Fracture Assessment of Weld Joints of High-Strength Steel in Pre-Strained Condition

Abstract

Unstable fractures tend to occur after ductile crack initiation or propagation. In most collapsed steel structures, a maximum 15% pre-strain was recorded, at the steel structural connections, during the great earthquake of 1995, in Japan. Almost-unstable fractures were observed in the beam-to-column connections, where geometrical discontinuities existed. Structural collapse and unstable failure occurred after large-scale plastic deformations. Ship structures can also suffer from unstable fractures in the welded joints. The fracture resistance of butt-welded joints subjected to tension in the pre-strained condition was estimated by considering the toughness deterioration, due to pre-strain and toughness correction for constraint loss in a tension specimen. The target specimen for this fracture assessment was a double-edged, through-thickness crack panel, with a crack in the weld joint (heat-affected zone (HAZ)). The critical fracture toughness value (crack tip opening displacement (CTOD)) of a large structure with pre-strain, which was applied to the HAZ region, was estimated from a small-scale, pre-stained, three-point bend specimen. Fracture toughness values, evaluated by a CTOD test, were recently mandated for shipbuilding steel plates. The critical fracture toughness value is a very useful parameter to evaluate the safety of huge ship structures.

1. Introduction

In general, 600–780 MPa-class high-strength structural steels are developed by a thermomechanical controlled process (TMCP), in steel mills [1]. Recently, the demand for large steel structures has been increasing, and in order to build a large steel structure, the strength must be increased. The application of high-strength steels, such as the 600 MPa-class steel, can be found in recent structures. These steels are used to increase the design stress and reduce transportation cost, owing to their light weight. Steel structure designs generally employ high-strength steel as the structural element. For the safe application of these steels in steel-framed structures, safety during earthquakes (dynamic loading and pre-strain occurrence) and toughness of the weld joint, especially in the heat-affected zone (HAZ), should be ensured. In general, the toughness decreases in the HAZ as the steel undergoes cycles of heat and strain, during multi-pass welding. In multilayer welding, the texture and toughness of the HAZ due to multithermal cycles show very complex distributions. As a result of measuring the toughness change by the crack tip opening displacement (CTOD) test for the multilayer welded HAZ, the decrease in toughness in the HAZ was remarkable [2,3,4]. The HAZ is characterized by unequal growth of the austenite formed in the dual phase, where C and Mn are not evenly distributed and concentrated on both sides, which thus, remains unstable in the low temperature region, without being decomposed into ferrite at high temperatures. The austenite instead transforms into martensite. This is defined as the M–A (Martensite–Austenite) constituent, which is a very poor phase. From a metallographic viewpoint, the reason why the HAZ is the most vulnerable phase is due to the M–A.

The promotion of a brittle fracture is a very dangerous problem that occurs, owing to dynamic and pre-strain effects. The fracture driving force increased, owing to the dynamic loading effect, and the fracture toughness decreased, owing to the pre-strain effect.

Unstable fractures of high-toughness steel structures in weld joints tend to occur after ductile crack initiation or propagation. During the great earthquake, as reported in [2,5,6], damage to steel structures occurred in weld joints after large strain deformation. The pre-strain in the cyclic loading, during the earthquake, occurred at the strain concentration area, like the weld toe in the weld structure. The great earthquake occurred in 1995. In most collapsed steel structures, a maximum of 15% pre-strain was recorded at the steel structure connections [3,4]. An almost-unstable fracture was observed in the beam-to-column connection areas where geometrical discontinuity exists. In the collapsed structures, unstable failures occurred after large-scale plastic deformation. This earthquake was characterized by the brittle fracture of structural steel under high-speed ground motion (104 kines (cm/s)) and large ground displacement (27 cm).

As a complement to the classical, unstable crack-initiation/propagation failure toughness, brittle crack-arrest failure toughness has become an important mechanical property to address material fracture and failure mechanisms, to avoid unstable fractures in structures [7,8,9,10,11,12,13,14,15,16,17].

The objectives of this study are as follows: (i) investigation of the change in the strength and toughness of the TMCP steel welds, with pre-strain and dynamic loading; (ii) development of a fracture assessment procedure for TMCP high-strength steel welds, under seismic conditions; and (iii) verification of the fracture assessment procedure by a large-scale component test. These can establish the procedure for the fracture performance assessment of TMCP high-strength steel welds, under seismic conditions. Attention was focused on the impact of the HAZ softening on structural integrity after pre-straining. The applicability of WES 2808 [18] to 780 MPa TMCP steel and its welds were examined.

