Journal or Materials Proeesshag Teehnelogy ELSEVIER Journ,il of .Materials Pr~lce,,,,iug -i'ccht~olo~ 68 11997~ 43 49 On crack susceptibility in the submerged arc welding of medium-carbon steel plates M.E. Khallaf M.A. Ibrahim h, N.A. EI-Mahallawy h M.A. Taha u., " I lw El,'.vption lrms and .S'tCt'/ I I'o.r/~.+. th'/wtm. ('uiro. Et.'ff,t h Dtpartmt'nl ~1 Design atul Producth,n En.eim,criog. Facultl. t , / E n g b w o l t t g . . Jitl-Sham ~ I itil o ~ i l i..41,tt.~t,la. ('~tir,. [?g.ipt Rccci~etl g Seplember 1995 Abstract The present paper describes cracking behaviour during the submerged art: x~cldingof medium carbon steel plates (11.45".,(~. It discusses the results of tests made to examine the effect of welding variables (current. welding speed and wire feed rate}, plate fabricatiop, conditions (rolling reduction ratio) and plate thickness o,1 cracking susccptibilily using trans-varestraint tests. It is foun6 that the cracking susceptibility increases with an increase in the welding current, and decreases with an increase in the welding sF~.ed or the electrode wire feed ,-ate. It also increases with increases in the plate rolling :'eduction ratio and ~vith decreases in the plate thickness. These results have a practical signilicance t\~r industrial ticlds, especially where welded machine spate parts are concerned. :(;, 1997 Elscvier Science S.A. Kcj'u'ord.~: Submerged arc welding: Welding of mediom carbon ~,teel: Welding of steel plales: Crack suso.'ptibilit} in x~elding I. Introduction Engineers concerned with the design and manufacture of welded machine ;pare parts arc aware of the existence of cracking problems in the welded joints. In the submerged arc welding process, using higher deposition rates and greater heat input leads to welding conditions that are more likely to lead to cracking. Previous investigations have been concerned with the nature and mechanisms of solidification cracking. During solidification of weld metal, grains begin to grow from the fusion boundary towards the central region of the weld pool. Many alloying elements, as well as impurities, are rejected ahead of the growing crystals and become concentrated in the last part of the liquid to solidify. Their presence lowers the freezing temperature of this liquid. As solidification is taking place, the weld and surrounding material progressively cool, thu~ giving rise to contraction strains across the weld [1]. The presence of residual low-melting-point liquid may then lead to lowering of the ductility, so that the * Corr~-sponding author. Fax: mtaha@ASUNET.SHMS.eun.eg + 20 4152991: c-mail: 0924-0136!97!$17.00 ~t) 1997 ElsevierScience S.A. All rights reserved. PII S0924-0136(96)02530-7 contra,:fion strains produce ,:racking [2]. In other words, the ductility is low due to the residual liquid, but. on the other hand. the amount of residual liquid is insufficient to lill any voids produced by cracking. It is c~ideut that this problem must depend on the weld metal composition [3,4]. In the present work the effect of welding variables, plate thickness and the plate forming reduction ratio (during fabrication) on the crack susceptibility is presented and discussed. 2. Experimental work 2. I. Wehli:lg equipment The automatic welding system used for the DC submerged arc welding ~wocess has a control unit by which the welding variables (current, u'a~el speed and ~virc feed rate) can be controlled automatically. The main technical data of the welding equipment arc: wire dia m e t e r = 3--7 ram; travel speed = 0-120 m:h: welding current = 0-1250 A; and maximum voltage = 44 V. Copper coated steel electrodes of 3.2 mm diameter were used throughout the experiments. The angle be- 44 M.E. Khallaf et aL / Journal of Materials Processing Technology 68 (1997) 43-49 tween the electrode and the test-piece is 40*; the overhanging length of the electrode beyond the nozzle is 25 nun and the distance between the electrode tip and the test-piece is 3 ram, all of these values being kept constant. 