SPECIFICATION FOR CARBON STEEL ELECTRODESAND FLUXES FOR SUBMERGED ARC WELDING

SFA-5.17

r{Dm.(Identical with AWS Specification AS.17·89)

1. Scope

This specification prescribes requirements for theclassification ofcarbon steel electrodes (both solid andcomposite) and fluxes for submerged arc welding.

SECflON A - GENERAL REQUIREMENT

2. Classification

2.1 The welding electrodes and fluxes covered by thespecification are classified according to the following:

(a) The mechanical properties of the weld metal ob­tained with a combination of a particular flux and aparticular classification ofelectrode, as specified in Ta­bles 5 and 6;

(b) The condition of heat treatment in which thoseproperties are obtained, as specified in para. 8.4 (andshown in Fig. I)

(c) The chemical composition of the electrode (forsolid electrodes) as specified in Table I, or the weldmetal produced with a particular Oux (for compositeelectrodes) as specified in Table 2.

2.2 Electrodes classified under one classification shallnot be classified under any other classification in thisspecification, except that solid electrodes meeting thechemical composition requirements of both the EL8and ELI2 classification (Table I) may be given bothclassifications. Fluxes may be classified under any num­ber of classifications, for weld metal in either or boththe as-welded and poslweld heat treated conditions, orusing ditl'erent electrode classifications. The classifica­tion system is shown in Fig. 1.

295

2.3 The electrodes and fluxes classified under thisspecification are intended for submerged arc welding,but that is not to prohibit their use with any otherprocess for which they are found suitsble.

3. Acceptance

Acceptance of the materials shall be in accordancewith the provisions of the latest edition of AWS AS.OI,Filler Metal Procurement Guidelines' (see AppendixA3).

4. CertIfication

By affixing the AWS Specification and Oassificationdesignations to the packaging, the manufacturer certi­fies that the product meets the requirements of thisspecification (see Appendix A4).

5. Units of Measure and Rounding.()ff Procedure

5.1 U.S. Customary Units are the standard units ofmeasure in this specification. The Sl Units are given asequivalent values to the U.S. Customary Units. Thestandard sizes and dimensions in the two systems arenot identical, and, for this reason, convenion from astandard size or dimension in one systems will 'notalways coincide with a standard size or dimension inthe other. Suitable conversions encompassing standardsizes ofboth can be made, however, ifappropriate toler­ances are applied in each case.

5.2 For the purpose of determining conformancewith this specification, an observed or calculated valueshall be rounded to the nearest 1000 psi for tensile

tAWS standards may be obtained (rom the American WeldingSociety.''11 N.W. LcJeune1lold, P.O. Bo. 3'1040. Miami, Florida33m.

PART C - SPECIACATIONS FOR WELDING RODS.

ELECTRODES, AND ALLER METALS SFA·5.17

"; TABLE 1CHEMICAL COMPOSmON REQUIREMENTS FOR SOUD ELECTRODES

Electrode UNS wt. ptreentlU W

Classification Numbe,o' C Mn Si S P Cit"

law Manganese Electrodes

ELS KOI008 0.10 0.25/0.1>0 0.07 0.030 0.030 0.35ELSK KOIOll'l 0.10 0.25/0.1>0 0.10/0.25 0.030 0.030 0.35ELl2 KOIOl2 0.04/0.14 0.25/0.60 0.10 0.030 0.030 0.35

Medium Manganese Electrodes

EMI2 KOUI2 0.06/0.15 0.80/1.25 0.10 0.030 0.030 0.35EMI2K KOU13 0.05/0.15 0.80/1.25 0.10/0.35 0.030 0.030 0.35EM13K K01313 0.06/0.16 0.90/1.40 0.35/0.75 0.030 0.030 0.35EMI4K K01314 0.06/0.19 0.90/1.40 0.35/0.75 0.025 0.025 0.35

1TI 0.03/0.17)EMI5K KOl515 0.10/0.20 0.80/1.25 0.10/0.35 0.030 0.030 0.35

Higll Manganese Electrodes

EHUK KUI40 0.07/0.15 1.40/1.85 0.80/1.15 0.030 0.030 0.35EHI2K K01213 0.06/0.15 1.50/2.00 0.25/0.65 0.025 0.025 0.35EHI4 KU585 0.10/0.20 1.70/2.20 0.10 0.030 0.030 0.35

NOTES:(1) The filler metal shall be analyzed for the specific elements for which values are shown in this table. If the presence of ot."er elements is

indicated, in the course of this work, the amount of those elements shall be determined to ensure that their total (excluding iron) doesnot exceed 0.50 percent.

(2) Single values are maximum.(3) SAE/ASTM Unified Numbering SyStem for Metals and Alloys.(4) The copper limit Includes any copper coating that may be applied to the electrode.

TABLE 2CHEMICAL COMPOSmON REQUIREMENTS FOR

COMPOSITE ELECTRODE WELD METAL

7. Retest

Ifany test fails to meet its requirement, that test shallbe repeated twice. The results of both tests shall meetthe requirement. Samples for retest may be taken fromthe original test assembly or from one or two new test

NOTES:(l) The weld metal shall be analyzed for the specific elements for

which values are shown in this table. [f the presence of otherelements Is Indicated, In the course of this work, the amountof those elements shall be determined to ensure that their to­tal (excluding Iron) does not exceed 0.50 percent.

(2) Single values maximum.(3) A low dilution a of the groove weld of Rg. 3 or the frac-

tured tension test specimen of FIg. 5 ...., be substituted forthe weld pad, and shall meet the above requirements. In caseof dispute. the weld pad shall be the referee method.

assemblies. For chemical analysis. retest need be onlyfor those specific elements that failed to meet theirrequirement.

8. Weld Test AssembUes

8.1 No weld test assemblies are required for classili­cation of solid electrodes. One weld test assembly isrequired for classification ofcomposite electrodes. It isthe weld pad in FIg. 2 for chemical analysis of the lowdilution weld metal.

In addition to the above, one weld test assemblyis required for each classification of an electrode-fluxcombination. This is the groove weld in Fig. 3 for me­chanical properties and soundness of the weld metal.Note (3) to Table 2 allows the sample for chemicalanalysis in the case ofa composite electrode to be takenfrom a low dilution area in the groove weld, FIg. 3. orfrom the fractured tension test 8pecimen, FIg. S. there­by avoiding the need to make the weld pad. In case ofdispute, the weld pad shall be the referee method. Whena Certificate ofTest is prepared, it shall indicate wheth-

Mn Sf S PC

0.15 1.80 0.90 0.035 0.035 0.35

ElectrodeClassification

ECI

297

PART C - SPECIACATIONS FOR WELDING RODS,ELECTRODES, AND ALLER METALS

~i-+_,_m_ln_,__'2-+ml-__in~'~__=~__=~6~-_=~:~~::..;Point at which

temperaturei.measured

AII-weJd metaltenlion specimen

~1mi".

(A) Joint Configul'lltlon end Location of Ten Specimen,

SFA-5.17

Width of bocklng

2

Nole: All dimensions except angles are in inches.

~fjir -- -Weld ct. Weld~

(B) Location of Impact TestSpecimens

(c) Location of Tension TestSpecimen

51 EQUIVALENTSin. mrn

1116 1.6114 6.4318 9.51/2 13314 19

1 255 1276 152

12.. _ 305 __

Wekfing Conditions tor SoUd Electrodes •.b

ElectrodeSizec:

65 10 325°F 275 to 325"F(1810 163"C) (13510 163"C)

Amperage" Voltage. Welding Travel Speed

..:i;.:n.'-----.::m:::m:.:-_-'I"'±..:2::5:c) -'I"'±'--"')'---_..:C::urrenl ipm (± 1) mmls (± 0.41

1/16 1.6 350 28 12 5.15164 2.0 400 28 13 5.53132 2.4 450 28 D.C. 14 5.91/8 3.2 500 28 eijher 15 6.35132 4.0 550 28 polarilY 16 6.83116 4.8 600 28 17 7.27/32 5.6 650 28 18 7.61/4 6.4 Not Specified Not Specified Not Speafledh

Baset

Metal

ASTMA36.A285 Grade C,

A515 Grade 70,or

A516 Grade 70

PreheatTemperatureg

InterpassTemperature9

a. The first layer shall be produced in either 1 or 2 passes. All other layers shall be produced in 2 or 3 passes per layer except thelast. which shall be produced in 3 or 4 passes. These welding conditions are intended for machine or automatk: welding.b. Welding conditions for composite electrodes shall be as agreed between purchaser and supplier.c. Classification is based on the properties of weld metal with 5/32 in. (4.0 mm) electrodes or the closest size manufactured. if 5/32in. (4.0 mm) is not manufactured. The conditions given above 'or sizes other than 5/32 in. (4.0 mm) are to be used when .classification is based on those sizes, or when they are required for lot acceptance lesling under AS.01. Fill~, M~/(ll Procu,~mtnl

Guid~lin~s (unless other condilions are specified by the purchaser).d. Lower amperages may be used for the first layer.e. Contact tube·to·work distance is 1/210314 in. (13 to 19 mm) for 1/16 and 5/64 in. (1.6 and 2.0 mm) electrodes; 3/4 10 III.. in. (191032 mm) for 3/32 in. (2.4 mm) eleclrode~: and I 10 JIll in. (25 to 38 mm) for 1/8,5/32,3/16. and 7/32 in. (3.2,4.0.4.8 and 5.6mm) elcclrodes. When an electrode manufaclurer recommends a comact tube·to·work dislance outside the range shown, thoserecommendations shall be followed within 1/4 in. (6 mm).r. In case of dispule, A36 steel and DeEP (electrode positive) shall be used as the referee base metal and current.g. The first bead shall be produced with the assembly at any temperature between 6.5 and 325°F (18 to 136°C). Welding shall continue.bead by bead, until a temperalure within the interpass temperature range has been anained. Thereafter, production of subsequent beadsmay begin only when the assembly is within the interpass temperature range.h. When a lest assembly is required for 1/4 in. (6.4 mm) electrodes, Ihe welding conditions shall be as agreed between purchaser andsupplier.