2. Manufacturing of the HSB 600 High-Performance Steel Welds

2.1. Material Properties of the Specimens Used

The test specimen was produced by a conventional welding process, namely, submerged arc welding (SAW) [19]. This welding method was selected because SAW applies a relatively large heat input in the construction method of large steel structures, and adversely affects the safety of the welded joints. The mechanical properties and chemical composition of 25 mm-thick, high-strength steel (HSB600 high-performance steel for bridge structures) are listed in

Table 1. Mechanical properties of the HSB600 steel plate used.

Specimen Symbol	Young’s Modulus E, (MPa)	Yield Strength YS, (MPa) *	Tensile Strength TS, (MPa)	Yield-to-Tensile Ratio Y/T *
BM*-1	204,600	620	679	0.91
BM-2	207,800	613	693	0.88
BM-3	205,000	632	691	0.91
BM-4	206,700	576	684	0.84
BM-5	205,100	593	680	0.87
BM-6	208,300	589	688	0.86
Average	206,250	604	686	0.88

* BM: base metal, YS: 0.2% proof stress, Y/T= YS/TS.

and

Table 2. Chemical composition of the HSB600 steel plate used (wt. %).

Steels	C	Si	Mn	P	S	Fe
HSB600 (25 mm)	0.15≤	0.75≤	2.00≤	0.30≤	0.007≤	bal.

, respectively. The tensile test was carried out six times in the rolling direction of the steel plate by ASTM E8 [20], and with round-bar type test specimens, with a strain rate of 0.007 (1/s), at room temperature, (approximately 20 °C). The average yield strength and tensile strength were 604 MPa and 686 MPa, respectively. The Charpy impact energy was carried out by ASTM E23 [21], and the value was 47 J at −5 °C, the temperature required for the test condition of the steel bridge by Korean Industrial Standards (KS D3868) [22]. Figure 1 shows the groove geometry and macro section of the weld joint. The weld joint was made by base metal plates, with the dimensions 1,000 mm in length × 500 mm in width × 25 mm in thickness, with a single bevel groove angle of 17°. The SAW technique was used and the weld joint had a seven-pass welding layer. The general weld joint macro section is shown in Figure 1b. The applied heat input was 50 kJ/cm, which was used in the wide production filed to construct the steel welding structure. The welding conditions are listed in

Table 3. Welding condition for 25 mm steel plate.

Welding Process	Current (A)	Voltage (V)	Speed (cm/min)	Heat Input (kJ/cm)	Preheat Temp. (°C)	Interpass Temp. (°C)
SAW	700	34	29	50	100	100

. The chemical composition of the welding consumable is shown in

Table 4. Chemical composition of the welding consumable (wt. %).

Welding Consumable	C	Si	Mn	P	S	Fe
SAW wire	0.08	0.32	1.67	0.01	0.007≤	bal.

[14] and the used flux was AWS A5.23 F8A4-EA3-G.

2.2. Double-Edge Through-Thickness Crack Panel within a Crack in the HAZ

The fracture resistance of the butt-welded joints subjected to tension in the pre-strained condition was estimated by considering two effects: (1) toughness deterioration due to pre-strain (ductile-to-brittle transition temperature shifts due to pre-strain), and (2) toughness correction for constraint loss in the target specimen.

The target specimen for fracture assessment was a double-edged, through-thickness crack panel (ETCP) with a crack in the HAZ, as shown in Figure 2. The crack length 2a ranged from 10 mm to 100 mm. It was assumed that 3.0% tensile pre-strain was applied to the HAZ region. The pre-strain amount was measured with line interval lengths near the HAZ area. Lines were drawn at intervals of 3 mm in the loading direction of the specimen, before tensile loading, to measure the pre-strain amount in the HAZ. As the width of the HAZ was 3 mm, an interval of 3 mm was used to measure the pre-stain amount of HAZ. The tensile load applied until the 3 mm interval became 3.09 mm, especially near the HAZ.