2.2. Test equipment The principles of the trans-varestraint test equipment of Farrer and Garland [5] have been used to design a simplified test rig for this work. Fig. l(a) shows the trans-varestraint test equipment used in the present work and the applied forces, as well as the sample dimensions and positioning. In the test equipment, the specimen is mounted by means of two clamping forces at its ends. The required strain with a suitable deflection (1.5 mm) is exerted in the middle of the specimen during the welding process, 4 r 2 -]-~\ 3 ~ P R F7 5 6 L.. (a) 300 - - [ Oetail o f t= 12,16,20mm. .groove where the electrode tip is at point (A), as shown in Fig. l(a). The deflection is kept constant (1.5 mm) during all experiments. 2.3. Welding specbnens Medium carbon steel, C 45, has been used for welding, its chemical analysis (wt%) being: C = 0.46, Si = 0.29, Mn = 0.6, P = 0.032 and S = 0.03. The electrode used was of type ASW A5. 17-80 EMI2, which has the following chemical analysis (wt%): C = 0.12, Si = 0.1 and Mn = 1.0. The welding specimens have a rectangular shape with different thicknesses (12, 16 and 20 ram). They were fabricated by rolling to different reduction ratios (11.3, 8.5 and 6.8). Fig. l(b) shows the details of the welding specimen. All of the test specimens were selected from the rolled plates such that the direction of welding is the same as the direction of rolling. This was considered in order to eliminate any possible effect of the rolling direction on the mechanical properties, and consequently on the cracking susceptibility. In the present experiments, the current varied from 450 to 550 A, the travel speed (s) varied from 24 to 36 m/h, the wire feed rate (wfr) varied from 10 to 20 kg/h and the voltage varied from 35 to 37 V. In order to study each welding parameter individually, the other parameters were kept constant. To study the effect of plate thickness, specimens with different thicknesses (t) of from 12 to 16 m m were obtained by machining from a plate which was already fabricated using a rolling reduction ratio of 6.8 with an original thickness of 20 ram. To study the effect of the rolling reduction ratio (RR), plates of 20 and 16 mm thickness were reduced to 12 mm thickness by employing reduction ratios of 6.8 and 8.5, respectively. Further to the above plates, there was the original plate of 12 m m thickness, which had been subjected to a reduction ratio of 11.3, together with another specimen of zero reduction ratio, obtained by shaping the original as-cast slab to the same thickness, 12 ram. 3. Results and discussion welding direction ~un.in_~o~ ,,,, Ill Fig. 1. Details of the experimental set-up. (a) Trans-variant strain equipment: 1, clamping plate; 2, specimen; 3, flux; 4, torch; 5, present travel distance; and 6, die block. (b) Test-piece: F i = holding forces; F2 = exerted force; P = penetration; and R = plate resistance thickness. The observed cracks are characterized by single route cracks (without branching), as shown in Fig. 2(a), microscopic observations revealing that the cracks are inter-granuPar, as indicated in Fig. 2(b). 3.1. Effect of welding current on crack length Fig. 3 indicates that, for all plate thicknesses used, the crack length increases with an increase in the welding current. The curves obtained in Fig. 3 are almost M.E. Khalh!