved obviously after heat treatment compared with the unheated case. Heating process H2 had only half heat-treating time duration but with similar strength modifying effect as that of H1. Yield strength reduction was more evident compared with corresponding maximum strength. This reduction pattern can facilitate early entrance of heated steel to plastic state to dissipate energy, simultaneously maintaining a good post yielding strength. Among the four steel types, Q345 had the most remarkable strength reduction on both yield (around 45% reduction) and ultimate strength (around 30% reduction), whereas the least effect was presented on Q235 steel (about 25% reduction on yield strength and 11% reduction on ultimate strength). In addition, heat treatment presented similar strength reduction effect on Q390 and AG235 steels.

2.5. Cyclic behavior Cyclic responses of tested specimens are given in Fig. 5, with two heat-treated specimens and room-state steel plotted together to compare strength development. Due to comparatively small strain range, all specimens had stable loops except for Q345-H1, which coupon buckled during compression at 2.5% strain. Under cyclic loading, yield plateau gradually disappeared and strength developments closely depended on strain amplitudes, cycles, and loading histories. Yield strengths and post cyclic strengths of

471

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

(a) Cyclic loading device

(b) Cyclic tension random (TR)

(c) Cyclic ascend high amplitude(AS-H)

(d) Cyclic ascend low amplitude (AS-L)

(e) Cyclic alternate high amplitude(AL-H)

(f) Cyclic alternate low amplitude (AL-L)

Fig. 3. Cyclic loading device and protocols.

Table 2 Loading conditions of different steel grades. Steel type (designed thickness)

Q235 (16 mm) AG235 (10 mm) Q345 (12 mm) Q390 (14 mm)

Cyclic random TR p p p p

Cyclic ascend AS-H p

Cyclic alternate AS-L

AL-H p

p p p

AL-L p p

p

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

Q235 Stress (σ/MPa)

600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

R H1 H2 0.1

0.2

0.3

0.4

0.5

AG235

R H1 H2 0.1

0.2

0.3

Strain

(a) Q235

(b) AG235 Q345

R H1 H2 0.1

600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

Strain

Stress (σ/MPa)

Stress (σ/MPa)

Stress (σ/MPa)

472

0.2

0.3

0.4

0.5

600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

0.4

0.5

Q390

R H1 H2 0.1

0.2

Strain

0.3

0.4

0.5

Strain

(c) Q345

(d) Q390

Fig. 4. Monotonic stress-strain curves of different steel types.

Table 3 Mechanical properties of specimens under monotonic loading. Steel type

Case

Yield strength fy (MPa)

Ultimate strength fu (MPa)

Ultimate strain eu

fy/fu

fy/fy(R)

fu/fu(R)

eu/eu(R)

Q235

R H1 H2

280.00 205.86 207.24

437.38 383.48 390.74

0.194 0.201 0.199

0.640 0.537 0.530

1 0.735 0.740

1 0.877 0.893

1 1.036 1.026

AG235

R H1 H2

270.30 167.11 168.90

447.72 368.10 375.81

0.249 0.291 0.275

0.604 0.454 0.449

1 0.618 0.625

1 0.822 0.839

1 1.168 1.104

Q345

R H1 H2

395.80 226.70 221.40

542.74 378.88 377.37

0.139 0.221 0.215

0.729 0.598 0.587

1 0.573 0.559

1 0.698 0.695

1 1.590 1.547

Q390

R H1 H2

401.70 250.40 269.90

543.68 430.62 442.59

0.156 0.183 0.183

0.739 0.581 0.610

1 0.623 0.672

1 0.792 0.814

1 1.172 1.172

heated steels were lower than that of corresponding untreated steels. Normal state R specimens had their strength mainly increased with the strain amplitude, whereas only slight strength increase at certain strain level. On the other hand, for heated steels, expansion of stress-strain loops was evident with every increased strain level and expansions were also apparent during constant amplitude cycles with gradual saturation, indicating isotropic hardening behavior. In this study, we defined strain amplitude related hardening as strain hardening and expansion during constant amplitude cycles as cyclic hardening. Due to significant cyclic hardening effect, comparatively large strength gap at yielding state between heated cases (H1 and H2) and R specimen gradually got narrowed with increasing strain amplitudes and loading cycles. These phenomena are commonly observed in general low-yield point steel [15].

2.6. Residual behaviors After the completion of low frequency cyclic tests, those specimens were unloaded firstly and then further pulled out in tension until complete fracture. These post-cycle tension tests were performed in Engineering Center of Tianjin University and stressstrength relations of the final pull-out stage were defined as residual performance. Due to lack of long-range extensometer this time and to protect the electronic device, the extensometer was removed from tested specimens once stress-strain curve passed maximum stress value and presented obvious strength reduction. Residual tension curves are given in Fig. 6. Both Q235 and AG235 untreated and heated specimens still presented obvious yielding behaviors after cyclic ascend or cyclic alternate loading process with sudden tangent stiffness change on stress-strain relation.

473

400 Q235 Cyclic ascend

400 Q235 Cyclic alternate

300

300

300

200

200

200 100 0 -100 -200 R H1 H2

-300 -400 0.0

0.5

1.0

1.5

2.0

2.5

100 0 -100 -200 R H1 H2

-300 -400

3.0

-3

-2

True strain (%) 400

AG235 Tension random

400

True stress (MPa)

100 0 -100 -200

R H1 H2

-300 -400 0.5

1.0

1.5

2.0

2.5

-3

AG235 Cyclic ascend

400 300

200

200

100 0 -100 -200

R H1 H2

-400

3.0

-1.5

600

Q345 Tension random

-1.0

-0.5

0.0

R H1 H2

-600 0.5

1.0

1.5

2.0

2.5

0.5

1.0

-400

1.5

-1.5

600

Q345 Cyclic ascend

150 0 -150 -300 R H1 H2

-600 -1.5

-1.0

True strain (%) 600

600

-0.5

0.0

0.5

1.0

-1.5

-600 1.0

1.5

True strain (%)

-1.0

2.0

2.5

3.0

600

Q390 Cyclic ascend

150 0 -150 -300 -450

R H2

-600 -3

-2

-1

0

1

-0.5

0.0

0.5

1.0

1.5

True strain (%)

True stress (MPa)

