w(cid:127),=o~ AD-A213 390 " 390Wnt~.i~ghouse WNSD/HYDRO-89-06 METAL TRANSFER IN GAS METAL ARC WELDING Donald M. McEligot Hydrothermodynamics Research and Technology Westingbhouse Electric Corportion- Naval Systems Division (focmerly Gould/Ocean Systems Division) 62 Johnnycake Hill Middletown, Rhode Island 02840 30 September 1989 Final Report for Period 21 September 1988 to 30 September 1989 Contract Number N00014-88-C-0393 APPROVED FOR PUBLIC RELEASEi "DISTRIBUTION IS UNLIMITEDI Prepared for .. I MATERIALS SCIENCE PROGRAM OFFICE OF NAVAL RESEARCH (Code 1131H) 800 North Quincy Street Arlington, Virginia 22217-50o0 OT 04 1989 , 89 10 4 040 UNCLASSIFIED SECURIfY CLASSIFICATION OF TMIS PA=3E REPORT DOCUMENTATION PAGE , REPORT SECURITY CLASSIFICATION lb RESTRICTIVE MARKINGS Unclassified ,_ _ _ __ _ 2a SECURITY CLASSIFICATION AUTHORITY 3 DISTRIBUTION /AVAILABILITY OF REPORT 2..b /A PPR.O..V ED .. .. . . FOR PUBLIC RELEASE: 2b OECLASSIF;CATION / DOWNGRADING SCHEDULE DISTRIBUTION IS UNLIMTITED. 4 PERFORMING ORGANIZATION REPORT NUMBER(S) S MONITORING ORGANIZATION REPORT NUMBER(S) WNSD/HYDRO-89-06 66 NAME OF PERFORMING ORGANIZATION [6b OFFICE SYMBOL 7a NAME OF MONITORING ORGANIZATION Westinghouse Electric Corp. (If dPo(cid:127)icabl) Naval Systems Division DCASPRO-Westinghouse 6C ADDRESS (City. State, and ZIP Cod) 7b ADDRESS (City. State, nd ZIP Code) Hydrothermodynamics Research and Technology c/o Westinghouse Oceanic Division 62 Johnnycake Hill 18901 Euclid Avenue Middletown, RI 02840 Cleveland Ohio 44117 0668 Ba NAME OF FUNDING/SPONSORING 8b OFFICE SYMBOL 9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMSER ORGANIZATION (if (cid:127)pphcbble) Office of Naval Research Code 1131M N00014-88-C-0393 - Sr ADDRESS (Ci(cid:127)y, Stfe, and ZIP Code) 10 SOURCE OF FUNDING NUMBERS 800 N. Quincy Street PROGRAM PROJECT TASK WORK UNIT Arlington, VA 22217 ELEMENT NO NO ACCESSION NO 1i TIILE (Include Security Clahssicetion) Metal Transfer in Gas Metal Arc Welding 12 PFRSONAL AUTHOR(S) Donald M. McEligot T 131 TYPE OF REPORT 13b TIME COVERED 114 DATE OF REPORT (Year, Monthi, Day) 15PGE COUNT Final Report .FROM 21 Se.o30 Se8 89 Sep 30 124 '6 SUPPLEMENTARY NOTATION ,1 COSAT CODES IS SUBJECT TERMS (Continue on reverse if necessary and identif) by block number) fIELD GROUP ISUBGROUP._j IWelding - GMAW yHeat transfer J Metal transfer1 Dropletz Detachment 'Melt in7Numerical predictions, I1 1aBSTRACT (Continue on reverse if necessary and identify by block number) (cid:127)-' The ultimate goal of the present work is to develop sufficient fundamental under- standing of generic GMA welding to enable prediction of the weld quality to be made in terms of the material properties and control parameters. Evidence from experiments and simple analyses shows that the melting rate is controlled by the thermofluidmechanic behavior of the solid electrode and the molten electrode tip. Consequently, an ana- lytical and numerical study of the transient, thermalfluidmechanic behavior of the electrode region in GMA welding continues. The present report discusses two idealized regions separately: (1) ateady thermal conduction in a moving solid electrode and (2) transient droplet formation. The method of superposition and classical techniques were used to describe the temperature discribution in an electrode heated ,oan the side by electron condensation. A transient numerical analysis for the liquid dropl'e or streaming jet is being developed using LpGrangian governing equations...,.. ) 20 DISTR4IUTION AVAILABILITY Or ABSTRACT 21 ABSTRACT SECURITY CLASSIFICATION 0l a)uNk.LAsSIFIkO IUNLIMIT ED 0 SAME AS RPT DTIC USE RS I&* NAME OF nESPONSI&LE INDIVIDUAL 22b TELEPHONE (Include Area Code) 22c OFFICE SYMBOL Dr. George R. Yoder (202)696-0282 ONR Code 1131M DO FORM 1473,84 MAR 86 APR edition may bet utd untl ehavttd __SECURIT Y CLASSIF iC *,ON OF THIS PAGE All other edit-ont are obsolete UINCL 6,SIFIED -I METAL TRANSFER IN GAS METAL ARC WELDING I Abstract 3 Gas metal arc welding has become the most common