FIG. 3 GROOVE WELD TEST ASSEMBLY

PART C - SPECIACATIONS RlR WELDING RODS.ELECTRODES. AND ALLER METALS SFA·5.17

metal. Both surfaces of the test assembly, in the areaof the weld, shall be smooth enough to avoid di1Iicultyin interpreting the radiograph.

10.2 The weld shall be radiographed in accordancewith ASTM method E 142 Controlling Quality ofRa­diogrrzphk Testing. The quality level of inspection shallbe 2-2T.

10.3 The soundness of the weld metal meets the re­quirements of this specification if the radiograph doesnot show:

(a) Cracks, incomplete fusion, or incomplete pene­tration;

(b) Slag inclusions longer than y. in. (6 mm) or Y, ofthe thickness of the weld, whichever is greater, orgroups ofslag inclusions in line that have an aggregatelength greater than the thickness ofthe weld in a length12 times the thickness of the weld except when thedistance between the successive inclusions exceeds sixtimes the length of the longest inclusion in the group.

(e) Rounded indications in excess ofthose permittedby the radiographic standards in Fig. 4.

One inch ofthe weld, measured from each end of thetest assembly, shall be disregarded from radiographicevaluation.

11. Tension Test

11.1 One all-weld-metal tension test specimen shallbe machined from the groove weld described in para.8.4 and shown in Fig. 3. The dimensions of the speci.men shall be as shown in Fig. S.

11.2 The specimen shall be tested in the mannerdescribed in the tension test section of the latest editionofANSIIAWS B4.0, Standard Me/hodsfar MechanicalTesting of Welds.

11.3 The results of the tension test shall meet therequirements specified in Table S.

12. Impact Test

12.1 Five Charpy V-notch impact specimens (Fig. 6)shall be machined from the test assembly shown in Fig.3 for those classifications for which impact testing isrequired in Table 6.

12.2 The five specimens shall be tested in accordancewith the impact test section of ANSIIAWS B4.0. Thetest temperature shall be that specified in Table 6 forthe classification under test.

12.3 In evaluating the test results, the lowest and thehighest values obtained shall be disregarded. Two of

J01

the remaining three values shall equal, or exceed, thespecified 20 ft·lb (27 1) energy level. One of the threemay be lower, but not lower than IS ft·lbs (201), andthe average of the three shall not be less than the re­quired 20 ft·lb (27 1) energy level.

SECrION C - MANUFACI'URE,IDENTJFICATION AND PACKAGING

13. Method of Manufacture

The electrodes and flues classified according to thisspecification may be manufactured by any method thatwill produce material that meets the requirements ofthis specification.

14. Electrode Requirements

14.1 Standard Sizes. Standard sizes for electrodes inthe different package forms (coils with support, coilswithout support, and drums) are shown in Table 7.

14.2 Finish and Uniformity

14.2.1 The electrode shall have a smooth finishwhich is free from slivers, depressions, scratches, scale,seams and laps (exclusive of the longitudinal joint incomposite electrodes), and foreign matter that wouldadversely afl'ect the welding characteristics, the opera­tion of the welding equipment, or the properties of theweld metal.

14.2.2 A Each continuous length ofelectrode shallbe from a single heat or lot ofmaterial, and welds, whenpresent, shall have been made so as not to interfere withthe uniform, uninterrupted feeding of the electrode onautomatic and semiautomatic equipment.

14.2.3 Core ingredients in composite electrodesshall be distributed with sufficient uniformity through­out the length of the electrode so as not to adverselyafl'ect the performance ofthe electrode or the propertiesof the weld metal.

14.2.4 A suitable protective coating, such as cop­per, may be applied to any electrode covered in thisspecification.

14.3 Standard Package Forms

14.3.1 Standard package forms are coils with sup­port, coils without support, and drums. Standard pack­age dimensions and weights for each form are given inTable 8. Package forms, sizes and weights other thanthese shall beas agreed between purchaserand supplier.

PART C - SPECIACATIONS FOR WELDING RODS.ELECTRODES. AND ALLER METALS SFA·5.17

I-------c------.IF 11\I G

-- -

~ '-0 /(See Note (2))

Dimenslons 01 Specimen. in.

Test Plate ApproximateThickness 0 G C B F.MIN Area. in.2

0.500 ±0.010 2.000 ± 0.005 2-1/4 3/4 0.375 1/5(3/8)

Dimensions of Specimen. mm

Tes1 Plate ApproximateThickness 0 G C B F,MIN Aree. nun'

25.4 12.7 ± 0.25 SO.8 ± 0.13 57.1 19.1 9.5 129

NOTES:(1) Dimensions G and C shall be as shown, but the ends may be as required to fit the testing

machine holders as long as the load is axial.(2) The diameter of the specimen within the gage length shall be slightly smaller at the center

than at the ends. The difference shall not exceed one percent of the diameter.(3) The finish of the surface within the C dimension shall be no rougher than 63 ... in. (1.6 f.L m).

AG. S DIMENSIONS OF TENSION TEST SPECIMEN

TABLESTENSION TEST REQUIREMENTS

Yield Strengthmin.wAux

ClassificationCU

Tensile Strength

psi MPa psi MPaElongation% min.w

F6XX·EXXXF7XX·EXXX

60 000/80 00070 000/95 000

415/550480/650

4800058000

330400

2222

NOTES:(1) The letter "X" used In various places In the classifications in this table stands for, respectively, the

condition of heat treatment, the toughness of the weld metal, and the classification of the electrode~~n ..

(2) Yield strentth .t 0.2 petUflt offset and elongation In 2 in. (51 mm) gage length.

303

PART C - SPECIACATIONS FOR WELDING RODS.ELECTRODES. AND AUER METALS

TABLE 7STANDARD ELECTRODE SIZES AND TOLERANCEsm

Size(Diameterl Tolerance

Solid lD Composite (EelIn. mm ±In. ±mrn ~In. ±mm

II. (0.0625) 1.6 0.002 0.05 0.005 0.13¥~ (0.078) 2.0 0.002 0.05 0.006 0.15II. (0.094) 2.4 0.002 0.05 0.006 0.15II (0.1251 3.2 0.003 0.08 0.007 0.180/,. (0.156) 4.0 0.004 0.10 0.008 0.20l\. (0.188) 4.8 0.004 0.10 0.008 0.20Y,z (0.219) 5.6 0.004 0.10 0.008 0.20Yo (0.250) 6.4 0.004 0.10 0.008 0.20

NOTE:III Other sizes and tolerances may be supplied as agreed between purtllaser and supplier.

.~

SFA·5.17

aged to ensure against damage during shipment andstorage under normal conditions.

14.7 Marking of Paclalges

14.7.1 The following product information (as aminimum) shall he legibly marked so as to he visiblefrom the outside of each unit package.

(a) AWS specification and classification number(year of issue may he excluded);

(b) Supplier's name and trade designation;(c) In the case of a composite electrode, the trade

designation of the ftux (or lIuxes) with which its weldmetal composition meets the requirements of Table 2;

(d) Size and net weight;

TABLEBSTANDARD DIMENSIONS AND WElGHTSllI

Electrode Siz....

HetOt

Weight ofCoil

InsldeDIameterof Uner

Width ofColI, Max.

OutsideDIameter

of Coi~ Max.

in.

Coils With Support

mm Ib kg in. mm in. mrn in. mm

~.-Y. 1.&-0.4

Coils Without Support

1.6-4>.4

Drums

1.6-4>.4

[25 ll} -C211 65 17Y: 44550 23 12zY. 30Sz3 _ 4¥e 120 17 43060 2765 30

{100 :}150 (4) 5 125 31Y, 800200 91

- - - As agreed between purchaser and supplier - - -

- - - As agreed between purchaser and supplier - - -

NOTES:(u Other dimensions and weights may be supplied as agreed between purchaser and supplier.(2) The range is inclusive.(3) Net weights shall not vary more tIIa!' ± 10 percent.(4) The diameter of tile liner shall be as agreed between purchaser and supplier.(s) Outside diameter of drums Is nominal, not max1mum.

305

PART C - SPECIACAll0NS FOR WELDING RODS.ELECTRODES, AND AUER METALS SFA-S.17

• Wear correct eye, ear and body protection.• Do not touch live electrical parts.• See American National Standard Z49.1. Sofety in

Welding and Cutting, published by the AmericanWelding Society. 550 N.W. LeJeune Road, P.O.Box 351040, Miami, Florida 33135; OSHA Safety

307

and Health Standards, 29 CFR 1910, availablefrom the U.S. Government Printing Office, Wash­ington, DC 20402

DO NOT REMOVE THIS LABEL

AppendixGuide to ANSI/AWS AS.17-89, Specification for CarbonSteel Electrodes and Fluxes for Submerged Arc Welding

(Ibis AppendiJI is not a part of ANSI!AWS AS.17-S9, speci/iCJJtion for Qubon Steel Elet:ttodes and F7wces for SubmergedAn: Welding, but is included for information purposes only.)