True stress (MPa) R H2 0.5

R H1 H2

-600

1.5

300

-450

1.5

Q345 Cyclic alternate

-450

300

-300

1.0

-300

300

-150

0.5

0

450

0

0.0

-150

450

150

-0.5

150

450

0.0

-1.0

True strain (%)

Q390 Tension random

-0.5

R H1 H2

True strain (%)

-450

3.0

3

AG235 Cyclic alternate

-300

300

-300

2

-200

300

-150

1

-100

300

0

0

0

450

150

-1

100

450

0.0

-2

True strain (%)

-300

True stress (MPa)

True stress (MPa)

-400

3

450

-0.5

R H1 H2

True strain (%)

-450

True stress (MPa)

2

-200

300

True strain (%) 600

1

True stress (MPa)

True stress (MPa)

200

0.0

0

0 -100

True strain (%)

300

-0.5

-1

100

-300

True stress (MPa)

-0.5

True stress (MPa)

400 Q235 Tension random

True stress (MPa)

True stress (MPa)

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

2

True strain (%)

3

Q390 Cyclic alternate

150 0 -150 -300 -450

R H2

-600 -3

-2

-1

0

1

2

3

True strain (%)

Fig. 5. Cyclic responses of different steels.

Specimens of these two grades also presented faster strength growth than corresponding direct pulling specimens; data are shown in Fig. 4. In contrast, for Q345 and Q390 steel specimens, yield plateau disappeared after cyclic loading process. Specimens of AG235, Q345, and Q390 displayed similar strength-developing patterns despite different cyclic loading protocols and heat treatment processes. Q235 gave similar residual strength behavior between two heat-treated conditions but presented different

strength degradation patterns between tension random loading and the other two loading cases. After experiencing cyclic ascend and alternate loading cycles, heated Q235 steel specimens had their residual strength developed fast with an early achievement of maximum strength. However, its strength deterioration was also quick, presenting weak loading sustaining ability. Random loaded specimens had smaller but more gradual strength hardening performance, as well as better loading sustaining ability.

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

Q235

R-TR R-AS R-AL H1-TR H1-AS H1-AL H2-TR H2-AS H2-AL

0.1

0.2

0.3

0.4

Stress (σ/MPa)

Stress (σ/MPa)

474

0.5

600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

AG235

R-TR R-AS R-AL H1-TR H1-AS H1-AL H2-TR H2-AS H2-AL

0.1

0.2

600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

Q345

R-TR R-AS R-AL H1-TR H1-AS H1-AL H2-TR H2-AS H2-AL

0.1

0.2

0.3

0.4

0.5

Strain

0.3

0.4

0.5

Stress (σ/MPa)

Stress (σ/MPa)

Strain 600 550 500 450 400 350 300 250 200 150 100 50 0 0.0

Q390

R-TR R-AS R-AL H2-TR H2-AS H2-AL

0.1

0.2

0.3

0.4

0.5

Strain

Strain Fig. 6. Residual stress-strain relations of different steels.

Table 4 Residual properties comparing to direct tension data. Steel type

Specimen

R(yielding)

R(ultimate)

R(strain)

Steel type

Specimen

R(yielding)

R(ultimate)

R(strain)

Q235

R-AS R-AL R-TR H1-AS H1-AL H1-TR H2-AS H2-AL H2-TR

1.29 1.29 1.11 1.70 1.70 1.28 1.70 1.76 1.38

1.01 1.02 0.99 1.08 1.08 1.02 1.07 1.08 1.02

0.76 0.82 0.96 0.63 0.63 0.88 0.65 0.61 0.84

Q345

R-AS R-AL R-TR H1-AS H1-AL H1-TR H2-AS H2-AL H2-TR

1.04 1.06 1.07 0.92 1.08 0.93 1.21 1.20 1.07

1.00 0.99 1.00 1.04 1.00 0.98 1.03 1.03 1.02

1.07 1.07 1.04 1.01 0.87 1.03 0.98 1.00 1.00

AG235

R-AS R-AL R-TR H1-AS H1-AL H1-TR H2-AS H2-AL H2-TR

1.29 1.28 1.22 1.89 1.74 1.64 1.79 1.73 1.53

1.01 1.01 1.01 1.09 1.09 1.07 1.07 1.07 1.05

0.61 0.59 0.60 0.48 0.50 0.52 0.50 0.53 0.57

Q390

R-AS R-AL R-TR H2-AS H2-AL H2-TR

1.02 1.01 0.92 1.24 1.23 1.06

1.02 1.01 1.00 1.05 1.05 1.03

1.00 0.95 1.01 0.95 0.87 1.02

To better understand cyclic loading history influence on mechanical properties, yield strength (r0.2), ultimate strength (maximum stress), and corresponding ultimate strain (strain when ultimate stress happened) were captured and compared with those of directly tensioned cases. Comparisons were presented as the ratio of different factors (as in Table 4), where R (yielding) gives residual yield strength ratio to direct yielding strength. Four steel grades all presented slight increase of ultimate strength after small amplitude cycles, but the elevation was weak. While the strain hardening and cyclic hardening effect during hysterics loading can indeed increase the residual yield strength. This increase of yielding strength was especially obvious for Q235 and AG235

steels. Heat-treated specimens were more sensitive to cyclic loading protocols and history, with more than twice strengthening effect for heated specimens than normal specimens. For Q345 and Q390 steel grades, residual mechanical factors were nearly unchanged (R and H1 serial specimens) or slightly increased (H2 specimens), indicating little influence from cyclic loading history. The forwardness of ultimate strain was prominent for AG235 steel, obvious for Q235 steel, and nearly impervious for Q345 and Q390. Therefore, residual behaviors indicated that this strength-reducing heat treatment was more applicable for Q345 and Q390 on seismic design usage. Whereas for heated Q235 and AG235 steels, obvious strength hardening and quick deterioration behavior after cyclic

475

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

strength at each strain amplitude levels were obtained and plotted in Fig. 7. Both cyclic hardening and softening behaviors were closely related to steel material types, cyclic strain amplitudes, and loading protocol histories. Under random tension loading conditions, strain range varied between 1% (±0.5%, 1.5–2.5%, 0.5–1.5%) to 2% (0–2%), with an increasing mean strain level. Therefore, during the first two cyclic stages (± 0.5%, 1.5–2.5%), slight cyclic hardening was observed and the majority hardening extent was limited to the first two cycles of each strain range. During 1.5–2.5% strain