/'et al..,rJotti'Hal OJ Materials Processing Techm,!o~,q. 6g (19971 43-49 45 Fig. 4. Effect of the welding current on the penetration ~lt: (~1) 45[) A" (b) 5(1it A, and It) 550 A. (a) (b) Fig. 2. Micrographs showing: (a) a single-route crack ( x 75): (b) an inter-granular crack ( × 150). parallel, except at the lowest current. In this latter case. little penetration is obtained, which corresponds to a great plate resistance thickness, the plate resistance thickness being the unmelted part of the base metal under the weld bead, Fig. 1. In contrast, in the case of greater currents, greater penetration is obtained resulting in a weaker plate resistance thickness, as shown in Fig. 4, these results being due to the direct effect of the current on the heat input during the welding process. The amount of heat put into the weld joint has many effects, one of the most important of which is the 50 40 nilulion ratio % Cr~.k k.lUa ( u ) 80 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S-24m/b w . f . r : 16kK/l~ 10 0 ___~; _._..~--- ~ - "-,. . . . . . . . . . . _ / ~ --4"-~ /- -~ 30 20 change in the dilution rati(,~ i.e. the ratio of the crosssectional area of the molten base metal in the weld bead to the total cross-sectional area of the weld bead. which in turn affects the chemical composition of the weld bead, the latter having a major role in changing the cracking susceptibility [1,4,6-9]. This means that an increase in the heat input (which is caused by an increase in the current) leads to an increase in the dilution ratio, i.e. a greater part of the base metal is molten and added to the weld bead. Therefore. by increasing the current the share of the base metal in the chemical composition is increased relative to the wire. Fig. 5 shows that, under the present experimental conditions, the dilution ratio increases with current from about 50 to 75%. As a result of the increase in dilution ratio, the carbon content in the weld metal increases, Fig. 6(a), and the crack length consequently increases, Fig. 6(b). The carbon content was also calculated and the results are super-imposed on Fig. 6(a), indicating that the calculated values arc very close to the measured values. As the carbon content increases, the solidification temperature range, which is :'ritical for the occurrence of hot cracking [6], increases [I,8.10]. Fig. 7 shows that as the carbon content measured in the weld metal increases, the crack length increases. This is in agreement with the relationship between the dilution ratio and the crack length, shown in Fig. 6(b). This figure shows that any increase in the dilution ratio is accompanied by an increase in the crack length. For a welding current starting from 450 A up to 550 A, micro examinations reveal that, as the current increases, the grain orientation becomes steeper, as shown . . . . ~-.~_ ........ ~5_-~.. ~ 72.?". . . . - 40 ~ / 450 475 R.R ..--~f SOl) Curr~t 20 + llJ' &|2 m.m 8.6 & 16 m.m o 6 . 8 & 2 0 m.m 525 550 (amp) Fig. 3. Effect o f the current on the crack length for different plate thicknesses and different reduction ratios. 0 . 450 &t m.m. o I1.3 &12 m.m. + 8.5 & 16 m.m. $-24m/h w.Lr.=iQkl~lh . . . * 6.8 475 .S00 525 & 20 m . m . 550 Cur~nl (Imp) Fig. 5. Effect o f the current on the dilution ratio for different different plate thicknesses and rcdvetion ratios. 46 M.E. KhaUaf et al./ Journal of Materials Processing Technology 68 (1997) 43-49 'trice hi weld 0.4 metJl ("k) w.f.r=lOkglb o Actual 0.3 61 64 71 (a) 76 74 mlu~on (%) SO Crick length (m.m.) . f f 40 jJ 30 F i g . 8. Micrograph showing the grain orientation observed at 5 5 0 A . J for a high depth/width ratio of lhe weld bead ( x i 50). R.R-6.8 S=24 m/b w.f.r.=lO kglk t=12 m.m. 20 56 58 60 (b) 62 66 aU.aon (%) F i g . 6. E f f e c t o f the dilution ratio on: (a) the carbon content in the welding metal; and (b) the crack length. in Fig. 8. Such steeper orientation is found to promote solidification cracking, similar to that found in a previous investigation [! l]. 