At. introduction

The purpose ofthis guide is to correlale the electrodeand flux classification with their intended applicationsso the specification can be used etrectively. Reference toappropriate base metal specifications is made wheneverthat can be .done and when it would be helpful. Suchreferences are intended only as examples rather thancomplete listings of the base metals for which eachelectrode and flux combination is suitable.

Al. Oassificatlon System

Al.t Oasslficatlon of Electrodes. The system foridentifying the electrode classifications in this specifi­cations follows the standard patlem used in other AWSfiller metal specifications. The letter "E" at the begin­ning of each classification designation stands for elec­trode. The remainder of the designation indicates thechemical composition of the electrode, or, in the caseof composite electrodes, of the low dilution weld metalobtained with a particular flux. See Fig. 1.

The letter "L" indicates that the solid electrode iscomparatively low in manganese conlent. The letteruM" indicates a medium manganese content, while theletter "H" indicates a comparatively high manganesecontent. The one or two digits following the manganesedesignator indicale the nominal carbon content of the .electrode. The letter "K", which appears in some desig­nations, indica1eS that the electrode is made from a heatofsilicon-killer steel. Solid electrodes are classified onlyon the basis oftheir chemica! composition, as specifiedin Table I of this specification.

309

A composite electrode is indica1ed by the letler "C"after the "E", and a numerical sullix. The compositionof a composile electrode is meaningless and the user istherefore referred to weld metal composition (Table 2)with a particular flux, rather than to electrode compo­sition.

Al.2 Oasaificatlon of Fluxes. Fluxes are classifiedon the basis of the mechanical properties of the weldmetal they produce with some certain classification ofelectrode, under the specific test conditions called forin Section B of this specification.

As examples of flux classifications, consider the fol-lowing:

F6AO-EH14F7P6-EMI2KF7P4-ECI

The prefix "F' designates a flux. This is followed bya single digit representing the minimum tensile strengthrequired of the weld metal in 10 000 psi increments.

When the letter "A" follows the strength designator,it indicates that the weld metal was tested (and is classi­fied) in the as-welded condition. When the letter "P"follows the strength designator, it indicates that theweld metal was tested (and is classified) after postweldheat treatment called for in the specification. The digitthat follows the A or P will be a number or the letter"Z". This digit refers to the impact strength ofthe weldmetal. Specifically, it designales the Iemperature at(and above) which the weld metal meets, or exceeds,the required 20 ft-lb (27 J) Charpy V-notch impactstrength (except for the letler Z. which indica1eS thatno impact requirement is specified - see Table 6).These mechanical property designations are followed

PART C - SPECIFICATIONS FOR WELDING RODS.ELECTRODES, AND FILLER METALS SFA·5.17

weld metal chemical analysis as a result of a largechange in the arc voltage, and thus, the arc length.

The primary use for neutral fluxes is in multiple passwelding, especially when the base metal exceeds I in.(25 mm) in thickness.

Note the following considerations concerning neu­tral fluxes:

(a) Since neutral fluxes contain little or no deoxidiz­en, they must rely on the electrode to provide deoxida­tion. Single pass welds with insuflicient deoxidation onheavily oxidized base metal may be prone to porosity,centerline cracking, or both.

(b) While neutral fluxes do maintain the chemicalcomposition of the weld metal even when the voltageis change, it is not always true that the chemical compo­sition of the weld metal is the same as the chemicalcomposition of the electrode used. Some neutral fluxesdecompose in the heat of the arc and release oxygen,resulting in a lower carbon value in the weld metal thanthe carbon content of the electrode itself. Some neutralfluxes contain manganese silicate which can decomposein the heat of the arc to add some manganese andsilicon to the weld metal even though no metallic man­ganese or silicon was added to these particular fluxes.These changes in the chemical composition of the weldmetal are fairly consistent even when there are largechanges in voltage.

(c) Even when a neutral flux is used to maintain theweld metal chemical composition through a range ofwelding voltages, weld properties such as strength leveland impact properties can change because of changesin other welding parameters such as depth of fusion,heat input, and number of passes.

A6.1.2 Active Fluxes. Active fluxes are thosewhich contain small amounts of manganese, silicon, orboth. These deoxidizers are added to the flux to provideimproved resistance to porosity and weld crackingcaused by contaminants on or in the base metal.

The primary use for active fluxes is to make singlepass welds, especially on oxidized base metal.

Note the following considerations concerning activefluxes:

(a) Since active fluxes do contain some deoxidizers,the manganese, silicon, or both in the weld metal willvary with changes in arc voltage. An increase in manga­nese or silicon increases the strength of the weld metalin multiple pass welds but may lower the impact prop­erties. For this reason, voltage must be more tightlycontrolled for multiple pass welding with active fluxesthan when using neutral fluxes.

(b) Some fluxes are more active than others. Thismeans they offer more resistance to porosity due tobase metal surface oxides in single pass welds than a

311

flux which is less active, but may pose more problemsin multipass welding.

A6,l.3 Alloy Fluxes. Alloy fluxes are those whichcan be used with a carbon steel electrode to make alloyweld metal. The alloys for the weld metal are added asingredients in the flux.

The primary use for alloy fluxes is to weld low alloysteels and for hardfacing. As such, they are outside ofthe scope of this specification. See the latest edition ofANSVAWS A5.23, Specification for Low Iflloy SteelElectrodes and Fluxes for Submerged Ifn: Welding. fora more complete discussion of alloy fluxes.

A6.1.4 Wall Neutrality Number. The Wall Neu­trality Number is a convenient relative measure of fluxneutrality. The Wall Neutrality Number addressesfluxes and electrodes for welding carbon steel with re­gard to the weld metal manganese and silicon content.It does not address alloy fluxes. For an electrode·f1uxcombination to be considered neutral, it should have aWall Neutrality Number of 40 or less. The lower thewall neutrality number, the more neutral is the flux.

Determination of the Wall Neutrality Number (N)can be done in accordance with the following:

(a) A weld pad of the type required in the specifica­tion is welded with the electrode-flux combination be­ing tested. Welding parameters are the same as thosespecified for the weld test plate for the diamter electrodebeing used.

(b) A second weld pad is welded using the sameparameters, except that the arc voltage is increased by8 volts.

(c) The top surface of each of the weld pads isground or machined smooth to clean metal. Samplessufficient for analysis are removed by machining. Weldmetal is analyzed only from the top (fourth) layer ofthe weld pad. The samples are analyzed separately forsilicon and manganese.

(a) The Wall Neutrality Number depends on thechange in silicon, regardless of whether it increases ordecreases, and on the change in manganese, regardlessof whether it increase or decreases. The Wall NeutralityNumber is the absolute value (ignoring positive or neg­ative signs) and is calculated as follows:

N = 100 (ld%Sil + Id%Mnj)

where:d %Si = is the difl"erence in silicon levels of the

two pads andd%Mn = is the corresponding difl"erence in manga­

nese levels.

PART C - SPECIACATIONS FOR WELDING RODS.ELECTRODES. AND ALLER METALS

"

SFA·S.17

tions of those beads. For this reason, the properties ofa single-pass weld may be somewhat different fromthose ofa multipass weld made with the same electrodeand Dux.

The weld metal properties in this specification aredetermined either in the as-welded condition or after aposlWeld heat treatment (one hour at 115O"F [621'C),or both. Most of the weld metals are suitable for servicein either condition, but the specification cannot coverall ofthe conditions that such weld metals may encoun­ter in fabrication and service. For this reason, the classi­fications in this specification require that the weld met­als be produced and tested under certain specificconditions encountered in practice. Procedures em­ployed in practice may require voltage, amperage, typeof current, and travel speeds that are considerablydifl'erent from those required in this specification. Inaddition, difl'erences encountered in electrode size, elec­trode extension, joint configuration, preheat, interpasstemperatures, and postweld heat treatment can have asignificant dect on the properties ofthejoint. Extendedpostweld heat treatment (conventionally 20 to 30 hoursfor extremely thick sections) may have a major inftu­ence on the strength and toughness of the weld metal.

313

Both can be substantially reduced. The user needs tobe aware of this and of the fact that the mechanicalproperties of carbon steel weld metal produced withother procedures may difl'er from the properties re­quired by Tables 5 and 6 of this specification.

A6.4 Diffusible Hydrogen. Submerged arc weldingis normally a low hydrogen welding process when careis taken to maintain the Dux in a dry condition. Insubmerged arc welding with carbon steel electrodesand Duxes classified in the specification, weld metal orheat-afI'ected zone cracking associated with difl'usiblehydrogen is generally not a problem. Exceptions mayarise when joining high carbon steels or when usingcarbon steel electrodes to weld on low alloy highstrength steels (e.g., for a joint of carbon steel to lowalloy steel).

If an assessment of the diffusible hydrogen contentis to be made. The method of ANSIIAWS A4.3-86,Standard ProceduresforDetermination ofthe DiffUSibleHydrogen Content ofMartensitic. Bolnitic. and FerriticSteel Weld Metal Produced by Arc Welding. is appro­priate.

i•!

Effect of Electrochemical Reactions onSubmerged Arc Weld Metal Compositions

Weld metal composition is controlled by chemical reactionsin four separate areas during welding

BY J. H. KIM, R. H. FROST, D. L. OLSON AND M. BlANDER

(1)

I.;

,· .