3. Results and discussion 3.1. Cyclic material strengths To better compare strength developments and relations between different treated state cases, yield strength and ultimate R

H1

H2

Q235-Ascend ce

True stress (MPa)

1.5%-2.5% ±0.5%

0

0.5%-1.5%

0-2%

4

8

12

500 400 300 200 100 0 -100 -200 -300 -400 -500

16

0.125% 0.25%

R

0.5%

0.75%

AG AG235-Random

0

H1

H2

4

8

12

16

AG AG235-Ascend ce

40 400

1.25%

20

28

24

R

H1

H2

20 200

200

0.5%-1.5%

0-2% ±0.5%

-100 -200

True stress (MPa)

200 1.5%-2.5%

10 100 0 0.125% 0.25%

0.5%

-100 10

0.75%

1%

1%

1.25%

-200 20

8

12

0

R

H1

4

8

12

Q345-Ascend ce

H2

1.5%-2.5% 0.5%-1.5% 0-2%

±0.5%

0

4

8

12

500 400 300 200 100 0 10 -100 20 -200 30 -300 40 -400 50 -500

16

0.125% 0.25%

4

1.5%-2.5%

0

4

0.5%-1.5%

8

Cycle

20

24

H1

12

16

500 50 400 40 300 30 200 20 100 10 0 -100 10 -200 20 -300 30 -400 40 -500 50

20

25

R

H1

H2

100 0 0.5%

10 -100

0.5% 1%

1%

1.25%

20 -200

0

28

5

10

8

12

0.125% 0.25%

0.5%

500 400 300 200 100 0 -100 -200 -300 -400 -500

1%

16

1%

1.25%

20

24

4

8

25

0.75%

16

R

H1

H2

1%

5

1%

1.25%

10

15

20

25

Cycle

R

12

20

0.5%

0.5%

0

28

Q390-Alternate lt

H2

1%

1.25% 2.5%

0

15

Q345-Alternate lt

H2

0.5%

Q390-Ascend ce

H2

0-2%

±0.5%

AG AG235-Alternate lt

Cycle R

15

Cycle

0.75%

0

True stress ss (MPa)

Q390-Random

16

R

Cycle 500 50 400 40 300 30 200 20 100 10 0 -100 10 -200 20 -300 30 -400 40 -500 50

10

Cycle

True stress (MPa)

Q345-Random

5

40 -400

-400 40

16

cycle

500 400 300 200 100 0 -100 -200 -300 -400 -500

1.25% 2.5%

True stress (MPa)

4

0.5%

30 -300

-300 30 0

H2

1.5%

400 300

0

H1

cycle

30 300

100

R

0.75%

0

300

True stress (MPa)

True stress (MPa)

R

-400

True stress (MPa)

1%

500 50 400 40 300 30 200 20 100 10 0 -100 -200 -300 -400 -500

cycle

-300

True stress ss (MPa)

Q235-Alternate

H2

2.5%

Cycle 400

H1

20

24

28

cycle Fig. 7. Ultimate strengths during each cycle.

True stress ss (MPa)

True stress (MPa)

Q235-Random

500 400 300 200 100 0 -100 10 -200 20 -300 30 -400 40 -500 50

True stress (MPa)

loads could weaken strength reduction effect from heat treatment and even bring negative effect on post-earthquake performance.

500 400 300 200 100 0 -100 10 -200 20 -300 30 -400 40 -500 50

R

H2

0.5%

0.75%

1.25%

1.5% 2.5%

0

5

10

15

Cycle

20

25

476

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

and cyclic hardening or softening behavior were closely related to strain amplitude. Under cyclic ascend loading conditions, tested specimens all presented strength hardening during each strain amplitude elevation and the following constant strain loading stage. Both strain hardening and cyclic hardening effects were evident at large strain amplitude loading period (2.5% cycles for Q235 steel and Q390 steel). The last cyclic stage had a sudden tightened loading range, especially for Q235 and Q390 steels (strain amplitude decreased from 2.5% to 1.25%). Then, both strain softening and cyclic softening behavior were obvious and leaded to a gradual

range, mean strain level was elevated but with tightened strain amplitude; all steel specimens presented a cyclic softening effect for tension stress, but displayed nearly unchanged or slight cyclic hardening behaviors for compression strength. On the other hand, during 0.5–1.5% loading stage, strain amplitude was unchanged. However, mean strain decreased from 2% to 1%, tension strength was unchanged or presented a slight cyclic hard3ening behavior, and compression strength displayed a cyclic softening effect. Cyclic ascend and alternate loading cases had symmetrical loading with the mean strain kept at zero. Steel strength development 1.1

1.0 0.9 0.8 Q235-Random R 0.7 H1 H2 0.6 0 4

8

12

16

Cycle

1.1

Nominal stiffness (E/E0)

Nominal stiffness (E/E0)

Nominal stiffness (E/E0)

1.1

1.0 0.9 0.8 Q235-Ascend R 0.7 H1 H2 0.6 0 4 8

16

20

24

0

28

0.9 0.8 AG235-Random R H1 H2

8

12

0.9 0.8 0.7 0.6

16

AG235-Ascend R H1 H2

0

4

8

10

15

20

25

20

25

20

24

20

25

0.9 0.8

12

16

20

25

AG235-Alternate R H1 H2

0

5

0.8

1.0 0.9 0.8 0.7 0.6

12

16

20

24

Q345-Alternate R H1 H2 0

28

5

10

Nominal stiffness (E/E0)

0.9 0.8

12

16

1.1

1.0 0.9 0.8 0.7 Q390-Ascend R H2 0.6 0 4 8

15

Cycle

1.1

1.0

15

1.1

0.9

Q345-Ascend R 0.7 H1 H2 0.6 0 4 8

10

Cycle

1.1

8

0.7

Cycle

1.0

Cycle

Cycle

0.8

28

Nominal stiffness (E/E0)

Nominal stiffness (E/E0)

1.0

0.7 Q390-Random R H2 0.6 0 4

16

1.1

8

0.9

Cycle

1.1

Q345-Random R 0.7 H1 H2 0.6 0 4

1.0

0.6

12

Nominal stiffness (E/E0)

4

Nominal stiffness (E/E0)

Nominal stiffness (E/E0)

Nominal stiffness (E/E0)

5

Cycle

1.0

Cycle

Nominal stiffness (E/E0)

Q235-Alternate R H1 H2

0.7

1.1

0

Nominal stiffness (E/E0)

0.8

Cycle

1.1

1.0

0.6

0.9

0.6

12

1.1

0.7

1.0

12

16

20

24

28

Cycle Fig. 8. Evolution of stiffness E under cyclic loading.

