DCSP TIG WELDING OF AA2219 ALUMINUM ALLOY EFFECT OF MISMATCH & WELDING SPEED SGK.Manikandan, S.K.

.Govindarajan, Mohammed Muslim LPSC/ISRO, Mahendragiri. 1.0. INTRODUCTION Considering the high strength at low temperature and good weldability AA2219 aluminum alloy was selected as the material of construction for the cryogenic tanks. Welding technology for aerospace application is available for commercial grade aluminum alloys. Hence there is a need to develop the technology for the welding of AA2219 aluminum alloy. Weldability of aluminum alloy is defined as the resistance of an alloy or alloy/filler metal combination to hot cracking during welding. When the weldability of an aluminum alloy system is monitored in terms of total crack length (TCL) by varying the concentration of major solute element (Copper), a peak hot tearing susceptibility is usually observed. Since the liquid film strength varies inversely with film thickness, an alloy with low copper content and a corresponding low volume eutectic will have a high resistance to tearing. At high solute contents (more than 5.8% Cu) there is always sufficient eutectic liquid available for back filling. The alloy containing more than the maximum soluable Cu level (5.68 wt %), the last liquid freezes at constant temperature very similar to a pure metal and hence acts as a reservoir of molten metal to feed the cracks. It has been suggested that the critical solidification range depends on the amounts of impurities present. The lesser the impurities, lesser the solidification range. Tensile components are required to form the crack, since liquid can not hold the tensile stresses. The weldment will have the sufficient strength to withstand the shrinkage stresses after complete solidification.Only the phase region is more prone for cracking. Hence reducing this region certainly reduces the susceptibility to cracking. The temperature variation in the job during welding induces remarkable change in the precipitation sequence. The thermal energy input during welding and the associated temperature gradients influence the Heat Affected Zone (HAZ) grain growth, eutectic formation and the precipitate of CuAl2. During welding the sequence of precipitation along the parent metal depends on time-temperature profile and parent metal structure. The growth of weld grains and their size depend on the original parent metal grain size. When a heat-treated aluminum alloy is welded, it losses its strength due to unfavorable microstructure changes. The strength can be recovered by solution heat treatment. But elongation can not be recovered fully. In order to achieve the base metal properties in weldment, it is necessary to limit thermal energy. In launch vehicles various components are integrated inside the tank. After completion of all the activities, the man-hole entry in the tank is closed by welding. This paper mainly discusses about the process development for the above and various problems encountered during actual tank welding especially the effect of mismatch in the weld joint. Also this technical paper addresses the improvement of strength by implementing new techniques in the existing process and implementation of the same during production. Details discussed in this technical paper are generated from the various studies carried out in lab level and production level. Also this paper deals with the development of Direct Current Straight Polarity TIG welding process

for the AA2219 aluminum alloy welding and advantages of this process. In this case study, lower thickness (below 5mm) weldments are fabricated and strength values are obtained. Microstructures are studied for weldments made by AC Orbital TIG, DCSP Orbital TIG and DCSP manual TIG processes. 2.0. LITERATURE SURVEY: The load carrying capacity of the fusion welded AA2219 alloy depends on the weld geometry, resistance to plastic flow and resistance to fracture. The weld process produces more dispersed weld structure which enhances the strength of the weld joint. The ductility of AA2219 welds is limited by a shallow, brittle, copper rich cellular (CRC) structure. This structure has unusually high copper content (30-35 wt %) and large volumes of continuous eutectic constituents. This structure is located at the top bead and root corners and extends from the surface. By removing the CRC structure, ultimate strength and ductility values can be increased. In general weld properties are reported in terms of weld efficiency wit h respect to the parent material properties. AA2219 alloy is available in different temper condition. Irrespective of the temper condition, when welding was carried out the properties of the weld are same since the weld is having cast structure and it fails within the weldment. This cast structure is the combination of filler and parent metal chemical composition. Hence the weld properties will depend on the conditions under which weld got solidified. So to get better mechanical and metallurgical properties in the weldment, it is essential to reduce the thermal energy supplied at that time. This parameter is known to be TIME AT TEMPERATURE. Factors affecting the weld strength are as follows, 1. Weld thermal effects - Time/Temperature effect - Cooling rate 2. Defects in the weld 3. Residual stresses 2.1. WELD THERMAL EFFECTS Detailed studies were carried out at the Marshall Space Flight Centre in the weld thermal effects. The results show that the heat input during welding is having an effect on the mechanical and metallurgical properties of the weldment. Higher weld strengths are obtained by using lower heat input. This held true regardless of the material thickness and welding process used. If heat input is reduced, this results higher strength metallurgical structure in HAZ. This phenomenon may be due to a reduced thermal effect and favorable geometry of narrow weld. A joint with a HAZ that has lower strength than the base metal, but with a very narrow weld-metal area, can still have nearly the same UTS as the base metal. This is due to the restriction of plastic deformation in weld metal and HAZ by surrounding base metal. As the heat input increases, there is an increase in grain size and in theta or copper aluminide agglomeration at the grain boundaries.