·,

1

j

ABSTRACT. The purpose ohms work is toinvestigate the relative influence of elec­trochemical and thermochemical reac­tions on the weld metal chemistry in a d~reet current submerged arc welding pro­cess. Chemical analyses were carried outon the melted electrode tips. the de­tached droplets and the weld metal forboth electrode-positive (reverse) polaritywhere the welding wire is anodic andelectrode-negative (straight) polaritywhere the welding wire is cathodic. Theresults suggest that both thermochemicaland electrochemical reactions are impor­tant in altering the composition of theweld metal. The anodic electrochemicalreactions include the oxidation of iron andaHoy elements. and the discharge andpickup of oxygen anions from the moltenflux. Cathodic electrochemical reactionsinclude the reduction of iron and aDoy el­ements from the flux to the metal phase,and the refining of oxygen. Therm0chem­ical reactions occur and move the overaUcomposition toward the thermochemicalequi6brium.

Introduction

The submerged arc welding processuses a protective flux cover with a con­sumable welding wire. The welding cur­rent is carried largely by the submergedarc and to some extent by conduction inthe molten flux layer. Information on theinteraction between slag and weld metitlis significant for a better understanding ofthe arc welding process. Although s1ag/metal reactions are of importance in manyfusion welding processes and are oftenreferred to in the literature, comparativefy

/. H. KIM. R. H. FRosrondD. L. OI.SONorewilhthe Ct>I1ter for Welding Resedrch. Dep.Jrtmentof Melo/lurgic./ Engineering. CoIorMJo SChoolof Mines. Colden. Colo. M. BJ.ANDEII is wilh~ N.ttioniJl L.Jbor.Jtory. Chemic~ T«It­noIogy Division/MJleri.J!s Science Mld Tech­nology ProgrMn. "'OJ"""', •.

P4p«' ~ed" lhe 6&h AInIM AWSMeeting. heldMJrt:h 11-17. 1987," O,bgo,..

lillie is known about the mechanisms thatcontrol them.

The chemistry in steel making tends tobe dominated by thermochemical reac­tions because they generally involve largesurface· areas, low current densities andalternating current power. But a directcurrent submerged arc wekfong processinvolves small surface areas and high cur­rent densities. Therefore, ft is expectedthat electrochemical reactions, as wei asthermochemical reactions, will exert a sig­nifICant influence over the fonal chemistryof the weld.

The overall composition o.f the weld iscontrolled by the composition of themetal droplets that enter the weld pooland by the amount of dilution of the weldpool by the base plate. These composi­tions are controlled by chemical reactionsat four separate zones: the melted elec·trocle tip. the detached droplet. the hotweld pool immediately below the arc, andthe cooling and solidifying weld pool be­hind the arc. Thermochemical reactionsoccur at all four zones and move thechemistry in the direction of chemicalequilibrium. But electrochemical reactionsoccur only at the melted electrode lip andin the hot weld pool immediately below.he arc. The electrochemical reactionswould result from ionic conduction of aportion of the welding current throughthe molten slag layer. Therefore, onlythermochemical reactions are expectedto occur at the surface of the detached

KEYWORDSSubmerged Arc WeldingWeld Metal CompositionElectrochemical ReactionThermochemical ReactionChemical AnalysesSAWAlloy ElementsFlux CompositionElectrocle Co.mpositionWelding Speed Effect

droplets and solidifying weld pool sincethese parts are no longer carrying current.

The most important chemical consider·ations for submerged arc welding includethe control of oxygen, oxidation losses ofalloy elements, and the pickup of undesir­able elements from the slag. A runber ofinvestigations have been made concern­ing the interaction between the slag andthe metal in submerged arc welding ofsteel. /IIIost of these investigations werebased solely on thermochemical reactions(Refs. 1-9). although lately a few investi­gators have considered the electrochem­ical reactions that occur when direct cur­rent is used in welding. Frost (Ref. 10)considered the different chemical effectsat the anode and cathode in electroslagwekfong. Blander and Olson (Ref. 11)pos­tulated an electrochemical mechanism forthe alteration of weld metal chemistry insubmerged arc welding.

Electrochemical andThermochemical Reactions

Electrochemical reactions could be animporlant factor governing the chemistryof weld metal in direct current wek:fmg.Electrochemical effects can occur at theinterfaces of the slag with the hot weldpool and the melted electrode lip. Therelative numbers of ions and electrons inthe slag govern the importance of theseeffects, which could be major factorscontrolling weld metal chemistry. Whenliquid slag has an interface with the !noItenmetd!. the electrochemical reactions arerelatively simple to understand. This typeof interface is present in submerged arcwelding. The possible anodic reactions in'clude the oxidation of iron and alloy ele­ments, and the discharge and pickup ofoxygen anions from the slag

M (metaO + n()2- (slag) =MO" (slag) + 2ne-

where M is iron or an anDy element at theelectrode tip/slag or the weld pool/slaginterface. Thus. substantial oxidationlosses of oDoy elernents and pickup ofoxygen are expected at the anode. The

T.able l-Compositions of Steel Pl.Jte ~nd

welding Wft (wl-'Xo)

Fig. 1- 5chem.1tic ofexpected vari.1tionsin oxygen and siliconcontents for anelectrode-positivepolarity submergedarc weld.

Experimental Procedure

Direct current electrode·negative(straight) and direct current elect ...ode­positive (reverse) polarity submerged arcwelds were made with 2.38 mm (l/n in.)commercial low-carbon steel weldingwire on ASTM A36 steel plates and .a wa~

ter-cooled pure copper plate. The com­positions of the steel plate and the 'Neld­ing wire are given in Table 1. and thecomposition of the commercial flux usedis shown in Table 2.

Figure 2 is a schematic drawing of across-section of the submerged arc 'Neld­ing process showing the melted electrodetips. the detached droplets and the baseplate for both electrode-positive and elec­trode-negative polarities. The melted elec­trode tips, detached droplets and weldmetal were collected and analyzed. Thewater-cooled copper plate was used as abase plate to collect the melted electrodetips and the detached droplets. The weld­ing process was operated at a constantwelding current of 585 A and a constantpotential of 28.5 V. The constant currenl

ygen pickUp and silicon loss. The thermo­chemical reactions cause more oxygenpickup at the melted electrode tip and thedetached droplet while the dro~let ispassed through the molten flux. A.t thecathodic weld pool. oxygen refining andsilica reduction were expected, and thenweld composition would vary with travelspeed due to the difference in the arnounlof lhermochemical deoxidation reactionAs the welding speed decreases # theamount of thermochemical deoxidalionreaction increases because the heat inputand the solidification time increase_There­fore, weld pool composition wa 5 ex­pected to go to equilibrium compositionas the travel speed decreases. .

/

"-.. ThermochemicalReactions inthe Weld Pool

Silicon (wt. %)

Cathodic 1'feld Pool

Expected Composition Variationstor a DCEP Submerged Arc Weld

ThermochemicalEquilibrium

AnodicElectrode

r::ClJbJJ:>­><o

tionsto occur. A number of investigationshave been made concerning the nature ofchemical reactions at the melted elec·trode tips and the detached droplets(Refs. 5.8.12-18), and several investiga­tors indicated that oxygen was transferredto the metal in these two zones (Refs. 2,3. 13. 14, 18). In the hot weld pool imme­dialely below the arc. the detached drop­lets become "diluted" with molten metalirom the base plate. Although the slaglmetal interfacial contact area available forreaction is much smaller than that of thedetached droplets, the strong turbulanceexisting in the hot weld pool creates aneffective stirring of the liquid metal. More­over, pickup of oxygen at this zone is fa­vored by the increased time available forreaction_ It is believed that the pickup ofoxygen is the result of the oxygen solubil­ity being exceeded in the hot weld pool.The zone of the molten weld pool behindthe electrode starts to cool and solidify asthe electrode moves away from it. Uponcooling of the metal in the weld pool froma high temperature to the solidificationtemperature, a supersaturation with re­spect to silicon and manganese deoxida·tion reactions occurs initially. Therefore,silicon and manganese will react with OiS­

solved oxygen, and then the compositionwill move rapidly towards equilibrium.

Figure 1 shows the schematic of ex­pected variations in oxygen and siliconcontents for an electrode-positive (re­verse polarity) submerged arc weld due toelectrochemical and thermochemical re­actions. The curve shows the oxygen andsilicon contents of steel in equilibrium withsolid silica at some temperature. Beforeexperiments were done, the compositionvariation was expected to follow thestraight lines for an electrode-positive sub­merged arc weld. The anodic reaction atthe melted electrode tip causes some ox-

(5)

(6)

Welding Wire

0.0591.380.0503300730.1100'80770.0330.0120.0150.0020.00<

balance

Steel Plale

0.181.250.050.030.0770.060.00<0.0360.00'00'80.0310.0020.00<

balance

EtemPnls

·:~C

Mn5.MoCrN,AIellT;p5oNFe

possible cathodic reactions indude reduc­tion of metallic cations from the slag and,10 some extent, the refining of nonmetal­lic elements such as oxygen and sulfur

Ml+ (slag) + 2e- = M (metal) (2)

Si<+ (slag) + 4e- = Si (metal) (3)

o (metal) + 2e- = 0'- (slag) (4)

where M and Si represent electrodepes·ited metals at the interface. The large d;s­parity in areas between the electrode tipand the weld pool results in different cur­rent densities. Current density is muchhigher at the electrode tip interface thanat the weld pool interlace. Thus. reactionsat Ihe electrode tip may exert a greaterinfluence on the final weld metal chemis­try than those at the weld pool.