1.0 0.9 0.8 0.7 Q390-Alternate R H2 0.6 0 5 10

15

Cycle

477

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

strength reduction during this period. AG235 and Q345 had a comparative weak strain amplitude reduction extents (from 1.25% to 1%), therefore strength softening effect was inconspicuous; AG235 steel even presented slight strength hardening during the last loading stage. Similar to varying trends for cyclic ascend case, specimens under cyclic alternate loading experienced strain hardening and cyclic hardening behavior as loading amplitudes increase. Otherwise, strength softening behavior was presented, and varying extents were still closely related to strain amplitudes. For all cyclic loading conditions, cyclic hardening or softening during same strain range cycles mainly resulted from steel isotropic hardening effect. This cyclic hardening or softening was mainly achieved dur-

σ

B

Ei C O

D

ε

E'i A

3.2. Steel stiffness

Fig. 9. Representative hysteretic loop and key mechanical index calculation diagram.

0-2

1.5-2.5

0.125 0.25 0.5 0.75 1.0

0.5-1.5

0-2

2.5 1.25

0.75

1.5-2.5

Strain history (%)

0.5-1.5

Equivalent damping coefficient

AG235-R AG235-H1 AG235-H2 Q345-R Q345-H1 Q345-H2

0.60 Cyclic Ascend 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

1.5

Q235-R Q235-H1 Q235-H2 Q390-R Q390-H2

0.5

2.5

1.25

Strain history (%)

2/π

2/π

±0.5

Q235-R Q235-H1 Q235-H2 Q390-R Q390-H2

0.60 Cyclic Alternate 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

Strain history (%)

Strain history (%)

0.60 Tension random 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

2/π

Equivalent damping coefficient

0.60 Cyclic Ascend 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

2/π

AG235-R AG235-H1 AG235-H2 Q345-R Q345-H1 Q345-H2

0.125 0.25 0.5 0.75 1.0 1.25 1.0

Strain history (%) Fig. 10. Comparison of energy dissipation indexes.

Equivalent damping coefficient

±0.5

Q235-R Q235-H1 Q235-H2 Q390-R Q390-H2

Equivalent damping coefficient

Equivalent damping coefficient

0.60 Tension random 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

During ductile failure, local micro-cracking and micro-voids formation led to progressive deterioration on mechanical

2/π

2/π

Equivalent damping coefficient

ing the first one or two cycles for each strain amplitude, followed with a steady state during remaining cycles. Commencement and saturation of cyclic hardening or softening during each strain amplitude change were partially attributed to maximum plastic strain memory effect of isotropic hardening. Specimens of both heat-treated types presented obvious strength reduction during or after yielding state. Similar to monotonic performances shown in Fig. 3, the two-heated steel presented almost the same strength development, with H2 being slightly larger than H1. Compared with corresponding untreated steel, heattreated Q235 and AG235 steels presented larger cyclic hardening behavior during early cycles, leading to fast strength development, with the strength gradually catching up with corresponding R specimen. Yield strength reduction after heat treatment was weak for Q235 steel. In addition, after experiencing apparent cyclic hardening, strength of the heated steel was nearly the same with R specimens during later loading cycles. This significant hardening behavior might be beneficial in terms of load-carrying capacity and energy dissipation, however, it has potentially negative implications for capacity design in seismic applications at the same time. Heated Q345 and Q390 steels had almost coordinated strength varying trends with corresponding R specimen, presenting a stable strength reduction performance during yielding state and post-cyclic loading state. From cyclic strength development point, strength reducing heat treatment was more suitable for low alloy Q345 and Q390 steels on seismic design application.

0.60 Cyclic Alternate 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.5

1.0

AG235-R AG235-H1 AG235-H2 Q345-R Q345-H1 Q345-H2

0.5

1.25

Strain history (%)

1.0

478

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

performance, where one of the damage representative factor is stiffness. Some nonlinear damage models for ductile metals described the damage as a constitutive variable that related material degradation to stiffness reduction [14,17]. Therefore, stiffness or elastic modulus of both tensioning and compressing processes during each cycle are captured and plotted in Fig. 8. The results indicated that steel stiffness remained unchanged during elastic state. When steel yielded and entered plastic state, steel stiffness decreased and the reduction extent was also closely related to loading protocols, strain history, and strain amplitudes. In general, elastic modulus during unloading process from tension to compression (Ei0 ) was higher than that of reverse loading process as from compression to tension (Ei). This difference was prominent during large amplitude cycles. Heated steels were more likely to have stiffness reduction and larger disparate Ei0 and Ei than that of untreated R specimens under symmetric cyclic loading conditions. Under random tension loading, when kinematic hardening or yielding surface shift dominated, the heated steels presented less stiffness reduction than that of R specimens. Most stiffness degradation happened during cycles within ±0.5% strain range, then stiffness varied in different way for different loading conditions. Under cyclic random loading case, stiffness presented staggered Ei0 and Ei with their average value kept nearly unchanged (Q235 and Q390) or followed a slight decreasing trend (AG235 and Q345) during 0–2% strain range cycles. The next two loading stage had similar tightened strain amplitude but varied mean strain; also, no obvious stiffness reductions were presented with their values fluctuated at 0.9 times of the original elastic mod-

900

900

Q235-R

800

Q235-H2

800

0.125%

700

700

2.5%

True stress(MPa)