3.2. Effect of the welding travel speed Crzek length (re.m) R.R~ 6.8 S=24 mlh tO ~ + t-12 m.m o t=16m.m * l= 20 m.m 1=550 A w.f.r - 1 0 kg/h 0 24 2"1 J.l ,t0 (a) 36 Third sped (m/b) ~enetration (re.m) 12 IO ,*----~.~... , - - ~----_.__._..__ ---. S: 6 4 + * R.R-6.8 1-550 A 2 24 t=lZm.m t-16m.m o l~2Om.m ,m..(.r~10kg/b 0 0.305 0.311 Carbon In wdd um,,tad(%) 0.1 w . L r - I O kg/h t--20 mn.m. ~- 0.321 0,3.51 Fig. 7. E f f e c t o f the carbon content in the weld metal on the crack length. R.R.-6.8 I0 27 30 Travel speed (m/b) ,t6 33 C,arl~n in weld metal (%) 0.2 15 0.29 20 ,/ 20 0 l 0.3 25 S Crm~k lemgth (re.m) ..10 O.4 j~J 30 SO 40 Fig. 9(a) describes the effect of welding travel speed on the crack length, from which figure it is noted that, as the travel speed increases, the crack length decreases. The results obtained can be attributed to the decrease in the heat input, which is related directly to the decrease in the travel speed. As the heat input decreases, both the penetration and the dilution ratio decrease, the carbon content in the weld metal as well as the depth/width catio of the bead also decreasing. Fig. 9b and c represent the effect of the welding travel speed on the penetration and the carbon content in the weld metal depth/width respectively. The results of Fig. 9a-c are mutually supportive, as it is indicated that the crack length increases as the penetration ratio or the carbon content in the weld metal increases. 35 Microscopic observations indicate that the microstrtlcture of the weld metal obtained at the lowest speed (2.4 m/h) consists of coarse columnar grains, as shown in Fig. 10(a). According to the strain theory for cracking susceptibility, such a structure is more susceptible to hot cracking [12,13]. At the highest travel speed, * t-12m.m o t-16m.m + t=20m.m R.R-6.8 i-550A w.f.r=10ks/h 0 (C) 24 27 JO 33 36 Trlvel speed {m/h) F i g . 9, E f f e c t o f the welding travel speed on: (a) the crack length; (t the penetration; and (c) the carbon content in the weld metal. M.E. KhallaJet al. . d o u r m d o/" Mtttcrittls Prot c~il~.: ' ;g (a) t'~.~' ~1997) 43 "" ,v dilution ratio decreases with a decrease in the wire feed rate, as illustrated in Fig. I I(b), this greatly affecting the resulting chemical composition of the weld me~.al. The wire used is characterised by depletion in the carbon content combined with enrichment in the Mn content. As the decrease in the dilution ratio means an increase m the wire content in the bead, then the bead composition will also be characterised by depletion in carbon and enrichment in manganese. This is supported by chemical analysis of the welded beads, Fig. I llc) indicating that increasing the wire feed rate results in increased %C and decreased %Mn in the weld metal. Previous investigations [4,6,10] have reported that the Mn content in the welded bead tends to decrease its cracking susceptibility. Another effect of wire feed rate, which is thought to play a considerable role in decreasing the cracking susceptibility, is the change in the bead dimensions and microstructure, which arc associated with the change in the wire feed rate. It is observed from the macroscopic observations of the welded beads that the greater the wire feed rate, the wider is the weld bead, whilst the weld depth remains constant, Fig. 12. This interpretaCrack I~gth (mm) 35 (tO Fig. 10. M i c r o g r a p h s ( x 150) s h o w i n g the effect o f the w e l d i n g t r a v e l s p e e d o n the w e l d m e t a l g r a i n size using c u r r e n t = 550 A. w i r e feed r a t e = I0 k g / h , a n d S = 7.4 m ; h (a), S = 36 m ' h (b). ....................................... 