2.2. TIME-TEMPERATURE EFFECT Time-Temperature effects in welding of AA2219 were analyzed by F.L.Jackson and details are debated in AWS Journal, Vol-45, No-8, Aug1966.This concept is used for the study of the effects of heat input on the strength of aluminum welds. This study has considered the following, (1) Maximum temperature: Peak temperature which the material being joined experiences during the welding cycle (2) Time at temperature: Time that the material being joined is above the temperature that adversely affects strength properties.Strength properties of AA2219 alloy are found to be adversely affected above 4500 F. The strength characteristics, YS and UTS varied with maximum temperature and inversely with time at temperature. Elongation varied with both maximum temperature and time at temperature. 2.3. TIME-TEMPERATURE CONTROL Strength of welded joint in aluminum could be increased by shortening the time at temperature. This can be achieved by the absorption of heat by forced cooling of the base metal by impingement of cryogenic liquids such as LN2, etc. A detailed study was conducted in Harvey Aluminum. Chilling reduced the extent of HAZ and reduced the grain size of the cast structure. Yield strengths are substantially increased by chilling from the front (torch) side. The greatest increase in average YS for welds in AA2219-T87 plate was 8.8%.Porosity could be reduced by cryogenic cooling (90% of chilled welds are free of porosity).Distortion is also minimized in front side cryogenic cooling. This study is initiated in our lab recently. 3.0. EXPERIMENT DETAILS AC TIG welding is most widely used to join aluminum alloys. The success of AC TIG process to weld aluminum lies in the effective removal of oxide layer which occurs in EP cycle of the AC waveform using Argon of Argon rich gas mixture (up to 25% Helium) as shielding gas. TIG welding in DCSP mode with Argon as shielding gas could not be used successfully to weld aluminum alloys as the aluminum oxide layer could not be removed effectively in DCSP mode. The flow of electrons from the electrode to the plate and the usage of helium as shielding gas together ensure a greater heat density in the arc which causes rapid melting of base metal and an excellent penetration is obtained. Moreover the weld bead is narrower than AC TIG process for the same welding current. POLYCAR system available is meant for cir-seam welding and hence specimen plates were fabricated as segments of circle. Bead-on trials were carried out. The near parameters were arrived and trials have been done in the specimen level. These weldments are qualified as per ASME Sec IX. WPS and PQR were generated. With this confidence, actual tanks were welded and weldment was qualified to 100% radiography as per standard generated by us. Details of the welding trials carried out are discussed here.

3.1. SELECTION OF TIG PROCESS VARIANTS 1. 2. 3. 4. 5. DCSP TIG Process in Orbital mode with pulsing DCSP TIG Process in Orbital mode with rapid pulsing DCSP TIG Process in Manual mode DCSP TIG Process in Manual mode with pulsing AC TIG Process in Orbital mode

3.2. SELECTION OF WELD PARAMATERS The weld parameters selection was based on the fact that the decrease in heat input increases the weldments properties and distortion. Following primary parameters are considered, (i) Weld speed (ii) Welding current (iii) Arc voltage (iv) Wire feed rate (v) Weld joint fit-up (Mismatch & Root Gap) In addition to the above, certain secondary parameters are also considered for successful welding and the parameters are, (i) Pulse duration (ii) Wire entry angle (iii) Orthogonality of the electrode with respect to the weld joint (iv) Electrode stick-out (v) Arc gap (vi) Wire to electrode distance To determine the minimum current required to weld for a given welding speed and process, prior to coupon welding, Bead-on plate trials were carried out. 3.3. WELDING SYSTEM POLYSOUDE make AC/DC TIG welding machine with POLYCAR welding system was used for coupon welding and actual tank welding. Capacity: 375A 3.4. JOINT DETAILS Joint type : Square Butt Material : AA2219 T87 Thickness : 3.2mm and 4.2mm 3.5. OPERATIONAL PARAMTERS: 3.5.1. TIG TORCH 250A capacity water cooled TIG torch was used for these welding trials. The TIG torch was held with a slight forward inclination of about 3.5 from the vertical axis. The torch was also fitted with a gas lens for smooth laminar flow of shielding gas. In addition, for maintaining the orthogonality with the spherical dome of the propellant tank, corresponding angle was calculated and set an inclination of 11 in the radial direction of the torch 3.5.2. ELECTRODE WOLFRAM 2% thoriated tungsten electrode of diameter 3.2mm with truncated tip of diameter 0.8mm 3.5.3. CERAMIC NOZZLE 5/8