True stress(MPa)

ulus before yielding. Under elevated mean strain loading, difference between Ei0 and Ei decreased and both stiffness values (Ei0 and Ei) at this loading stage were similar to (all the R specimens, Q235-H1, AG235-H1, AG235-H2, Q345H2) or larger than (Q235H2, Q345-H1, and Q390-H2) the high value of Ei0 at last cyclic loading stage. It0 s interesting to notice that AG235 and Q345 steel specimens had the two stiffness Ei0 and Ei staggered again during the last cyclic loading period. Under gradually increasing amplitude loading, the stiffness of all steel types gradually decreased before ±0.5% amplitude cycles. During later large stain amplitude cycles, the decreasing trend stopped and the nominated steel stiffness fluctuated steadily around 0.9. Strain amplitude of Q235 and Q390 experienced a dramatic increase (from ±1% to ±2.5% during the 20–24 cycles), resulting in sudden aggravated difference between E0i and Ei, as well as later decrease on average stiffness. Strain amplitude dropped back to a half during the last loading stage (24–28 cycles); then, averaged modulus stopped decreasing and in-between difference got narrowed. The average stiffness of AG235 and Q345 also presented slight decrease during elevated strain amplitudes (±1.25%) and remained unchanged during later tightened loading stage. Q345 displayed a constantly small disparity between E0i and Ei, however, difference in AG235 steadily increased where the gap did not close during the following tightened amplitude cycles. Similar varying trends were indicated for alternate loading cases. However, during 10–15 cycles and 20–25 cycles, the reduced strain amplitudes led to less disparity between E0i and Ei, as well as less reduction on average stiffness.

600 500

1.25%

400

1% 0.75% 0.5%

300 200

500

1.25% 1%

400

0.75% 0.5%

300 200

0.25%

100

2.5%

600

0.25% 0.125%

100

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

Strain (%)

Strain (%)

(a) Q235-R 1000 900

(b) Q235-H2 1000

Q345-R

900 800

True stress(MPa)

True stress(MPa)

800 0.125% 700 600 500

1.25%

400

1%

300

0.75% 0.5% 0.25%

200 100 0 0.0

Q345-H2

0.5

1.0

1.5

2.0

Strain (%)

(c) Q345-R

2.5

700

0.125%

600 500

1.25%

400

1%

300

0.75% 0.5%

200 100 3.0

0 0.0

0.25% 0.5

1.0

1.5

2.0

Strain (%)

(d) Q345-H2

Fig. 11. Relative hysteresis loops for Q235 and Q345 steel specimens (Cyclic ascend case).

2.5

3.0

479

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

3.3. Material dissipation behavior In seismic design, one of the key index for member or material performance is energy dissipation capacity, which reflects the ability to reduce seismic effect. Equivalent damping coefficient he is often used to quantify and compare cyclic behavior and energy dissipation ability. As shown in Fig. 9, enclosed area S(ACD+BCD) represents the practical dissipated energy. SOBD and SOAC are the triangular areas, and their summation represents a reference elastic strain energy quantity that is convenient for normalization. Then, he is given as the following equation and calculated indexes for tested specimens are given in Fig. 10.

SðACDþBCDÞ 2pSOBDþOAC

ð1Þ

‘‘Masing” properties are discussed in reference [18] on cyclic deformations in metals. If a material demonstrates ‘‘Masing” property, its hysteresis loops obtained for different strain ranges coincide if they are plotted in relative coordinates. The coincidence indicates unchanged yielding and strength hardening characteristics. To investigate the ‘‘Masing” properties of tested materials, stabilized loops under different strain amplitudes or loading stages were translated to zero point. Cyclic ascend cases of Q235(R, H2) and Q345(R, H2) steels were selected as representatives, as shown in Fig. 11. The relative hysteresis loops for Q235 steel specimens did not coincide with each other, which reveals non-Masing properties. Separation indicated increase in elastic domain, which mainly attributed to increase in the saturated state of isotropic hardening. Heated specimen (Q235-H2) presented more yield surface expansion. On the other hand, for Q345 steel, stable loops obtained at different strain ranges nearly coincided with each

400

400

1.5%-2.5% 300 0.5%-1.5%

300

200

200

True stress (MPa)

True stress (MPa)

For both normal unheated specimens and heated specimens of all steel grades, equivalent energy dissipation index increased with strain amplitudes. he presented slight reduction trend under repeated cycles at the same strain level; this decreasing trend was evident in early cycles, where the effect mainly came from obvious cyclic hardening at the beginning stage. Given the gradual saturation after certain cycles, he index remained unchanged across cycles at the same strain. For Q345 and Q390 steel grades, energy dissipation indexes of heat-treated specimens were larger than those of corresponding untreated ones, especially at early loading stages. Specimens of both heating processes had similar energy dissipation performances. For untreated specimens, significant difference was presented between energy dissipation index

3.4. Masing properties and yielding surface moving characteristics

100 0 -100 -200

1.5%-2.5% 0.5%-1.5%

100 0 -100

0-2% ±0.5%

-200

0-2% -300

±0.5%

-400 -1.0

-0.5

0.0

0.5

-300

Q235-R 1.0

-1.0

0.0

Strain (%)

(a) Q235-R

(b) Q235-H2

0.5

1.0

500

1.5%-2.5%

400 300

300

True stress (MPa)

-0.5

Strain (%)

500 400

Q235-H2

-400

200 100

0-2%

0

±0.5%

-100 -200

True stress (MPa)

he ¼

magnitudes of different steel grades. While after heat treating process, the he index variations trended to be similar for all four steel grades. When strain amplitude was elevated above 1%, he of the heated specimens could generally reach 0.45 (about 70% of the full rectangular cyclic loop condition), which was quite close to that of low-yield point steel LYP160 according to reference [15]. Then, from material energy dissipation performance, both heat-treating methods achieved a satisfying performance, especially for Q345 and Q390 steel grades.

1.5%-2.5%

200 100 0 -100

0-2%

-200

±0.5%

-300

-300 -400

0.5%-1.5%

-500 -1.0

-0.5

0.0

0.5

Q345-R 1.0

0.5%-1.5%

-400

Q345-H2

-500 -1.0

-0.5

0.0

Strain (%)

Strain (%)

(c) Q345-R

(d) Q345-H2

Fig. 12. Relative hysteresis loops for Q235 and Q345 steel (Tension random case).