method for arc welding 3. steels, superalnlys, aluminum and reactive metals, yet the behavior of the interacting physical phenomena Wtich determine the weld quality are not yet understood well. Ab)ut forty percent of the production welding in this country is accomp±ished by this process; in naval ship construction worldwide the percentage is higher and more will be necessary for HY-100 and HY-130. The ultimate goal of the present work is to develop sufficient fundamental understanding of generic GMA welding to enable prediction of the weld quality to be made ia terms of tha material properties and control parameters. Evidence from experiments and simple analyses shows that the melting rate is controlled by Lhe therr,ofluidmechanic behavior of the solid electrode and the molten electrode Uip. Consequently, an analytical and numerical study of the Stransient, thermalfluidmechanic behavior of the detachment phase for metal transfer in GMA welding has been initiated. Current. studies concentrate on two idealized regions to ease physical insight into the overall problem: (1) steady thermal conduction in the moving electrode and (2) transient droplet formation. Emphasis is on applications for steel electrodes. The previous report, for the first half year of the study, reviewed the related background and then concentrated on approximate and time scales analyses preparatory to the development of a transient, axisymmetric numerical ' analysis to describe droplet detachment in globular and spray modes. These analyses did not reveal any significant simplifications which would be appropriate for the governing partial differential equations. Highlights are included in Appendix A of the present report. To account for heating of the side surfaces by electron condensation in argon shielding of steel electrodes, a closed form analysis of thermal conduc- tion in moving electrodes was derived. The method of superposition was applied with a basic solution for uniform internal energy generation (resistive heat- ing) with a step change in surface heat flux (electron condensation). The resulting superposition integral allows treating any reasonable axial surface heating distribution chosen to represent the phenomenon. IScales to non-dimensionalize the droplet variables were defined in terms of the liquid steel properties; they are approximately 1/2 inch, 400 in/min and 30 ms. If the non-dimensional product rwVin is much less than one, droplets are expected to form; if greater, then steady jets should Por occur. This corresponds to globular-spray and streaming transfer, respectively. A one-dimensional, transient numerical analysis, using a LaGrangian representation, is being developed to describe this range of metal -I. transfer modes.'. . . o , B; 1ý 5t TABLE OF CONTENTS Title Page Abstract ..................... ............................. ii Nomenclature ................... ........................... .iv 1. Introduction ................... ......................... 1 1.1 Related studies .............. ...................... 2 1.2 Goal and objective ......... .................... 5 1.3 Tasks .............................. 6 1.4 Prior studies at Westinghouse Naval Systems Division/ . . 7 Newport, RI 1.5 Conspectus ........ ..................... ......... 10 2. Current approach ............... ....................... 11 3. Heat transfer in solid electrode ......................... 12 4. Fluid mechanics of droplet formation and detachment ........ .15 4.1 background ................. ........................ 15 4.2 General approach ............. ..................... .16 4.3 Governing equations (Prof. J.F. Foss) .... ........... .17 4.4 Scales and non-dimensional parameters .... ........... .19 4.5 Non-dimensional problem statement ..... ............. .22 4.6 Method of solution ............. .................... 23 4.7 Initial results ............. ..................... ..30 4.8 Next steps ................. ........................ 