In addition to the electrochemical reac­tions. thermochemical reactions also playan important role in controlling the weldmetal chemistry in submerged arc weld~

ing. The thermochemical reactions occurvery rapidly because of high temperature.These thermochemical reactions includedeoxidation reactions. such as those en­countered in steel making, and reactionsthat lead to a closer approach to equilib­rium between the flux and metal phase.Examples of such reactions would be sili·con pickup from a high silica flux or theoxidation loss of transition elementsthrough a deoxidation reaction.

Si02 (slag) + 2M (metal) = Si (metal)+ 2MO(slag)

FeO (slag) + M (metal) = Fe (metal)+ MO(slag)

where M can be aluminum or calcium inReactions 5 and 6 and manganese in Re­action 6 for the flux system used in this re­search. The droplet forms at the electrodetip and then travels through the moltenflux and plasma. The entire process occursinafew miliiseconds(Refs.12,13).andthelemperature of the droplet is very highDue to the high temperature, it is thermo­dynamically possible for chemical reac-

WElDING RESEARCH SUPPLEMENT I 447-5

(a) Reverse Polarily (DCEP) (b) Straight Polarily (DCEN)

F;g. 2 - SchetNlic of the CToss~lionof oJ submerged Me welding process showing the melted-._ "". lhe det«heddroplets and the basepUre for bolh eledrode-positWe andeIec­_galWe poIariries.

meltedwire tip

of the high current densities. and thermo­chemical reactions are expected becauseof the high temperature and the generallylarge temperature-dependent differencesin chemical potentials of the various reac·tants and products in the flux and metalphases.

An examination of the welding wire,steel plate and flux compositions in Tables1and 2 shows that the wire and steel platehave very low silicon and oxygen con­centrations, and a relatively htgh manga­nese concentration, while the flux has ahigh silica activity and a relatively lowmanganese-oxide/iron-oxide ratio. whichis far from equilibrium with the weldingwire and steel plate. Thus, the manganesecontent of the welding wire and steelplate would be expected to drop as a re­sult of thermochemical oxidation losses tothe flux, and the silicon and oxygen con­tent would be expected to increase be·cause of the reaction with the flux.

The reactions at the melted electrodetip. the detached droplet and the weldpool were considered. The chemical anal­ysis results show that a substantial differ­ence exists between the chemistries ofanode and cathode. These differences arethe resuk of electrochemical reactions.

F'llure 3 shows the plot of the averageoxygen contents of the welding wire. themelted electrode tips. and the detacheddroplets for both e1ectrode-negative andelectrode-positive polarity weldingmodes. This plot shows a very low oxygencontent in the welding wire (20 ppm) anda very significant oxygen pickup in themelted electrode tips for both polarities.The influence of thermochemical oxygenpickup is shown by the fact that signifICantoxygen pickup is observed in both elec­trode-positive and electrode-negativeconfigurations. This excess oxygen camefrom the surrounding atmosphere anddecomposition of oxide components inthe flux. The influence of electrochemicalreactions is shown by the fact that the ox­ygen content of the anode in the elec­trode-positive polarity power mode (591ppm) is over twice that of the cathode inthe electrode-negative polarity powermode (277 ppm). This oxygen contentdifference is due to oxygen pickup at theanode and oxygen refining at the cathode.The real difference is somewhat less sincemore wire is fed and melted at the cath­ode for a fixed current, thus diluting thetotal electrochemical and thermochemicaleffect at the cathode. If the electrochem­ical and thermochemical reactions wereconsidered as separate steps. the averagemelted electrode tip oxygen concentra­tion for the two polarities could be con­sidered to crudely represent the thermo­chemical contribution. and the separationof two concentrations from this meanwould repiesent the electrochemical ef­fects. However. the different wire feedrates doud this interpretation. After the

moltenflux

(-)

using a lECO interstitial analyzer, i.e.,fusion in a graphite crudble under inertatmosphere. A Baird - Atomic emissionspectrometer was used to determine man­ganese. silicon and other alloy elements inweld metals. welding wire and electrodetips. The electrode tips and wires wereflattened in a rolling mill to get enougharea (or analysis. The detached dropletswere analyzed using the wavelength dis·persive analyzer on a JEOl scanning elec­tron microscope.

A metallurgical model by Their (Ref. 5)is adopted to show the extent of elementtransfer between the weld metal and flux.Its application gives quantitative data forthe gain or loss of elements arising fromslag/metal reactions. The transfer effi­ciency into the weld can be expressed interms of the difference between analyticaland nominal compositions. The differencebetween analytical and nominal composi­tions is defined as delta (.0.) quantity. Theanalytical composition is obtained bychemical analysis methods and the nomi­nal composition can be obtained by cal­culations based on the detached dropletand plate compositions and dilution. Apositive concentration change (.0. > 0) in­dicates a gain or pickup of a particular el­ement. i.e., transfer of element from theslag to the weld metal. A negative change(.0. < 0) indicates a loss. i.e., transfer fromthe weld metal to the slag.

Results and Discussion

The purpose of this investigation is toconsider the relative influence of thermo­chemical and electrochemical reactionson the weld metal chemistry. Electro­chemical reactions are expected because

wire(2.4mm)

E

arcplasma

flux

T.bIe 2-Composilion of Welding Flux(wH.)

was obtained by adjusting the electrodevelocity, which was 72 mm/s (170 inJmin) for the anodic wire, and 99 mm/s(235 in./min) for the cathodic wire.

The electrode tips were collected bystopping the welding process. pulling theelectrode away from the weld pool andcutting off the tip. The liquid metal drop­lets are released from the electrode tipand travel through the molten flux. Thedetached droplets were collected fromwelds made at a high velocity over a wa­ter-eooled copper plate SO that the drop­lets are suspended in the molten flux. Thedetached droplets were extracted fromthe ground slag by magnetic separation.The speed of quenching attained is suffi­cient to freeze the metal composition es­tablished at high temperature. Single pass.bead-on-plate welds were made on steelplates at several welding speeds (1.55­34.67 mm/s) with constant voltage (28.5V) and current (580 A) for both polarities.The wire feed speed was adjusted tomaintain the constant current.

Analyses for oxygen were carried out

(+)

!Substance Amount

Si02 11.22AllO) 18.14MgO J323CaF1. 25.26e.o 6.92MnO 1,15

TiOl 090Nal0) 0.82fezO) 1.99e 0.37

!-.z.

-'"

448-s I DECEMBER 1990

30,..-----------------,

...:.:-:'

:f;g:Yi"~.;~:;~

~i~·>

iI:

.,:Ei~()

Fig. 5 -AveragerTldngdnese contentsof the initial wire.the electrode tipsdnd the detdcheddroplets for bothelectrode-negdtive(Cdthodic wire) dndeJectrode-positive(dnodic wire)polarities.

electrode tip, indicating that a large frac­tion of the silicon in the droplet has backreacted with more noble metal oxides inthe flux (e.g., Fe,oJl, One of the possibledriving forces for this reaction is related tothe possibility that the droplet is at a lowertemperature and has a higher oxysencontent than the tip. The average manga­nese content is further decreased bythermochemical reactions with more no­ble metal oxides in the flux when goingfrom the electrode t.ip to the detacheddroplet, which falls through and reactswith the flux.

Figure 6 compares the detached drop­let compositions for the various alloy ele­ments. Silicon, aluminum and manganese

Droplet

06

o 5 CJ Wire

Ii1iIIiill Anode

CD Cathode

'"04

•;:

Z03

0

'"-'V1 02

01

0

Wire Electrode DroplerTi p

fig. 4 - A verage SIlicon contents of the initi<ll wire. the electrode tips, endthe detdched droplets for both ele<:trode-negdtive (cathodic wire) .:Jndelectrode-positive (dnodic wire) pol.Jrities.

Wire ElectrodeTip

mochemical and electrochemical reac­tions are also indicated in the case ofmanganese; however, the high manga­nese content in the electrode and the lowmanganese-oxide (MnO)/iron-oxide(FezO]) ratio in the flux lead to thermo­chemical manganese loss at the meltedelectrode tip. Therefore, manganese oxi·dation and silicon reduction reactions aremajor thermochemical reactions in thissystem. The changes in the silicon andmanganese concentrations from the elec·trode tip to the detached droplet aremostly thermochemical. With sijicon. how­ever, there is a decrease rather than anexpected increase in silicon content of thedetached droplet compared to the melted

25 CJ Wire-AnOde

~ ~ Cothode

;: 20

wV1w 15z<t

'"Z'0<t

:;

05

0

1400

.,1200 IiiiE Anode

E CJ W,re

" 1000 [§J"~

"800

c.'u 600c0

U

c.' 400~,-"0

200

0

W,r e Electrode OropletTip

Fig. J - Averdge oxygen contents of the welding wire, the electrode rips.lnd the detdched droplets lOr both electrode-negdtive (Cilthodic wire)Jnd electrode-positive (.mooc wire) polarities

molten droplet separates from the wire.the electrochemical reaction ceases. Butthe thermochemical reaction continueswhile the droplet is tailing through themolten slag, Therefore, the detacheddroplet shows higher oxygen contentthan the melted electrode tip due to theincrease in oxygen content by the contin­uous decomposition of flux. Even thoughelectrochemical reaaion does not occurin detached droplets, the results of oxy­gen analysis still show the electrochemicalreaction effect at the melted electrodetips as indicated by the difference in oxy·gen contents between anodic and ca­thodic droplets,

Figures 4 and 5 show the average siliconand manganese contents in the initialelectrode, the melted electrode tip andthe detached droplet. The melted elec­trode tips show the increases in siliconconcentration (less increase at the anodeand more increase at the cathode) and thedecreases in manganese concentration(less decrease at the cathode and moredecrease at the anode) for both polarities,This is evidence that both thermochemicaland electrochemical reactions occurred atthe melted electrode tip. The initial weld­ing wire has a very low silicon content,while the flux has a high silicon-oxide con­tent. This causes the thermochemicalpickup of silicon from the flux. The elec­trochemical influence is Significant, as indi­cated by the fact that the cathodic elec­trode tip sjlicon content is about 0.06 wt­%higher than that of the anodic electrooetip, and by the fact that the cathode feedrate is higher than the anode feed rate,which means the total amount of silicon inthe cathodIC electrode tip is relativelyhigher than indicated in Fig. 4. Both ther-

WELDING RESEARCH SUPPLEMENT 1449-5

Q

Fig. 6 -Averdge droplet compositions for silKon. d/uminum. nickel. chromium, molybdenum, tita­nium and mimg.mese {or electrode-negdtive .Jnd electrode·poSl~ive polarities cOi7lPJred with theinitidl welding wire compositions.