0.5

1.0

480

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

Table 5 Calibration of Chaboche model parameters. Specimen Q235-R Q235-H1 Q235-H2 AG235-R AG235-H1 AG235-H2 Q345-R Q345-H1 Q345-H2 Q390-R Q390-H1

E Mpa

r0

R1 Mpa

b

C1 Mpa

c1

C2 Mpa

c2

C3 Mpa

c3

C4 Mpa

c4

Mpa

190,000 181,000 190,000 185,000 195,000 195,000 185,000 185,000 180,000 185,000 185,000

130 105 90 170 88.5 90 210 160 167 210 190

65 105 125 40 110 110 30 30 28 40 60

3.3 4 3 1 1 1 1 3 4 5 3.7

132,000 75,900 125,400 33,000 165,000 165,000 151,248 72,000 45,600 66,680 37,810

3300 3300 3300 3300 3300 3300 3151 3000 3151 2150 3151

60,000 61,750 57,000 78,300 69,300 69,300 87,000 15,000 25,000 82,800 45,000

1000 950 1000 900 900 900 1000 1000 1000 900 1000

8840 7980 8500 7200 4670 4670 24,500 8750 10,500 10,500 7200

170 210 170 120 80 80 350 250 350 150 120

2460 2300 1960 880 890 1470 3450 2875 2700 3450 1600

20 10 20 10 10 10 23 25 23 23 6.5

other. Heated specimen Q345-H2 displayed slight separation, however, the extent was small. Steel isotopic hardening describes yield surface enlargement, therefore, non-Masing separation indicated isotropic hardening extent on the performance of stabilized loops. Masing property can be used to evaluate isotropic performance, whereas kinematic hardening and yield surface shift can be indicated through tension random test data. Fig. 12 displays stable loop during different loading stages with their strain symmetric center reverting to zero point. Among the four loading stages in random tension loading case, the first and last two stages had similar strain amplitude as ±0.5% but different mean strain values. The first ±0.5% strain loading case presented relatively symmetric strength extent to the origin. While after experiencing ±1% cycles and relapse to ±0.5% amplitudes around 2% (1.5–2.5% period), the steel material displayed strain hardening and yield surface center shifting. Tension peak strength was larger than the corresponding compression peak. Furthermore, when the mean strain was reduced to 1% (0.5–1.5% period), material strengths were likely to go back to symmetric state. Material hysteretic behavior was determined by both kinematic and isotropic hardening. In addition, heating process presented slight reduction on center shifting behavior from kinematic hardening. 4. Calibration of cyclic constitutive models Material test data and related discussion on cyclic response indicated that the high temperature heating slow cooling treating method could present promising effect on reducing strengths (yield and maximum) and increasing energy dissipation ability, especially for Q345 and Q390 steel grades. Moreover, the heated specimens also displayed slight difference on hysteresis behavior when compared with corresponding untreated specimens. To better evaluate the practical application potential of this measurement on seismic design based on structural response, appropriate cyclic constitutive model is necessary for material numerical input in structural simulations. The nonlinear combined kinematic-isotropic hardening rule of Chaboche type was adopted for numerical simulation herein [19,20]. Chaboche model is widely applied in mechanical and engineering fields, and has also been included in general finite element software, such as ABAQUS and ANSYS. It is a two-surface model that utilizes hardening function to describe yield center movement (kinematic hardening effect) and yield surface expansion (isotropic hardening). The nonlinearity of the model is based on the concept of a yield surface in the stress space with position and size determined by the following hardening variables:

f ¼ Jðr  XÞ  k

ð2Þ

where the first part is the yield surface description and X is the back-stress tensor that gives the translation pattern. Chaboche model generally utilizes Von Mises yielding criterion and allows superposition of several independent back stress tensors to better describe kinematic hardening behavior:

X¼

X

Xi

ð3Þ

2 X_ ¼ cdep  cXdp 3

ð4Þ

where ep represents plastic strain and dp is the magnitude of plastic strain increment. Generally, three or four pairs of ci and ci can achieve a satisfying description of yield surface transitions. In ANSYS, yield expansion generally uses Voce law on nonlinear isotropic hardening, with an exponential saturation hardening term added to the linear term as follows:

k ¼ r0 þ R0 ep þ R1 ð1  ebep Þ

ð5Þ

where k is the yield surface size. r0 is the yield stress at zero plastic strain. R0 gives the stiffness coefficient for linear term. R1 is the maximum change in yield surface size and b defines the rate at which the size of the yield surface changes as plastic strain develops. From test data, isotropic hardening presented mainly nonlinear other than linear relation with its equivalent plastic strain. Then, R0 was obtained, and the following equation was derived:

k ¼ r0 þ R1 ð1  ebep Þ

ð6Þ

Therefore, together with elastic modulus, three factors for isotropic hardening and eight factors for kinematic hardening, a total of twelve parameters were calibrated from stabilized hysteresis loops of different strain amplitudes through data fitting [12,19]. Table 5 presents Chaboche parameters of tested heat-treated and normal steels. Material test simulations were performed in ANSYS with calibrated parameters for validation (comparison given in appendix). Isotropic hardening calibrated materials also indirectly indicated the comparatively lower yield strength on heated specimens (from yield strength r0) and comparatively larger cyclic hardening extents for Q235 and AG235 steels (from R1 values). Given that Chaboche model parameters were mainly calibrated from stabilized loops, precisely describing the initial small strain cycles before or around yielding was difficult because of considerable percentage of monotonic loading behaviors during early loading period. Moreover, classical isotropic hardening model in Eq. (6) was also unable to precisely describe the maximum plastic strain memory effect, which was obvious for the tested specimens. Despite these small deviations, good agreement was indicated from the comparison between test and simulation results, thereby indicating the accuracy of Chaboche hysteresis model and the applicability of those parameters in structural analysis.

481

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

heat treatment on seismic design, Chaboche cyclic constitutive model parameters were calibrated for future analyses on structural level. The following conclusions can be drawn from test data and discussions.