3.5.4. SHIELDING GAS Pure helium with flow rate range of 22+4lpm for DCSP TIG process High pure Argon (Grade I) with flow rate of 16+2 lpm for AC TIG process 3.5.5. WIRE FEEDING AND FILLER WIRE In-built wire feeding unit was used. ER 2319 filler wire as per AWS 5.10 of diameter 1.2mm was used. 4.0. RESULTS AND DISCUSSIONS 4.1. EDGE PREPARATION Wall thickness up to 8mm can be welded in one layer using square butt joints. In case of special requirements regarding throat thickness of the weld, a second layer may be required to correct the visual appearance of the weld. Normally a U joint is preferred for thicker walls. The quality of normal saw cut is insufficient for fully automated orbital TIG welding of aluminum components. The joint face is generally prepared to the welding with square butt joints using oil free tools/cutters. In case of coolant addition, alcohol is preferred, since the usage of alcohol effectively prevents the creation of circular tool marks. Circular tool marks lead to a transitory increase in weld porosity. This is also referred as mechanical porosity. Jigs and fixture are always recommended to align and fix the faces to be welded. Tack welds without addition of consumables will not normally cause any problems regarding weld porosity, though tack welds are always considered as critical area of concern in aluminum welds. Test specimen plates were machined to square edge with a chamfer of 1mmX45 in the under bead side. This practice is followed in some of the European countries stating that the slight chamfering in the bottom side edges helps a good under bead formation. Welding trials have been carried out even without bottom chamfering. Bead shape was remaining same. But the probability of lack of fusion exists. Each propellant tank has a minor difference in joint fit up. Sometimes there is an inherent root gap before welding. In such cases, 10% of the joint thickness is accepted for mismatch and root gap. Hence fit up was achieved to meet these requirements. In certain cases, the root gap and mismatch can not be minimized. This technical paper dealt a situation that was having deviations in mismatch and root gap in the order of 50% of the joint thickness. Hence process toolings were modified by conducting various trials on specimen plates. With this confidence level, five numbers of propellant tanks were welded successfully. 4.2. CLEANING Mechanical cleaning was employed for the weld joint interfaces. Scrapping was done on the weld joint interface up to 25mm from the edges both on top and bottom faces. Appearance of clean, shining surface ensured the weld joint interface is free from oxide layers. Before mechanical scrapping, the interfaces were thoroughly cleaned with industrial grade acetone to remove any dirt, paint marking, oil, etc., followed with pre-heating for the period of 5 minutes using hot air blower.

4.3. ARC LENGTH It is essential to maintain a very short arc length for obtaining best results in DCSP TIG process which necessitates the use of orbital or automated welding setup. In orbital welding, the arc length is adjusted such that the tip of the electrode remains at the surface level or even immersed below the top surface of the plate. In practical conditions, with mismatch and root gap, the arc length in the range of 1 to 1.5 mm is maintained. But for the fit up conditions exceeding the limits, the arc voltage in the range of 13 to 16.5v was tried. Based on the experiment results, bead profile/geometry is found acceptable with arc voltage range of 13.5 to 14 v which corresponds to the arc length of 1 to 1.5 mm. The maintenance of very short arc length enhances cleaning of oxide as the break down of the film of oxides depends upon the length of arc gap. It is to be noted that the inherent facility of oxide cleaning available in AC TIG process is not available in DCSP mode. Hence it is very much essential that the joint areas are thoroughly cleaned before welding. Ivanova.O.N and Kovalev.A.A reported the advantage of reduced porosity in DCSP TIG process with an immersed electrode in a narrow weld pool which is explained by the rapid circular motion of liquid metal in the plane of the weld edge resulting in degassing. 4.4. WIRE FEED SETTING AND WIRE FEED RATE The wire was fed to the weld pool at an angle of 68 with reference to the torch axis. This value was initially assumed in the range of 65-70. Bead-on trials were carried out with each wire entry angle setting in steps of 1. The wire entry angle was finalized by considering the following factors, (i) Arc gap (ii) AVC sensitivity (iii) Horizontal distance from centre of the electrode to wire entry point (iv) Vertical distance from tip the electrode to wire entry point (v) Length of wire stick out from the wire feeder nozzle The wire feeding nozzle was adjusted such that the tip of the wire touched the plate surface a few mm ahead of the leading edge of the weld pool in-line with the electrode or slightly to the side of the electrode. The size of the weld pool in DCSP TIG welding with Helium shielding is quite small as compared to the conventional AC TIG process. Hence proper positioning of wire tip is more important. There remains a probability of wire touching the electrode. Hence location of the wire tip with offset is always advantageous. The bead profile is having no major difference due to offset. Wire feed rate was initially calculated using the following empirical formulae to get the nearest value so that the number of welding trials can be minimized, G = 0.25 (F/S) (D) 2/ G- Weld bead cross section area, F- Wire feed rate (mm/min), D Diameter of Filler wire (mm) Various wire feed rate values were set and welding trials were conducted. Since wire feed rate is an important operation parameter, 2000 to 3500 mm/min range was considered. Welding trials were conducted for each setting of wire feed rate in steps of 50mm/min and at the rate 2500mm/min the bead geometry, weld strength and other requirements were also achieved. 4.5. WELD BEAD APPEARANCE