DROPLET COMPOSITION

I;

On slow cooling down to the solidificationtemperature. this will lead to spontaneousreaction and losses of dissolved silicon.The silicon loss must be due to the forma­tion of some deoxidation reaction prod­uct. It is generally accepted that formationoi silicon oxide (SiOz) takes place whensufficient oxygen is available. But siliconloss is recovered by electrochemical re­duction reaction at the cathode and in­creased at the anode by electrochemicaloxidation reaction. Therefore. gliconshows much loss at the anode. and a littlegain or almost no change at the cathode.

Figure 8 shows the variations in deltamanganese due to polarity. Similar to sili­con, oxidation of manganese is favored bythe lower temperature prevailing in thecooler part of the weld pool. Here, man­ganese will react with dissolved oxygen toform slags Ihat exhibit a large loss of man­ganese in both anode and cathode weldmetal. The reason why manganese showsextensive losses in both cases is becausethe flux has a little manganese oxide(MnO) and Ihe base plate has a high man­ganese content. Therefore, manganesecan be oxidized easily. Another reason formanganese loss can be a resuh of theevaporation because vapor pressure in­creases strongly with temperature. But thedifference in delta manganese due to p0­larity shows the effect of electrochemicalreaction. That is, the calhode shows lessloss than the anode because of the elec­trochemical reduction reaction at the cath­ode, and the electrochemical oxidationreaction at the anode.

Variations in delta aluminum are shownin Fig. 9. Aluminum. like smcon and man­ganese. is also oxidized while the temper­ature of the weld pool is lowered to the

2.50

2.00~;:

1.50 ~

Q)InQ)

1.00c0

'"c0

:::;;0.50

aMnTiMo

greater than those of molybdenum, andthese cause an increase in molybdenumconcentration at the anode and a de­crease at the cathode. Thus, this behaviorof molybdenum is also the result of elec­trochemical reaction.

Figures 7 through 14 show the results ofchemical analyses of the anoele and calh­ode weld metal for silicon, manganese,aluminum, molybdenum and oxygen as afunction of welding speed. The resulls forthe anode and calhode of each elemenlare plotted together. These results showthat electrochemical reactions cause sub­stantial changes in weld metal chemistry.

Figure 7 Jlustrates the variations in deltasilicon due to polarity. High transient con­centration of oxygen may exist in the liq­uid metal al the hot weld pool as a resultof high solubililY al elevated temperalure.

CrNiAI

_ Anodec=J WireiZ'Ll Co th ode

5i

show higher concentrations in the ca­thodic droplets than in the anodic drop­lets. This is largely caused by oxidationlosses to the flux al the anode and elec­trochemical reduction from the flux at thecathoele during the reaction at the meltedelectroele tip. The differences for nickel,titanium and chromium are of the order ofthe analytical uncertainties and are thusincondusive. Molybdenum shows Ihe 0p­posite trend. That is. the concentration inthe anode is higher than that of the cath­oeIe. This behavior is difficult to explain interms of molybdenum reaction alone. bulif iron reaction is also taken into account,it can be explai"led easily. Iron is the ma­jor eleetrode constituent and it is moreeasily oxidized than molybdenum. There­fore. the loss of iron at the anode and thegain of iron at the cathode are relatively

0.50

.'.~~ 0.40..;~~

z 0.300F«n::t- 0.20zw0z0 0.100

0

..

j•, .

I-!:i .•!

·I'

-!~,~­·i;

:.'L: .

i•·•1

0.15,-------------------, --'-tangan.5e Change In the Weld Pool

0.00

Cathode........... -0.05~ ...."i • •• • •......... -0.10 • •.. •In •~ -0.15 •0 •0>g -0.20

::lE

<:J -0.250

00

00 0 0 00 0

-0.300 0 00

0 Anode

oo 0

ooo

Cathode

o 0 0

•

00 0o 0

o

•• ••

0.05

c 0.00 +---."--.A...-..--'.'--~~--~~••,...-'_;.c-ia •u

Vi -0.05<1

Anode

-0.10

0.10

Silicon Chonge in Weld Pool

J

-0. 15 -h-~T"TO""TT~TT'~...,.,...T"TO""TT......_~TT......_l0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6

Travel Speed (cm/s)F•. 7 - Can or loss 01 silicon _ to the TNCIions in the weldpool (or

eI«tIodtM~ativt!and eleet,.ode-p0si6~polarities as a function 01~speed.

-0.35 -h-...TT"'rT'f......_~.,...,.~,....,""TT~TT'~-,rrrJ0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6

Travel Speed (cm/s)Ftg. 8-Cain or loss 01manganese _to the reactions in the weldpoolfor electr0de-neg8tivt! and electrode-posi#ve polarities as a function 01wefd;ng speed.

0.'5 0.15

Aluminum Chong. in the Weld Pool UOlvbdenum Change in 'he W.ld Pool

:-;0.10 0.10~

~ ~~&<:

"i ~0 Anode0.05 ~

0.05 0

00 00

E 0 0 0 0 0 0

E 0 o 0Cathode ~

" 0.00 c: 0.00 0

c: " 0 0 0-0 0 0 0 0

A 00

E • • • • • •• • • .0 0 •.' •OJ • ~ Cathode

:< -0.05 ,p~ -0.05

<I 0 0 0 <l0 0 0 00 o 000

-0.10 0 Anode -0.100>

nurn changes show the opposite trend asin the melted electrode tip. This behaviorof molybdenum is the resuft of iron lossesat the anode and pickup at the cathode asdiscussed previously.

Figures 11 through 13 show the varia­tions in delta nickel. chromium and tita­nium due to polarity. Those elements alsoshow a little electrochemical effect. Butthe differences between anode and cath­ode are of the order of the analytical un­certainties and are thus inconclusive.

Figure 14 shows the variations in deltaoxygen due to polarity. Oxygen losses areshown at both cathode and anode. andthe cathode shows less loss than the

Anode

Cathode

anode. This is the result of the differencein the slag/metal interfacial contact areaavailable for deoxidation reaction. Thearea of the cathode is smaller than that ofthe anode. Therefore, the separation ofdeoxidation products at the anode is eas­ier than that at the cathode. Thus, elec­trochemical effect may be swept out bythermochemical reaction.

Figure 1S is the composition map forboth electrode-negative and electrode­positive polarity submerged arc welding,showing the variations in oxygen and sili­con contents during welding. The curvesare the plots of the silicon and oxygencontents of steel in equilibrium with solid

0.10

0.\5 -,--------------------,

-0.10

Chromium Change in the Weld Pool

E ...E" 0 .00 +----..-~-..~~-.----,8;;---:--:---.--,•.-1

(lPOOO 0 0 00 ego ooL

.<:U -0.05

<l

-0.15 +rr............,.~.......,~.,..,..'TT"T.",..~0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6

Trovel Speed (em/')Fig. 10 - Gain or loss of molybdenum due to the reactions in the we'dpool for electnx}e·neg<1tive and electrodepositive pol<1rities as <1 func·lion of we'ding speed.

-0.15 +~rr~TT~-.-,~.....,,.,..~rr~,.,.~"~,,~rI0.0 0.' 0.8 1.2 1.6 2.0 2.' 2.8 3.'2 3.6

Trovel Speed (em/.)Fig. 12 - Gilin or loss ofchroIrWm due to the reoetiotls" the weldpoolfor eIecrrode-negative and efecrrode-positive polarities os • function ofwelding speed.

-0.10

Nickel Change in the Weld Pool

Cathode

0.\5 ,-------------------,

0.10

Ql 0.00 +--....~-~-~~-E- ..8,--~....'.,-~..-1.y. QJ:Jct 0 0 '='0 0 0000

o AnodeZ

-0.05<I

~ 0.05

-0.1 5 +-r~"""""~"~"""-~rT"'''''''''''-~Tr~.,..rrl0.0 0.04 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6

Travel Speed (em/.)Fig. 9 - Gain or loss of dluminum due 10 the reactions in the weld poolfor electrode--neg.uive .md electrode-positive po!ilniies as d function ofwelding speed.

solidification temperature. Aluminum ox·ide (AI,Oll is stable at high temperaturerather than siicon oxide (SiO,) and man­ganese oxide (MnO), as indicated in theElingham diagram. Therefore, eventhough the flux contains a large amount ofaluminum oxide, it is difficult to pick up thealuminum from the slag by electrochemi­cal reduction reaction at the cathode weldpool. So, the results show the aluminumlosses even at the cathode weld pool. Butthe lesser loss at the cathode is theevidence for the electrochemical reac­tion.