5. Conclusions

Q235-R-TR

400 300 200 100

500 300 200 100

0

1

2

3

-3

-2

-1

Strain(%)

Q235-H1-TR

400 300 200

0

1

2

3

-3

-100

Q235-H1-AS

1

2

3

200 100

200 100

-3

-2

-1

1

2

3

-3

200 100

Strain(%)

-200

-100

Test FEM

-400

AG235-R-TR

500

-500 500

1

2

3

300 200 100

300 200 100

-3

-2

-1

0

1

2

3 Strain(%)

-1.5

100

-200

-300

AG235-R-AS

200

-1.0

-400 -500

1

2

3 Strain(%)

0.5

1.0

1.5

Test FEM

-400 -500

500

-1.5

400

AG235-R-AL

300 200 100 -1.0

Strain(%)

-300

Test FEM

0

-300

300

0 -0.5 0.0 -100

-100 -200

Test FEM

400

0

Q235-H2-AL

400

Strain(%)

-500

Stress (MPa)

400

Test FEM

500

-300

-400

3 Strain(%)

0 0

-200

-300

2

-500

300

-1

1

-400

Q235-H2-AS

400

-2

0

-200

0 0

-100 -300

500

0

Q235-H1-AL

300

Stress (MPa)

300

Test FEM

400

Test FEM

-500

Q235-H2-TR

400

3 Strain(%)

500

Strain(%)

-400

Stress (MPa)

500

2

0 0

-200

Test FEM

1

-500

100 -1

0

-400

200

-2

-100 -200

-300

-500

Stress (MPa)

-1

-300

300

Strain(%)

-400

Stress (MPa)

-2

Test FEM

400

-300

-500

-3

0

-200

-400

3 Strain(%)

500

0

-200

2

-500

Stress (MPa)

Stress (MPa)

500

-100

1

-400

100

-1

100

Stress (MPa)

Test FEM

-500

-100

200

-300

-400

-1

300

0 0

-200

-300

-100

-100

Q235-R-AL

400

0

-200

-1

500

Stress (MPa)

-100

Q235-R-AS

400

0 -1

(1) Slow cooling heat treatment can effectively reduce both yield strength and ultimate strength, where strength reduction extents ranked decreasingly as Q345, Q390, AG235, and Q235, Q235, and AG235 steels presented obvious strain

Stress (MPa)

500

Stress (MPa)

Stress (MPa)

This paper investigated the application potential of a new strength reducing heat treatment on commonly used structural steel grades in China through a fundamental material cyclic loading tests. Focusing on stiffness, strain/cyclic hardening, and energy dissipation abilities, heat treatment influence on monotonic behavior, cyclic characteristics, and residual loading performances were studied. Finally, to explore the application potential of this

-0.5

0 0.0 -100 -200

0.5

1.0

1.5

Strain(%)

-300 Test FEM

Fig. 13. Comparison between test data and simulations with Chaboche model.

-400 -500

Test FEM

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

500

AG235-H1-TR

400 300 200 100

400 300 200 100

0 0

1

2

3

-1.5

-1.0

Strain(%)

-200

500

AG235-H2-TR

400 300 200

400

0

1

2

3

-1.5

200 100 -1.0

Strain(%)

-200

Test FEM

300 200

1.5

500

-1.5

400

0

1

2

3

-1.5

100 -1.0

-0.5

0.5

Test FEM

-400

Q345-H1-TR

500

-500 500

1.0

300 200 100

1.5

0

1

2

3 Strain(%)

-1.5

Q345-H1-AS

300 200 100 -1.0

0 -0.5 0.0 -100

0.5

1.0

1.5

Strain(%)

-200

-300

-300

Test FEM

Test FEM

-1.5

Q345-R-AL

300 200 100 -1.0

-0.5

0 0.0 -100

0.5

1.0

1.5

Strain(%)

-200 -300

400

0

1.5

Strain(%)

400

Test FEM

-500

Stress (MPa)

400

1.0

-500

-300

-400

0.5

-200

Strain(%)

-200

-300

0 0.0 -100

500

100

Strain(%)

-200

AG235-H2-AL

200

Q345-R-AS

200

-1.0

Test FEM

300

-400

300

0 -0.5 0.0 -100

1.5

Strain(%)

-500

Test FEM

400

0

1.0

-300

500

100

Stress (MPa)

1.0

-500

400

0.5

-400

Strain(%)

Q345-R-TR

500

-500

0.5

-400

Stress (MPa)

Stress (MPa)

-500

0 0.0 -100 -200

-300

-400

-400

0 -0.5 0.0 -100 -200

-300

-200

AG235-H2-AS

300

0

-100

-0.5

-300

-500

100

-1

-1.0

Test FEM

-400

Stress (MPa)

Stress (MPa)

500

-100

-1.5

100

Stress (MPa)

Test FEM

-500

-1

1.5

200

-300

-400

-100

1.0

AG235-H1-AL

300

Strain(%)

-200

-300

-1

0.5

400

Stress (MPa)

-100

0 -0.5 0.0 -100

500

Test FEM

-400 -500

500

Stress (MPa)

-1

AG235-H1-AS

Stress (MPa)

500

Stress (MPa)

Stress (MPa)

482

-1.5

Q345-H1-AL

400 300 200 100 -1.0

-0.5

0 0.0 -100 -200

0.5

1.0

1.5

Strain(%)

-300 Test FEM

-400 -500

-400 -500

Test FEM

Fig. 13 (continued)

hardening and cyclic hardening effects, however, reduced strength gradually caught up with untreated steels under cyclic loading. Heated Q345 steel and Q390 specimens presented obvious reduction on both yield and ultimate strengths; this reduction effect can be maintained during different cyclic loading protocols, especially for Q345 steel grade. Residual performance after cyclic loading present elevated yield strength, early strength growth, and quick deterioration behavior, which was prominent for AG235, obvious for Q235 steel, and nearly impervious for Q345 and Q390

steels. Strength reducing heat treatment was more applicable for Q345 and Q390 for seismic design usage. (2) Heated steel presented similar strain hardening and cyclic hardening trends for all four steel grades. Cyclic curves were affected by loading histories, where hardening characteristics were related to both plastic strain and strain amplitude. Maximum plastic strain memory effect indicated that cyclic hardening would initiate again after the increase of loading strain amplitude. Material elastic stiffness was reduced during yielding and elastic-plastic transition state and gradually

483

500

Q345-H2-TR

400 300 200 100

300 200 100

0 0

1

2

3

-1.5

-1.0

Strain(%)

-200

Test FEM

Stress (MPa) -1

-1.5

100 -1.0

-0.5

Strain(%)

600

Stress (MPa) 2

3

-3

150

300 150

1

2

3

-3

-2

-450

Q390-H2-TR

600

1

2

3 Strain(%)

0

-600

500

Q390-H2-AS

300 150 -1

3 Strain(%)

0

1

2

-150

Test FEM

-600

450

-2

2

-450

400

3 Strain(%)

-1.5

Q345-H2-AL

300 200 100

0 -3

1

-300 Test FEM

-300

Test FEM

-1 -150

Strain(%)