30 .......... 25 -LC - - ~ - .... 22Z - ~ -- -~-- _ - ..... .... ,,>_. _ -~ _ -- _ - -0 .... -Oe-_ - " .- ...... q 15 R.R =6,8 5 0 I ----5,5 O A S = 36m/b 10 125 15 (a) 70 Wire 175 20 feed rate (1¢~) Dilution ratio (%) rio - - " .t 50 --, . . __._...,~_. . . . . . .... . 40 i [ 30 R.R=6.$ l_-~r~0A S=~mTn 20 !0 * t=12m.m / o t=16m.m + t=-2Om.m ] 0 10 12.5 15 (13) I 3.3. Effect of the wire feed rate ~ 20 10 the microstructure of the weld metal consists of fine columnar grains, Fig. 10{b), due to a decrease in the heat input, this type of structure being less susceptible to cracking. Comparing the results given in Fig. 9(a) for the effect of the welding travel speed, with those given in Fig. 3 for the effect of the welding current, indicates that the crack length is more sensitive to variations in the current than it is to variations in the speed. From this it can be deduced that control of the crack length by changing the welding current is more effective than by changing the travel speed. Such a conclusion is of practical importance. --~ . . . . . . . . . . . ---~_3. . . . . 17.5 20 Wire feed rate (kg/h) C&Mn in weld metal Ma ~ -M ~ 0.8 - (%) ....... , _ i - 0.6 One of the factors that has an effect on the composition of the weld metal and therefore on the cracking susceptibility, is the wire feed rate. as the wire composition is different from that of the base metal. Fig. 11(a) shows the effect of wire feed rate on crack length: as the wire feed rate increases the crack length decreases. This can be attributed to the effect of the wire feed rate on the bead dilution and dimensions. The 0.4 c% O.2 --~-~--I,,..o (c) 10 ~--~-- . . . . . -'~...... t=2Om.m * 0 ~ t-16m.m 12,5 15 17.5 20 Wi~. feed rat~, ( l ~ J i a ) Fig. I I. Effect o f t h e w i r e feed rate on: (a) the c r a c k length: Ib) the d i l u t i o n ratio: a n d (c) t h e C a n d M n c o n t e n t s in t h e weld metal. 48 M.E. Khallaf et al./ Journal of Materials Process#;g Technology 68 (1997) 43-49 Creek leog~ (re,m) • R.R-I 1.3 + R.H-8.6 40 R.R=6.8 50 30 R.R=AI casted Fig. 12. Effect of the wire feed rate on the weld bead shape: (a) I0 kg/h; (b) 15 kg/h; and (c) 20 kg/h. 3.4. Effect of the plate thickness The effect of plate thickness on crack susceptibility was tested using different welding currents, keeping the reduction ratio constant at 6.8. The results of crack length versus the welding current are plotted in Fig. 13 for three different thicknesses: 12, 16 and 20 mm. The figure indicates that a greater crack length is obtained with decreasing plate thickness: the thinner the plate, the more sensitive the weld bead is to cracking• The plate thickness represents, then, a resisting distance to cracking. Previous studies about the relation between the weld metal cooling rate and the plate thickness by Kihara et al. [14] indicate that, for the same heat input, increasing the plate thickness leads to an increase in the cooling r~,'- of the weld metal. Thin plate is associated with a lower weld metal cooling rate during solidification. The strain theory of solidification cracking [12,13] indicates that the cracking susceptibility increases with increase in the solidification time, which is in agreement with the shrinkage-brittleness theory [13]. Therefore, the hot cracking is more probable in thin plates, as found in the present work. .... 475 ~ - ~ . ~ " ~~ ..rS ' ' ~ / I / ~ / - w't'r~lOklg TM ~ 450 tion agrees with the analysis reported by Blodgett [11], who reported that the weld bead width/depth ratio can significantly affect solidification cracking, deep and narrow beads being more susceptible to weld centerline cracking. ~ - ~ - 500 Curreut (Amp) 525 550 Fig. 14. Effect o f the welding current on the crack length for different reduction ratios. 