Minimizing Distortion by Transient Thermal Tensioning and Its Effect on Fatigue Crack Growth Behavior of Flux Cored Arc Steel Weld Joints

N Subeki, Jamasri, M.N. Ilman, P.T. Iswanto #4 Department of Mechanical and Industrial Engineering, Universitas Gadjah Mada Jl. Grafika No. 2, Yogyakarta, 55281, Indonesia 1) nursubeki@gmail.com 2) jamasri_tmugm@ugm.ac.id 3) Coresponding author: noer_ilman@ugm.ac.id 4) priyotri@ugm.ac.id * Department of Mechanical Engineering, University of Muhammadiyah Malang, Jl Raya Tlogomas 246 Malang 65144, East Java, Indonesia

such as TTT treatment would be more effective since it is carried out during process hence no additional work is required after welding. Therefore, the present investigation is aimed to study distortion and residual stress, and its effect on fatigue crack propagation resistance.

II. MATERIAL AND METHODS
Materials used in this research were A36 Steel plates with the dimensions of 5 x 120 x 300 mm. The steel plates have a yield stress of 248 MPa and a maximum tensile stress of 400 MPa with 20% in elongation. The filler metal used in this experiment was K-71T (AWS A5.20 / ASME SFA-5:20 E71T-1C) with a diameter of 1.2 mm. The experiment set-up is shown in Fig. 1.
The welding parameters including voltage, current and heat input used were 40 volts, 210 Ampere and 2.184 kJ/mm respectively. FCAW welding process was conducted by using welding speed of 3.846 mm/s. The secondary heating temperatures used were 100 °C, 200 °C, and 300°C. The thermal cycles were measured using thermocouples located at the distance of 10 mm, 35 mm, 55 mm and 80 mm from weld centerline.
After welding, distortion measurement was conducted with the dial indicator in longitudinal and transversal directions. Test on fatigue propagation rate was conducted with Servopulser machine. The specimen used was center crack tension (CCT) according to ASTM E 647-00 standard model. The frequency (f) and stress ratio (R) is maintained at 11 Hz and 0.1.

III. RESULTS AND DISCUSSION
The chemical composition of both weld metal and base metal are shown in Table 1. It can be seen that the amount of C in weld metal is low whereas the percentage of Mn is 1.583 % and this is considered to be the optimum Mn content steel weld metal. This weld metal is known as C-Mn steel where its tensile strength and toughness are controlled by the Mn/C ratio as reported by Surian, at al. [14]. Referring to Table 1 the percentage of C of weld metal is be lower with higher Mn content compared to the base metal. The weld thermal cycles under various secondary heating temperatures at 80 mm were measured by using thermocouple located in four points, namely TC1, TC2, TC3 and TC4 at the distance of 10 mm, 35 mm, 55 mm and 80 mm respectively as shown Figure 1. Figure 2 shows weld thermal cycles with and without TTT treatment. As shown in Figure 2a the temperatures of untreated welded plate are 782 °C, 302 °C, 180 °C and 143 °C of note is that the temperature distribution can affect distortion and residual stress whereas thermal cycles influence weld cooling rate which determine the mechanical properties of weld [15]. In TTT condition at heating temperature of 100 °C as shown in Fig. 2b, the peak temperature from the closest distance the weld centerline is 685 °C, while increasing distance up to 35 mm from the weld centerline decreases temperature to 290 °C. Further decrease in observed at the distance of 55 mm where the peak temperature at this position is 235 °C and finally at the distance of 80 mm the peak temperature is 166 °C. The effect of TTT on welded thermal cycles can be studied by comparing Fig. 2a-2d. It can be seen that temperature does not affect significantly the peak temperature of region near the weld area (TC1) but the effect of increasing TTT temperature as observed at the distance far from the weld TC2, TC3 and TC4 changes temperature distribution. These temperature changes resulted from TTT treatment are expected to generate thermal tension which help to reduce distortion and residual stress.
Results of 3D and 2D dimension curves of the welded plates prepared using the conventional welding and the TTT treatment can be seen in Fig. 3. As shown in Fig. 3 (a-d) the distortions for all welded plate specimens were measured along the longitudinal and transversal directions. The highest distortion in longitudinal and transverse direction are 10.56 mm and 5.8 mm respectively observed in as-welded condition as seen in Fig. 3 (a). TTT weld treatment could decrease distortion. These results are consistent with previous report which shows that distortion could be reduced by thermal tensioning [4]. In this research, the optimum TTT treatment occurs at temperature 200 °C marked by the lowest distortion. The results of measurements distortion at the edge of the welded plates along a longitudinal direction in Fig. 3 (e) shown the longitudinal distortions are seen to be concave with maximum distortions occur at the middle of the plates. At TTT temperature of 200 °C and 300 °C, the longitudinal distortions are very low where their values at the middle part of the welded plate length are -6.16 mm and 0.34 respectively. These findings indicate that the temperature distribution, especially temperature gradient (dT/dx) influences the formation of distortion. It can be seen that a low or almost no distortion is observed when the heating temperature of TTT achieves 200 °C [6][7] [8]. Microstructures of the weld metals with and without TTT treatment are shown in Fig. 4. It can be seen that microstructures of as welded metal consists of mainly bainite. In contrast, a TTT treated weld metals are composed of acicular ferrite (AF) as the major microstructure with a small amount of grain boundary ferrite (GF) and Widmanstaten ferrite (WF). It has been reported that acicular ferrite gives high strength, good impact toughness [16] and fatigue performance [2], due to its fine grained and interlocking structures. In comparison with the as welded weld metal, the amount of acicular ferrite in TTT treated weld metal is higher. Hardness test were conducted in weld metal (WM), coarse HAZ, fine HAZ, base metal (BM). It can be seen that the highest hardness values are observed in as welded weld joint. This is because the weld cooling rate in as welded condition is relatively fast resulting in bainite microstructure. Increasing TTT heating temperature decreases hardness. Results of tensile stress for as welded and TTT-treated welds under various heating temperatures are shown in Fig. 6. As shown in figure, increasing TTT heating temperature increase both yield and tensile strength. These results seem to confirm that cooling rate under TTT treatment plays decisive role in determining weld yield and ultimate strength due to formation of fine-grained acicular ferrite.  (Fig. 7a) and also the relation between propagation rate of fatigue crack (dɑ/dN) and stress intensity factor range (ΔK) as shown in Fig. 7b. The trend lines were taken from region II of da/dN -ΔK curves as shown in Fig. 7c. Figure 7a shows that the fatigue life (given in the number of cycles) of the weld joint under TTT treatment at heating temperature of 200 °C is the highest. In Fig. 7b-7c the fatigue crack growth rate is influenced by C and n values. Compared with as welded weld joint, fatigue crack propagation of the weld treated using TTT with 200 °C heating is lower. These results seen to be consistent with previous report [6] [7][8] [11]. The Paris constant of C and n can be seen in Table 2. The n value for the weld treated using TTT heated is smaller than that of as welded, suggesting that fatigue crack growth rate is improved.
Fracture surfaces of weld metals with and without TTT treatment are show in Fig. 8. As shown in this Fig. 8a, fracture surface of as weld metal specimen shows that they have large striations typically around 30 µm in which with the crack direction from left to right. On the other hand fracture surface of the weld treated by TTT at heating temperature 200 °C shows that they have finer striation spacing. It is also looked that fracture spacing was reducing by TTT treatment. This result could be correlates with the value of Paris constant as shown in Table 2.