The weld bead after DCSP TIG welding does not look bright and clean as in AC TIG process. The bead looks dull grey due to the presence of a light oxide film on the weld bead surface. This film can be easily removed by wire brushing and the presence of this film does not indicate any weld defect in radiography. 4.6. BEAD GEOMETRY The size of the weld pool observed in DCSP TIG is smaller than AC TIG process. The bead width in DCSP for the thickness of 4.2mm is found to be in the range of 10 to 11 mm. Because of the smaller size of weld bead, it is more essential to maintain the precise location of the filler wire tip ahead of the weld puddle to avoid the touching of the wire tip on the electrode. 4.7. EFFECT OF WELDING SPEED The welding speed is playing major role in welding and controlling heat input compared to any other parameter. The weld parameter selection was based on the fact that the lesser the heat input, better the weldment properties. Out of the various parameters, the current and welding speed has the pre-dominant effect on heat input. Hence the welding speed range was fixed and the corresponding minimum current requirement was finalized by carrying out bead-on trials. The welding speeds ranging from 200 to 340mm/min with an increment of 20mm/min were selected and welding trials were carried out with minimum current obtained from bead-on trials. In general, it is not advisable to weld with higher welding speeds, as it requires high current. Welding at higher current leads to lot of practical problems such as blockage of filler nozzle due to thermal expansion and tungsten spitting into the weld. Also welding at higher speed needs higher filler wire feed rate. Any change in filler wire feed rate causes arc instability and seldom electrode spitting. As discussed in chapter 4.1, the wire feed rate also affects the torch setting and thus the change in weld parameters. Higher the welding speed, lesser the heat input (Refer Fig.-4) and results in faster cooling of the weld metal. Faster the cooling rate, more uniform microstructure and more solid solution can be expected. And thus the mechanical properties of the weldment are enhanced. Based on the experiments carried out, it can be understand that UTS is practically constant with increase in welding speed. Moreover in heat treatable aluminum alloy weldment, UTS is not much affected by heat input. Whereas yield strength is highly sensitive to heat input. This is clearly seen in the graph (Fig-2) Welding speed has effect on the presence of porosities for a given concentration of hydrogen. Normally porosities will nucleate when welding speed is intermittent. At low weld speeds the solidification is slow and there will be sufficient time for hydrogen to escape. By the coalescence of small pores, larger pores are formed and reach a critical stage where it crosses the buoyancy forces and escapes from the weld bead. Presence of porosity will have the effect on the weld strength and thus reduces the weld efficiency. It is understood from the experiments that the lower band of the welding speed provides better weld efficiency as indicated in Fig-5 With higher welding speed, there will not be sufficient time for hydrogen to nucleate into pores and thereby fewer pores will be formed. Based on the experiment and trials conducted, it can be inferred that AA2219 alloy is insensitive to porosity formation within the welding speed range used during our trials.