31 4.9 Two-dimensional computational modelling of hydrodynamic . .32 droplet processes (Prof. B.E. Launder and Dr. M.A. Leschziner, UMIST) 5. Concluding remarks ............... ...................... 37 Acknowledgements ................. ......................... .41 Table 4.1 C, D and S terms for individual equations .......... .42 Figures ...................... ............................. 43 Appendix A. Analyses of electrode heat transfer in ......... .. A-i Gas Metal Arc Welding by, Y.-S. Kim, T.W. Eagar and D.M. McEligot Appendix B. Classical eigenvalue solution for thermal ... ..... B-i conduction in idealized moving electrode, by J.S. Uhlman, Jr. and D.M. McEligot References cited . . . . . . . . . . . . . . . . . . . . . . . . R-1 L Iii NOMENCLATURE S Acs Cross-sectional area As Surface area cC p Specific heat at constant pressure S D,d Diameter e Electron charge 9s Acceleration of gravity gc Unit conversion factor U H Specific enthalpy h Convective heat transfer coefficient, qs"/(Tw - Tf) II Specific internal thermal energy i Electrical current J Electrical current density, i/A.s I k Thermal conductivity L Characteristic length; also length of electrode from contact tip to arc-molten metal interface Th Mass flow rate; melting rate I P Period q Energy rate (power) or heat transfer rate; qG. total resistive heating; qEC, energy absorbed on side surfaces due to electron condensation; qL-S. required from liquid drop to liquid-solid interface q Electron charge 3 qG Volumetric energy generation rate (resistive heating per unit volume) Sq8 " Surface heat flux r Radial distance; ro, outside radius T Absolute temperature; Tra reference or room temperature; To, amplitude of temperature fluctuations *iv NOMENCLATURE - Continued t Time; relative temperature (e.g.,*C) u Velocity component in axial direction V Voltage; Vc, apparent condensation voltage; Va, anode voltage drop V Uniform or bulk velocity; Vw, electrode (wire) feed velocity v Velocity component in radial direction x Axial distance z Axial distance, measured from molten tip, L-x Non-dimensional parameters Bd, Bo, Bond numbers for thermocapillary phenomena (Lai, Ostrach and Kamotani, BOd 1985], see text Sec. 2. - Appendix A Bi Biot number, hVol/(Asek lid) so Pe Peclet number, PrRe z Vd/a Pr Prandtl number, cpu/k Re Reynolds number, 4vi/(vDv) Greek letters Fraction of total electron condensation absorbed on electrode side surface; thermal diffusivity, k/pCp 8 Volumetric coefficient of thermal expansion o Time scale; (cid:127)T, thermal; Ov, viscous W Viscosity V Kinematic viscosity, vi/p p Density; pt. liquid density Pe Electrical resistivity V NOMENCLATURE - Continued Greek letters - continued cr Surface tension; Gaussian distribution parameter Work function of material (volts) Subscripts i Liquid-solid interface in Evaluated at inlet to region considered, e.g., at x = 0 s Evaluated at surface conditions w Wire (electrode) Note: Some additional symbols used only locally in text are defined where they are used. vi METAL TRANSFER IN GAS METAL ARC WELDING 1. INTRODUCTION Gas Metal Arc (GMA) welding is the major method for welding steels, super-alloys and aluminum; about forty percent of the production welding in this country is accomplished by this process. In this process the thermal phenomena and melting of the solid electrode are coupled to the plasma arc and weld pool. Thus, the thermalfluid behavior of the electrode and detaching drops can have significant effects on the consequent weld quality and production rate. In naval ship construction a greater percentage of the welding is by GMA, an6 more will be necessary for HY-IO0 and HV-130. Yet the behavior ot the interacting physical phenomena, which determine the weld quality in this complicated process, are not yet understood well. GMA welding is impor. i' i- Lhe construction of Naval ships and submersibles. For example, the operating depth of submersibles can be increased by a factor of two by using titanium alloys [Masubuchi and Terai, 19761 if reliable welds can be insured. The gas tungsten arc (GTA) process has proved to be reliable but slow. In order to be used in a production shipyard environment, automatic end manual all-position GMA processes need to be developed. While a number of qualitative hypotheses concerning metal transf or have been suggested and in some instances accepted, quantitative proof of their DMH/ld/O59m -I- validity is still lacking [Eagar, 1989]. A purpose of the present study is to provide quantitative analyses, concentrating on the thermal behavior of the solid electrode and molten drop to aid fundamental understanding of the process. 