Figure 10 shows the variations in deltamolybdenum due to polarity. Molybde-

-0.15 +~...~...~..~...~...~,,~........,0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6

Trovel Speed (em/.)F". 11-Gilin or loss ofnickeldue to the fNetiotls .. the weld pool fore/eetJo'* ._tive and .trode-positive polarities os a function of~speed.

WELDING RESEARCH SUPPlEMENT 1451-5

100O)Cygen Change In th. W.ld Pool

aCathode A

AA

A AA A""":'-100 A A

00 00

E A0 0a. 0

a. 0 0.......... -200

• AC 00 0 Anode0 "01 •>--300x 0

0 A

<l '"-400

A

-500A

Anode

Cathode

~ 0.00 +-M09H!J·Hi_---<~_&_"_&i___e~___,Il_-___&_&___it__e_{

'c2~ -0.05

<l

-O.iO

0.15 ~--------------------,

0.10

Titanium Chonge in Weld Pool

-0.15 rr'0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6

Travel Speed (em/s)Fig. 13 - Gdin or loss 'of titanium o:Jue to the reactions in the weld poolfor electrode-negdtive and electrode-positivepo/drities ~s ~ function ofwelding speed.

- 600' f-,~..,.,..,.,..,.,..,.,..,.,..,.,..,.,~~~,.,-,.,-,.,-,.,-,.,-,.,..j0.0 0.' 0.8 1.2 1.6 2.0 2.' 2.8 3.2 3.6

Travel Speed (em/s)Fig. 14 - G.lin Of.JoSS of oxygen due to the reactions in the weldpool for~ectrocJe-negatlVeand electrode-positive polarities as d function ofweidinS speed.

Composition Mops for DCEP and DCENSUbmerged Arc Welding Processes

tents in the droplets are higher than thosein the electrode tips tor bath polarities.The weld metal shows decreases in siliconand oxygen contents by deoxidation re­adion and composition variation due towelding speed. As the welding speed de­creases, the composition of the weldmetal approaches the equilibrium compo­sition.

Condusions

1) Both electrochemical and thermo­chemical readions are significant in sub­merged arc welding.

2) The thermochemical reactions movethe flux and weld metal compositions to­ward chemical equilibrium. In this study,manganese activity was higher in themetal than in the flux; therefore, manga-

nese oxidation loss from the metal to theflux was observed. Smcon activity washigher in the flux; therefore, a thermo­chemical silicon pickup by the weld metalwas observed.

3) The electrochemical reactions at theanode include oxidation losses of alloy e~

emenls to the flux and the discharge andpickup of oxygen anions fram the flux.The electrochemical reactions at the cath­ade include the reduction of metal ionsfrom the flux and the ref.,ing of oxygen.

4) Thermochemical and electrochemi­cal composition changes are greater at alow than at a high welding speed. EIec'ro­chemical reactkms are enhanced byhigher, total current flow per unit volumeof weld metal. Thennochemical reactionsat a low welding speed are enhanced byhigher temperatures and longer reactiontime before solidification.

S) Compasilion paths for the silicon­oxygen equilibrium are different for directcurrent electrode-positive (reverse polar­ity) and direct current electrode-negative(straight polarity) welds.

6) The total gain or loss of alloy ele­ments in weld metal is influenced by thecompositions of flux, electrode and baseplate, and by welding process conditions.

Further experiments with syntheticfluxes, which are chosen to minimizethermochemical reactions, are plannedand should help to better define the rela­tive importance of electrochemical reaC­tions.

Acknowled{pnents

J. H. Km, R. H. Frost and D. l. Olsonwish 10 acknowledge the support of theU.s. Army Research OffICe at lhe C0lo­rado School of Mines. and M. Blander ac­knowledges the 5Upporl of the OffICe of

0.30

EleclrodeTip

Drople!

DeEN

DeEP

/+/ I

/ I/ I

0.'0 0.20Silicon (wt.%)

1800·C

/

/

/

1600·C / ,/,/" /

Weld ...........

0.15

0.05

c 0.10

"01>-Xo

Fig. 15- Thecomposition fTMp for

direct CUff'ffitelectrode-negative(DeEN) ~nddirect

aHrentelectrode-positive

(DeEP) polaritysubmerged arc

welding showing thevatUtions in oxygenand silicon contents

durintI weIdinfI·

silica al 1600'C (2912'F) and 1SOOA C(3272 A F). The compositions of weldingwire, melted electrode tips, detacheddroplets and weld metal are platted anthe diagram far bath polarities. The w~ehas low si6con and oxygen contents. Bathanodic and cathodic electrode tips showan increase in silicon content due to theoxidation of iron and manganese. and inoxygen content due to decomposition offlux and contamination from the atmo­sphere. But the cathodic electrode tipshows agreater increase in silicon and lessincrease in oxygen, due to the electro­chemical silicon reduction and oxygen re­fining reactions. Whtle the detached drop­lets are falling through the molten flux, thedroplets pick up the oxygen by flux de­composition and contamination from theatmosphere. Therefore. the oxygen con-

The two papers contained in this bulletin provide definitive information concerning the elevated tem­perature rupture behavior of 21/4Cr-l Mo weld metals.

WRC Bulletin 354June 1990

(l) Failure Analysis of a Service-Exposed Hot Reheat Steam line in a Utility SteamPlantBy C. D. LlIldin, K. K. Khan, D. Yang, S. Hilton and W. Zielke

1965. fusion of electrode metal and ilS interac­tion wilh 1M slag durirtg submerged arc weld­ing.Avt. SV.uJcd 10; 16 to 22.

13. POlapov. N. N.• and lyubavski. K. v.1971. Oxygen content of weld metal depositedby automatic submerged arc welding. Sv.arProiz. 1: 11to 15.

14. North. T. H. 1977. The distribution ofmanganese between slag and metal duringsubmerged arc welding. Welding Rese.archAbroad n 2 10 13.

15. Potapov. N. N.. and lyubavski, K. V1971. Interaction between the metal and slag inthe reactkm zone during submerged arc weld­ing.Svilf. Proiz. 7: 9to 11.

16. Nerin. P. A.. and Malyshev. N. I. 1982Losses of manganese from electrode dropletsin arc welding in air. SVo1r. Proiz. 2: 21 to 33.

17. Grong. 0., and Christensen. N. 1982Factors controlling weld metal chemistry. FinalReport Contract No. DATA 376·61-C-QJ09.European Research Office of the U.S. Army.

18. lau. T.• Weatherly. G. C. and Mclean,A. 1985. The source of oxygen and nitrogencontamination in submerged arc welding usingea0-AlzO j based fluxes. Welding journiJl&4(12): 343-, to 347-..

WRC Bulletin 352April 1990

6. North. 1. H.• Bell. H. 8.. Nowicki. A.• andCraig. 1.1978. Slag/metal inleraction. oxygenand 100000000' in submerged arc welding.Welding Journal 57(3): 63-s 10 7S-s.

7. Dav~. M.l. E.. and Cae. F. R. 1977. Thechemistry of submerged arc welding fkJxesThe Welding Institute. 39/19771M.

8. Milra. U. 1984. Kinetics of slag metal re·actions during submerged arc welding Of sleel.Ph.D. dissertation. MIT. Cambridge. Mass.

9. Blander. M., and Olson, D. l. 1984. Ther­modynamic and kinetic factors in the pyro­chemistry of submerged arc flux welding ofiron-·based aDays. Proc. IntI. Symp. on Metallur­gicalSlag, and Fluxes. The Mel. Soc. of AM.

10. Frost R. H.• Olson, D. L. and Edwards.C. R. 1983. The influence of electrochemicalreactions on the chemistry of the electroslagwelding process. Engineering foundation con­ferences. NfodeJing of C.IS/ing and WeldingProcesses H. eds. J. A. Dantzig and J. T. Berry.The Mel. Soc. of AM.

11. Blander. M.. and Olson. D. L. 1986. EIec·trochemical effects on weld pool chemistry insubmerged arc and D':,c. electroslag welding.Proc. /ntl. Coni. on Trends in WeIdHlg Rese.Jrch.

12. Pokhodnya, I. K., and Kostenko. B. A.

(2) The Influence of Flux Composition of the Elevated Temperature Properties of Cr-MoSubmerged Arc Weldments8y J. F. Henry, F. V. EDis and C. D. Lundin

Naval Research al Argonne National Lab­oralory under Navy Order No. N00014­85-F-()(l97.

The first paper gives a detailed metallurgical failure analysis of cracking in a longitudinally welded hotreheal pipe with 184.000 hours of operalion at 1050°F. The second paper defines the role of Ihe weldingflux in submerged arc welding of 2'/.Cr·1 Mo sleel.

Publicalion of Ihis reporl was sponsored by Ihe Sleering and Technical Committees on Piping Syslemsof the Pressure Vessel Research Council of the Welding Research Council. The price of WRC Bulletin 354is $50.00 per copy. plus $5.00 for U.S. and $10.00 for overseas poslage and handling. Orders should besent with paymenl to Ihe Welding Research Council, 345 E. 47th 51.. Room 1301. New York. NY 10017.