3.5. Effect of the plate rolling reduction ratio It has been reported in the literature that crack susceptibility in welding is greatly influenced by the mechanical properties of the plate [2]. As these properties are determined by the amount of reduction in the plate thickness during its forming process, studying the direct effect of the reduction ratio on cracking susceptibility, will therefore be of l:ractical significance. For such purposes, a special group of specimens was prepared with the same thickness (12 mm) but where the reduction ratio was different: zero (as-cast), 4.25, 6.8, 8.5 and 11.3. The results obtained applying different welding currents are presented in Fig. 14, from which figure ;,t is noted that, as the reduction ratio increases, the crack length increases. This is explained by the variation in ultimate tensile strength, as discussed previously by Garland [2]. in his work, Garland investigated the influence of the ultimate tensile strength on the crack length during the submerged arc welding process, using a heat-treatment process (quenching and tempering) to raise the ultimate tensile strength of mild steel. His results revealed that the greater the tensile strength of the base metal is, the more likely solidification cracking is to occur. 4. Conclusions SO Chick ksgtb (me) 40 R.R-6.$ S-24mth w.f.r-lO kllh 2O 10 0 450 475 500 Curv~ut (Amp) $25 SSO Fig. 13. Effectof the weldingcurrent on the crack length for different thicknesseshaving the same reduction ratio. Crack susceptibility in the welding of medium carbon steel plates by submerged arc welding is found to be affected by welding conditions, plate thickness and plate fabrication conditions, as follows. 1. The crack susceptibility decreases with decreasing welding current and with increasing welding travel speed or wire feed rate, the behaviour being related to variations in the heat input, the dilution ratio, the penetration and the carbon content. 2. Thin plates are more susceptible to hot cracking than are thicker plates• M.E. Khallaj' et al. ,r'JOlll'~l~il O~ Materiuls Prm'es.~ing Technology 68 (i997) 4 3 - 4 9 3. Plates that have been fabricated using greater forming reduction ratios are more susceptible to hot cracking due to their greater ultimate ~,ensile strength. Acknowledgements The authors wish to thank Dr. Eng. Ali Helmi, chairman of the Egyptian Iron and Steel works, Helwan, for his valuable support throughout this work. References [I] N. Bailey and S. Jones. WHdi..zg R,'~. Sup., {Aug. 1978) 217- 231. [2] B. Hems Worth, T. Boniszew.ki and N. Eaton, Me,. Con.strm't. Br. Weld. J., (Feb. 19691 5 17. 49 [3] R.G. Baker and R P. Newman, ,1.let. Con~tru,'t. Br. WehZ. J.. (Feb. 1969) I--4. [4] J.G. Garland, WeM. Re.s. lilt.. (Jan 19791 59 82. [5] J.C.M. Farrar and J.G. Garlal~d, DerehJpmenl m "/'ran.~ar¢,.~traint Test ./or asse,~sing Solidi[i~ation Crack Su.sc~7~tthility ql It'eh/ Metals. Weldinfg hist. Members Report M 7771. 1973. [6] N. Bailey. Weld. Re,,:. h~t.. 8 (March 1987) 215-238. [7] H. Kihara. H. Suzuki and H. Tamura. Re.war,,'he~ ,m IVehkd,A, High Strength Steels, BOth Anniversary Series, Vol. I, Soc. of Naval Architects of Japan. Tokyo, 1975. [8] A.Yu. Sterenbogen and P.F. Petropv, Aurora. Weld., 32 {July 1979) 34-37. [9] M.L.E. Davis and N. Bailey, The Wehling tn.~t. C'ot~i Wchl Pool Chemistry and Metallurgy, 1979. [101 J.G. Garland and N. Bailey, WeM. Res. int., 9 (Jan. 1979) 85- ! 05. [I 1] O.W. Blodgett, Weld. hmor. Q., 2 {3) (1985) 4. [12] A. Arne, Weld. Res.. {Sept. 1952) 58 66. [131 J.C. Borland, Br. Wehl. J., (Aug.. 1960) 508 512. [14] H. Kihara, H. Suzuki and H. Tamura, 60th Amlirersary Seric.~, Vol. I, Society of Naval Architects of Japan, Tokyo. 1957.