3. Pre-Strain Effect on Fracture Toughness

3.1. Critical CTOD of 3.0% Pre-strained HAZ Specimen

In order to evaluate the fracture resistance of the 3.0% pre-stained ETCP, toughness deterioration due to the pre-strain should first be estimated. Generally, fracture toughness is required by the CTOD value in the line pipe and the offshore structural steel plate [23]. Therefore, in this study, fracture toughness was evaluated using CTOD. The CTOD fracture toughness of the pre-strained HAZ (fusion line + 1 mm) was measured by a three-point bend (3PB) test. The 3PB tests were conducted to measure the CTOD fracture toughness of 3.0% pre-strained specimens, with full thickness. The thickness of the 3PB specimen, B, was 22 mm, because the specimens should be manufactured by surface finishing in welded joints, which have angular distortion. The 3PB specimen had a deep through-thickness crack; a/W = 0.5, where a and W are the crack depth and the specimen width, respectively. The 3PB test temperatures were −60 °C, −40 °C, −20 °C, and 0 °C. The specimen was cooled with liquid nitrogen in a cooling bath. The yield strength of HSB600 steel at the test temperature was used for the calculation of the elastic component of CTOD. The critical CTOD at the fracture was calculated in accordance with ISO12135 [24] and ISO15653 [25]. The fracture toughness results of the pre-strained HAZ are shown in Figure 3. The specific temperature at which the critical CTOD was 0.1 mm, tended to increase with increasing pre-strain. The 3.0% pre-strain shifted the specific transition temperature by approximately 40 °C. These results indicate that the deterioration of the fracture toughness due to pre-strain should be taken into account, during the safety assessment against fracture from the weld HAZ, in the pre-strained condition.

Table 5. Critical crack tip opening displacement (CTOD) data of 3.0% pre-strained HAZ obtained by the three-point bend test.

Temperature	Critical CTOD (3PB)
T (°C)	δcr (mm)	δcr ave. (mm)
−20	0.012	0.079
−20	0.049
−20	0.177
0	0.047	0.1
0	0.108
0	0.145

summarizes the critical CTODs of the 3.0% pre-strained HAZ used in the assessment [19].

The critical CTOD of the ETCP would be larger than that of the 3PB standard fracture toughness specimen that was subjected to tension, owing to the plastic constraint loss around the crack tip in the ETCP. ISO 27306 [15] provides an engineering method to correct the CTOD toughness for the constraint loss in structural components. The toughness correction ratio, defined as the equivalent CTOD ratio β, was standardized in ISO 27306. Figure 4 shows a monograph of β₀ for an ETCP, including a reference size of the crack. β₀ is a function of YR and m, where YR and m are the yield-to-tensile ratio and the Weibull shape parameter, respectively.

3.2. Estimate of Critical CTOD in Pre-strained ETCP

The equivalent CTOD ratio β_2a for the ETCP with crack length 2a was determined using Equation (1) [13].

$${\beta }_{2a\left(ETCP\right)}={\beta }_{0\left(ETCP\right)}·{\left(2a/11\right)}^{k_{ETCP}\left(m, YR\right)}, k_{ETCP}\left(m, YR\right)=\frac{-0.57+3.1YR-1.45Y{R}^2}{\exp \left\{-0.35\left(m-10\right)\right\}+1}$$

(1)

According to ISO 15635 [25], the yield strength used in the calculation of the CTOD—when located in (or is partially in) the transformed HAZ—is higher than the yield strength of the base metal and weld metal (WM) strength applied, as it is difficult to evaluate the strength of the actual HAZ. When under-matched, weld-joint, pre-strain is applied, most of the deformation occurs in the base material, and the strength of the base metal becomes almost similar to that of the WM. Therefore, the yield strength of the base material was used in the CTOD calculation in this study and was estimated from the stress–strain curves of base steel, without the pre-strain, as shown in Figure 5. The yield stress of the pre-strained HAZ σ_Y^pre can be obtained using Equation (2) [18].

$${\sigma }_Y^{pre}=S_Y\left({\epsilon }_{pre}\right)$$

(2)

where S_Y(ε_pre) is the true stress of the original HAZ at the true pre-strain ε_pre. The tensile strength of the pre-strained HAZ σ_T^pre can be determined by Equation (3) [18];

$${\sigma }_T^{pre}={\sigma }_Y\frac{1+{\epsilon }_T}{1+({\epsilon }_T-{\epsilon }_{pre})}$$

(3)

where σ_T and ε_T are the tensile strength and uniform elongation of the original HAZ, respectively. The estimated YR of the 3.0% pre-strained steel was 0.90, where σ_Y^pre and σ_T^pre are 684 MPa and 759 MPa, respectively. The equivalent CTOD ratio β₀ for the ETCP with YR = 0.90 was 0.08 for m = 20 (Figure 4). The equivalent CTOD ratios β_2a for the 3.0% pre-strained ETCP, with different crack lengths 2a, were calculated using Equation (1); the results are listed in

Table 6. Equivalent CTOD ratio β for 3.0% pre-strained ETCP (YR = 0.9, m = 20), as a function of crack length 2a.

2a	β0 (ETCP)	β2a (ETCP)
10	0.08	0.07
20	0.08	0.15
30	0.08	0.22
40	0.08	0.30
50	0.08	0.37
60	0.08	0.45
70	0.08	0.52
80	0.08	0.60
90	0.08	0.68
100	0.08	0.75

.