1.1 Related studies For a general review of recent work on metal transfer, the reader is referred to Lancaster's chapter in the text by Study Group 212 of the Inter- national Institute of Welding tLancaster, 1984]. A brief literature review is also included in our last report (McEligot and Uhlman, 1968]. The classical pioneering study of metal transfer by Lesnewich 11958a, b] has been recently summarized by him in a letter (Lesnewich, 1987]. Cooksey and Miller 119621 described six modes and Needham and Carter 119651 defined the ranges of metal transfer. Weld quality depends on the mode occurring. The axial spray transfer mode is preferred for gas-shielded metal-arc welding to insure maximum ar4 stability and minimum spatter, Analyses and exptriments have been conducted by Greone !150), Halmoy 119801, Woods 11980], Ueguri, Hara and Konira 11985), Allum 1985 and by Waszink and coworkors 11SP2. 1983, 1985, 19866. These studita have prefomi- nantly addressed steady or static conditions, although LAntasster and Allum did consider transient instabilities for possible exr/lanatiovs of the final stage of the droplet detachment process. Tests by AirCo approximately three decades ago showed that various additives on the surface of welding rod0s cold cause sigrificant variations in the deposition rate [Cameron an'i Baeslack, 1956). one likely effect of such additives would be a modification of the surfaca tension of ttU0 liquid metal in DMP/ld/0559m -.2- the droplet. In comparable experiments with GTA weld pools, Heiple and Roper 11982] demonstrated that such alteration of surface tension gradients fhanged the fluid flow patterns and the fusion zone geometry significantly. Thus, it is anticipated that surface-tension-driven flow ("Marangoni convection") [Levich and Krylov, 1969; Ostrach, 1977, 1983] induced by temperature gradients in the molten droplet will also have a significant effect on the droplet formation, flow and detachment and, therefore, on the metal transfer rate. DThC and INEL (Johnson. Carlson and Smartt, 19891 are condutting visual- ization experiments plus studies of signals from electric, audio and acoustic sensors. Droplet detachment events observed on film or video can be correlated with the sensor data. Audio, current and voltage data can be used to discrimin- ate between modes of transfer: globular, spray and streaming. Detachment frequencies tan be cbserved for globular and spray modes, but little informa- tion is observed in the digitized data for streaming transfer. Usually, no abrupt transition (Lesnewich, 1958) is observed (Morris, 1989; Xim, 1989). In recent studies Chen and colleagues 11989] have studied the moclunism of globular metal transfer from covered electrodes. Liu and Siewert 119891 oxamined metal transfer in the short circuiting %ode. Recently Kim 11980 and Appendix A of present report (part)) explored the effects of welding p~rameters on metal transfer phonom"a in GMkW. Droplet sizes were measured by high speed videography and were compared with sizes predicted by a static force balance theory and the pinch instabiliti theory. The coaparison snowed that the droplet size predicted from the static force balance theory can predict the droplet size reasonatly well in the t-ange of globular transfer, but deviated significantly in tho range of spray transfer. The cause of the deviation was found to be due to tipering of the ,,lectteva tip WHI/ld/0559m -3-