-300 Test FEM

Q390-R-AL

450

0 0

-150

Strain(%)

Test FEM

600

Q390-R-AS

300

-1

1.5

Strain(%)

-500

450

-2

1.0

-400

0 1

0.5

-300

-500

Q390-R-TR

0 0.0 -100 -200

Test FEM

-400

Stress (MPa)

Stress (MPa)

-500

600 500 400 300 200 100 0 -100 0 -200 -300 -400 -500 -600

1.5

200

-300

-400

-1

1.0

-200

-300

600 500 400 300 200 100 0 -100 0 -200 -300 -400 -500 -600

0.5

300

Stress (MPa)

-100

0 -0.5 0.0 -100

Q345-H2-AL

400

Stress (MPa)

-1

500

Q345-H2-AS

400

Stress (MPa)

500

Stress (MPa)

Stress (MPa)

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

-1.0

-0.5

0 0.0 -100 -200

0.5

1.0

1.5

Strain(%)

-300

-450

Test FEM

-600

-400 -500

Test FEM

Fig. 13 (continued)

fluctuated around 0.9 times of the initial elastic modulus under later large cycles. Elastic modulus of compressing process was higher than that of tensioning process; this difference became prominent during large amplitude cycles. (3) After the strength-oriented heat treating process, material energy dissipation indexes could generally reach 0.45 for all four steel grades, proving the ability of this method on creating low yield point similar to steel from high strength structural steels. The 2nd heating process had more acceptable peak temperature and shorter heating input. It presented similar material strength weakening effect for all four steel grades as well, thereby indicating the promising practical application potential of this heat-treating process on structural practice. (4) Based on cyclic stress-strain test data, cyclic constitutive models of both untreated and two heat-treated steels were calibrated. The related parameters are now available for use by the structural engineering community for nonlinear structural analysis.

Acknowledgements This work is supported by the National Natural Science Foundation of China (Grant No. 561272264), China Postdoctoral Science

Foundation (2016M590202) and Visiting Scholar program at Columbia University in United states from China Scholarship Council (File No. 201606250065). And the authors also appreciate the professional help from Prof. Tasnim Hassan at North Carolina State University in USA.

Appendix A. Validation of the Chaboche hysterical parameters See Fig. 13.

References [1] T.T. Song, G.F. Dargush, Passive Energy Dissipation Systems in Structural Engineering, John Wiley & Sons, New York, 1998. [2] D. Piedrafita, X. Cahis, E. Simon, J. Comas, A new perforated core buckling restrained brace, Eng. Struct. 85 (2015) 118–126. [3] A. Farzampour, J.A. Laman, Behavior prediction of corrugated steel plate shear walls with openings, J. Constr. Steel Res. 114 (2015) 258–268. [4] S.L. Jones, G.T. Fry, M.D. Engelhardt, Experimental evaluation of cyclically loaded reduced beam section moment connections, J. Struct. Eng. 128 (4) (2002) 441–451. [5] E. Saeki, M. Sugisawa, T. Yamaguchi, A. Wada, Mechanical properties of low yield point steels, J. Mater. Civ. Eng. 10 (3) (1998) 143–152. [6] H. Yan, P. Pan, Y.Q. Wang, T. Makino, X. Qi, Tests of buckling-restrained braces using low-yield point steel as core material, Appl. Mech. Mater. 166–169 (2012) 3159–3165.

484

Y. Yu et al. / Construction and Building Materials 140 (2017) 468–484

[7] S.-J. Chen, C. Jhang, Cyclic behavior of low yield point steel shear walls, ThinWalled Struct. 44 (7) (2006) 730–738. [8] M. Morrison, D. Schweizer, T. Hassan, An innovative seismic performance enhancement technique for steel building moment resisting connections, J. Constr. Steel Res. 109 (2015) 34–46. [9] M. Morrison, D. Schweizer, T. Hassan, Seismic enhancement of welded unreinforced flange-bolted web steel moment connections, J. Struct. Eng. (2016). 04016102. [10] Y. Yu, Z. Chen, X. Wang, Investigation and evaluation of a novel technique on CFRT column connection, J. Constr. Steel Res. 113 (2015) 195–208. [11] Y. Yu, Z. Chen, X. Wang, Effect of strength weakening oriented heat treatment on carbon steel and application on aseismic design, Constr. Build. Mater. (Under review). [12] Y. Shi, M. Wang, Y. Wang, Experimental and constitutive model study of structural steel under cyclic loading, J. Constr. Steel Res. 67 (8) (2011) 1185– 1197. [13] G. Shi, M. Wang, Y. Bai, F. Wang, Y. Shi, Y. Wang, Experimental and modeling study of high-strength structural steel under cyclic loading, Eng. Struct. 37 (2012) 1–13.

[14] F. Zhou, Y. Chen, Q. Wu, Dependence of the cyclic response of structural steel on loading history under large inelastic strains, J. Constr. Steel Res. 104 (2015) 64–73. [15] M. Wang, L.A. Fahnestock, F. Qian, et al., Experimental cyclic behavior and constitutive modeling of low yield point steels, Constr. Build. Mater. 131 (2017) 696–712. [16] L. Xu, X. Nie, J. Fan, M. Tao, R. Ding, Cyclic hardening and softening behavior of the low yield point steel BLY160: experimental response and constitutive modeling, Int. J. Plast 78 (2016) 44–63. [17] D. Bao, S. Zuyan, Analysis of damage cumulation for steel tower structure subjected to earthquake, Special Struct. (in Chinese) 13 (3) (1996) 30–32. [18] R.P. Skelton, H.J. Maier, H.J. Christ, The Bauschinger effect, Masing model and the Ramberg–Osgood relation for cyclic deformation in metals, Mater. Sci. Eng. A 238 (2) (1997) 377–390. [19] J.L. Chaboche, Time-independent constitutive theories for cyclic plasticity, Int. J. Plast. 2 (2) (1986) 149–188. [20] J.L. Chaboche, Constitutive equations for cyclic plasticity and cyclic viscoplasticity, Int. J. Plast. 5 (3) (1989) 247–302.