The critical CTOD δ_cr,ETCP of the 3.0% pre-strained ETCP with the double-edged crack in the HAZ was estimated from the critical CTOD δ_cr,3PB for the 3.0% pre-strained 3PB specimen, given in Table 5, in the form of δ_cr,ETCP = δ_cr,3PB/β_2a. In this estimation, the median of δ_cr,3PB at each temperature was used. Figure 6 presents the estimated critical CTOD for the 3.0% pre-strained ETCP, as a function of the crack length 2a, at the temperatures 0 °C and −20 °C. The critical CTOD δ_cr,ETCP decreased with the increasing crack length 2a.

4. Results and Discussion

In this estimation, the fracture toughness of the pre-strained HAZ obtained by the experiments was employed. WES 2808-2003 [13] provides the estimation method for the pre-strained toughness, as shown in Figure 7. The static fracture toughness of the base metal at the reference temperature of T − ∆T_PD can be used as the fracture toughness in the pre-strain and the dynamic conditions at the service temperature T. The ∆T_PD is the temperature shift of the fracture toughness caused by the pre-strain and the dynamic loading. In WES 2808, the temperature shift ∆T_PD correlates with the flow stress elevation, ∆σ_f^PD = (∆σ_Y + ∆σ_T)/2, by pre-strain and dynamic loading. ∆σ_Y and ∆σ_T are the increase in the yield stress and tensile strength, respectively. The temperature shift ∆T_PD is given by Equation (4) [18]. This formula is applicable to the structural steels of 400 to 590 MPa-strength class.

$$\Delta T_{PD}\left({}^{°}\mathrm{C}\right)=\{\begin{array}{c}0.4\Delta {\sigma }_f^{PD} :0\le {\sigma }_f^{PD}\le 100 \left(MPa\right)\\ 40 :100\le {\sigma }_f^{PD}\le 300 \left(MPa\right)\end{array}$$

(4)

According to WES 2808, the fracture toughness of the pre-strained HAZ of HSB600 steel was predicted from the HAZ toughness, without the pre-strain. First, the flow stress of the pre-strained HAZ was estimated under the assumption that the flow stress is the same as that of the pre-strained base steel. The yield stress and tensile strength, as well as the flow stress σ_f^PD, of the pre-strained weld HAZ are summarized in

Table 7. Mechanical properties of HSB600 steel with and without pre-strain.

BM-2	σY (MPa)	σ0.2 (MPa)	σT (MPa)	σfPD (MPa)	YR (=σY/σY)
εpre = 0%	605	613	693	649	0.87
εpre = 1.5%	649	–	748	699	0.87
εpre = 3.0%	684	–	759	722	0.90

σ_Y : Yield stress, σ_0.2: 0.2% proof stress, σ_T : Tensile strength, σ_f^PD : (= σ_Y + σ_Y)/2.

; these were calculated using Equations (2) and (3), on the basis of the stress–strain curve shown in Figure 5. Then, the temperature shift ∆T_PD for the pre-strained HAZ was calculated using Equation (4) [13], and the values are shown in Figure 8. The median of the critical CTODs estimated for the 3.0% pre-strained HAZ at −20 °C and 0 °C are listed in

Table 8. Critical CTOD for 3.0% pre-strained HAZ obtained by experiment and estimated from δ_c without pre-strain.

Temperature	Critical CTOD for 3.0% Pre-Strained HAZ, δcr,med (mm)
T (°C)	Experiment	Estimated from δcr,med without Pre-Strain
−20	0.049	0.035
0	0.108	0.11

.

Then, from the estimated critical CTODs, δ_cr,ETCP for the 3.0% pre-strained ETCP with a crack in the HAZ was estimated with the equivalent CTOD ratio β_2a. The β_2a used was the same as that listed in Table 6. As shown in Figure 9, the δ_cr,ETCP for the 3.0% pre-strained ETCP, derived from the toughness values estimated on the basis of WES 2808, was almost the same as that estimated from the experimental data of critical CTOD, for the 3.0% pre-strained 3PB specimen (Figure 6).

5. Conclusions

The pre-strain effect on the fracture assessment was determined in 25 mm HSB 600 high-performance steel plate and weld joints. The static fracture toughness of the base metal at a reference temperature of T − ∆T_PD can be used as the fracture toughness in the pre-strain and dynamic conditions at the service temperature T. The ∆T_PD was the temperature shift of the fracture toughness caused by pre-strain and dynamic loading. According to the WES 2808, HSB600 steel fracture toughness with pre-strained HAZ was predicted from the HAZ toughness without pre-strain. These results indicate that the prevention of unstable failure owing to pre-strain, should be considered in the failure safety assessment of the weld in the pre-strained HAZ. In addition, the fracture toughness values of large structures with the pre-strain effect can be estimated from small-scale specimens (3PB), despite the high-strength steel plate weld joints.