Considering the insensitive nature of AA2219, it is more suitable to select lower band of welding speed (200 220 mm/min) by taking into account of the restriction on the weld overlap region of a cirseam weld and increase in root gap. 4.8. EFFECT OF MISMATCH Design engineers certainly would prefer that mismatch is not present in the weld joint design. But in fabrication, mismatch can never completely eliminate. In welding of sheet metal, angular distortion and mismatch can be held to virtually zero conditions. While welding, mismatch is bound to occur. Problems of mismatch are not resolved even though the parts are clamped during welding. It was observed during welding, the given amount of mismatch in the pre-weld setup is increased. Mechanical difficulties and stresses in the fixing point (fixture) can contribute to this growth. But an additional cause of mismatch is the non-uniform heat input. The side of the weld that receives more heat tends to grow longitudinally. Correspondingly, the side of weld which losses heat at a lower rate increases in length. There is general practice to use rigid back-up bars for heat balance. But our study is to find out an alternative to the above solution, since back-up bar can not be used in our case. Hence we attacked the time-temperature relationships. Because mismatch at the time of welding will disturb the equality of heat distribution on each side of weld. Arc length will be modified and mechanical strengths are affected. Required heat input values were calculated and time at temperature was reduced. And hence there is an enhancement/improvement of the strength. This phenomenon can be observed in Fig-1. By arriving at optimized weld parameters with lesser time at temperature, we could weld the propellant tanks even with a mismatch of 50% of the joint thickness. And these are successfully performing in our developmental tests with high pressure and vibration environments. 5.0. CONCLUSION It can be concluded from the experiments conducted on AA2219 alloy weldments that, - DCSP TIG provides lesser heat input than conventional AC TIG process - Lower welding speeds (in the spectrum) are preferred for reducing the porosity - UTS is not affected by the change of heat input. However some marginal change in UTS is observed which can be explained by micro defects - Weld joints with larger (50%) mismatch can be welded by selecting the proper heat input considering the TIME at TEMPERATURE and hence reduction in UTS is minimized to 27% - Efficiency of the weld decreases as the mismatch increases - It is preferred to use manual DCSP TIG process for repair welding to minimize the loss of weld strength achieved by orbital DCSP TIG process. These trials were also reported in this technical paper (Please refer Fig-3a & 3b) 6.0. REFRENCES 1. ASME Sec. IX 2. ASM welding hand book 3. ASM hand book for aluminum alloys 4. NASA technical report on welding of AA2219 alloy

EFFECT OF MISMATCH
350 300 250 200 150 100 50 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 MISMATCH (mm) YS(Mpa) UTS(Mpa) Linear (YS(Mpa))
FIG-1

Linear (UTS(Mpa))

EFFECT OF WELDING SPEED

350 300 250 200 150 100 50 0 200

220

240

260

280

300

320

340

Welding Speed (mm/min) AC TIG - Auto UTS (Mpa) DCSP TIG-Auto UTS (Mpa)
FIG.-2

AC TIG - Auto YS(Mpa) DCSP TIG-Auto YS(Mpa)

EFFECT OF POLARITY ON UTS

FIG. -3a

EFFECT OF POLARITY ON YS
250 200 150 100 50 0 1 2 3 4 5 AC TIG-AUTO YS(Mpa) DCSP TIG -MANUAL YS(Mpa)
FIG. 3b

6 7 8 9 10 DCSP TIG-AUTO YS(Mpa)

WELDING SPEED Vs. HEAT INPUT

HEAT INPUT (kJ/min)

220

240

260

280

300

320

340

WELDING SPEED (mm/min) Heat input (kJ/min) Linear (Heat input (kJ/min))
FIG. 4

Heat input (kJ/min) Linear (Heat input (kJ/min))

SPEED Vs. WELD EFFICIENCY

220

240

260

280

300

320

340

Welding Speed (mm/min) weld efficiency w.r.t UTS Weld efficiency w.r.t YS

FIG. - 5 Table 1 Chemical composition of AA2219 plate as per ASTM B209 Elements Percentage Cu 5.86.3 Mn 0.20.4 V 0.050.15 Ti 0.020.1 Zr 0.10 Fe 0.30 max. Si 0.20 max. Zn 0.10 max. Mg 0.02 max. Al Bal.

Table 2 Mechanical properties of parent material AA2219 T87 plate UTS (MPa) 448 - 457 YS (MPa) 376 - 378 %Elongation 11 -12 Table 3 Details of torch setting Sl. No. 1 2 3 4 5 6 7 8 9 10 Variables for torch setting Electrode stick-out Arc gap Ceramic nozzle diameter Diameter of electrode Diameter of filler wire Wire entry angle Torch inclination towards filler wire Torch inclination towards spherical dome orthogonally Aligned distance of electrode tip to filler wire tip Linear distance of electrode tip to filler wire tip Set value 50.5mm 1 1.5 mm 5/8 3.2mm 1.2 mm 65+3 3.5+0.5 11 20.5mm 221mm