In Oclober 1987. the PVRC Sleering and Technical Commillees on Piping Systems eSlablished a taskgroup on independenl supporl motion (ISM) 10 evaluate the technical merils of using Ihe ISM melhod ofspectral analysis in the design and analysis of nuclear power plant piping syslems.

The results of the lask group evalualion culminated in a unanimous lechnical posilion Ihat Ihe ISMmethod of spectral seismic analysis provides more accurate and generally less conservative responsepredictions Ihan Ihe commonly accepted envelope response speclra (ERS) method. and are reported inIhis WRC Bulletin. The price of WRC Bullelin 352 is $25.00 per copy. plus $5.00 for U.S.• or $10.00 foroverseas, poslage and handling. Orders should be sen I wilh paymenl to the Welding Research Council,345 E. 471h 51.. Room 1301. New York. NY 10017.

References

1. Belton. G. R., Moore. T. J.. and Tankins. E.S 1963. Slag-metal reactions in submerged-arcwelding. Welding Jour""'42(7): 289-< 10 297-s.

2. Chai. C. 5.. and Eagar, T. W. 1982. Slag­metal reactions in binary CaFrmetal oxidewelding fluxes. Welding Joumd/61(7): 229-5 to232-s.

3. Chai. C. S.. and Eagar. T. W. ·'981. Slag'metal equilibrium during submerged arc weld­ing. Mf!rall. Trans. B 128(9): 539 10 547.

4. Indacochea. J. L Blander. M., Christen­SM. N .• and Okon. D.l. 1985. Chemical reac­tions ck;ring submerged arc welding with FeO­MnO-SiG,: fluxes. foAetaU. Trans. B 168(6): 23710245.

S. Their. H. 1980. Metallurgical reactions insubmerged arc welding. Proc. Inti Conf onWeld Pool Chemistry 'llJd Meldllurgy. TheWelding Institute. london. England. pp. 271 to278.

WElDING RESEARCH SUPPlEMENT 1453-5

BY J. H. DEVLETIAN, D. SINGH AND R. B. TURPIN

Nonconsumable guide electroslag welding provides a highly cost-effective means toweld T-joints in double-hull ship construction

ProcedureThe nonconsumable guide ESW process was uti­

lized to join two 25-mm-thick hull plates to a 25-mJ"!l­thick stiffener in a T-joint configuration as shown In

Fig. 1. The weld joint cavity consisted of a 25 x 25 mmspace between the three plates. The length of the weldwas 2.4 m (8 ft) plus run-on and run-off tabs. The 25­mm-thick steel plates used included 1) dual-certifiedA36 and AB5 Grade Aand 2) AB5 Grade C5 with com­positions given in Table 2. There were two water­cooled stationary shoes to form the two 6-mm (1/4­in.) fillets of the T-joint. To ensure smooth filletformation, the stationary shoes were machined with arecess, which permitted only about 30% of the cool­ing shoe surface to contact the steel plate. In this way,the copper shoes would slowly extract heat to allowthe fillets to fill and freeze with a smooth contour. Un­like commercial ESW systems, the snorkel or nozzle(Fig. 2) was not attached directly to the slidin~ shoe inorder to adjust the sliding shoe and nozzle md~pen­

dendy. The sliding shoe was machined to proVide a1.6-mm (1{16-in.) reinforcement on the flat face oftheT-joint.

The power source for ESW was a 1SOO-A. 100%duty cycle, DC constant voltage machine used withelectrode positive. Weld heat input averaged 25kJlmm (635 kl/in.) using 700 A, 47 V and 76 mmlmin(3 inJmin). An automatic flux feeder was attachedabove the slag pool (shown in Fig. 2) to compensatefor the thin layer of slag that was continuously con­sumed on the cold shoe surfaces during welding.

All-weld-metal tensile specimens and Charpy V­notch (CVN) impact toughness specimens were cutfrom the weld joint as shown in Fig. 3. In addition,nine 2.4-m-long electroslag welds were incorporatedinto a 6.4-m (21-fll I-beam for fatIgue testoog. Asshown in Fig. 4, each electroslag welded T.. joint was

fitted and submerged arc welded into a 21 :ft­long I-beam for fatigue testing. The four-polOtbending fatigue test was designed to evaluatethe electroslag weld joint performance in bend­ing fatigue. Each beam was .cyclically loadedwith a single 489 kN (110 kIps) dIgItally con­trolled, hydraulic actuator with an in-line 667kN (150 kips) load cell. Each actuator was cen­tered on 1.2-m (4 0 ft) spreader beams, which dis­tributed the load equally to load points on thetest specimen. The actuators were controlled bya computer to produce sinusoidal waveforms at

345 MPa (SO ksO min.49G-{oSO MPa (70-95 ksil20 J • -1 SoC (l5 ft-Ibs " (JOF)27 J(20 ft-Ibs) " -1 (JOC (l4°F)AWSlAASHTO 01.5 category B

for nonredundant strudures

Table I-Weld Metal Properly Requirements

S··, ince Congress passed the Oil Pollution Act of1990, all tankers entering U.S. harbors musthave double hulls by the year 2010 (Refs. 1, 2).

A typical double-hull design for an oil tanker or mil­itary ship has a greater number of hull-to-stiffener T­joints (Ref. 3) than a standard design, and these areideal applications for electroslag welding (E5W). Thefabrication of double hulls is extremely competitive(Refs. 4, 5) with nearly all current double-hull tankerfabrication being performed overseas. Although E5Wwas widely used in the construction of buildings (Ref.6) and bridges (Ref. 7), it has seen limited applica­tions in ship construction in this country (Refs. 8, 9).

Over the last two decades, many innovative im­provements in consumable guide ESW were devel­oped to improve the integrity and fracture toughnessof electroslag welds for bridges and buildings. By in­corporating a specially alloyed tubular filler metaland narrow groove techniques for reduced heatinput, substantial improvements in both weld metaland HAl toughness were achieved (Refs. 10-13). In1995, Juers and McConnell developed a procedureto weld T-joints in a single pass using the electrogasprocess (Ref. 14). This welding was incorporated intothe Navy's advanced double hull (ADH) design.These T-joints consisted of three orthogonal 25-mm(1-in.) thick plates of AB5 C5 grade steel and werejoined by electrogas welding i~ a single pass with acommercial 3.2-mm (1/8-10.) dIameter self-shIeldedwelding wire.

The objective of this work was to demonstrate thecost-effectiveness and design acceptability of joiningthree 2S-mm-thick structural steel plates in aT-jointconfiguration (similar to the ADH design) with a sin­gle full joint penetration uphill weld containing fil­lets using the nonconsumable guide ESW process.The weld metal property requirements are listed inTable 1.

Yield SlWlgthTensile strengthCVN toughnessCVN toughness (HAl)Fatigue strength

Electrcslag Welding of an AdvancedDouhie--I IuU.-Design Ship

.......~._ _--_ __ _ _.._ __ _----_.-._.__.._.._--.._.._--_._-----_ -

J. H. D~iTIAN is Ptof";";'-.nd R. B. TURPIN is Senior lIese.rch Engineer in t"!' Dept. ofMlteri.Is SCience ~ Engineering,Oregon Gradu.te Institute, Pottl.nd, Ore. D. 51NGH is Product Development Englfleer, Stoody Company, Bowl,ng Green, Ky.

WELDING JOURNAL I 49

Table 4-CVN Toughness of Weld Metal and HoU for EIectrosI"gwelds Deposited with EWT-G Metal-Cored Fil.... Metal on 25.mm·1hick CS Grade Plate at Different Levels of Heat Input

Average CVN Toughness at -1 O~C 04°F) (35 k)/mm 31 kJ/mm 25 kYmm

(890 k)lin.) (790 Win.1 (640 k)/in.)

52 J (38 ft-Ibs)94 (69)

67) (49 ft·lbs) 87 I (65 ft-Ibsl97 (71) 122 (90)

67 (49)

57 (42)

58 (43)

48 (35180 (59)

46 (34145 (33)45 (33)

Weld MetalWeld Center" 4 Distance

(5.. fig. 5)HAZ at mid thickness

Weld interfacelmm2 mm

Unaffected basemetal

:~ ABS Grade Sleeis Filler Metal Weld MetalGrade CS EWT-G EWT-G

fully Metal E70S-3 depositedKilled, Cored Solid on Grade

Grade A Normalized Wire Wire CS-'<. C 0.13 0.12 0.034 0.11 0.089

~!~fi ~.n1.1 01.3 1.2 1.12 1.23022 0.25 0.4 0.50 0.38

':\:~~:~~ ~~2.04 1.050.4 0.16

·:;-~f~ ~ 0.012 0.011 0.010 0.010 0.0110.008 0.006 0.010 0.009 0.013

AI 0.04 0.05 0.02 0.024Ti 0.06 0.03

Fig. 1 - r·joint geomelry for ESW lhree orthogonal 25-mnHhick sreelplates.

(

Fig. 2 - Top 'v~w of lhe nonconsumable guide ESW s)'srem ro weldflange-to-web T-joints. Automatic f7ux feeding system is shown abovethe snorkel.

,

6mmV(1/4in)... --

--:..---------..}, I, II II I 25mmI II I

(1 in)I I

l : II I_____________________.J

irequencies from 1 103hz, depending on the applied loadrange.

average weld metal yield strength was 452 MPa (65.5 ksi) com­pared to the base metal yield strength of 283 Mra (41.1 ksil.

Tensile Properties of Weld MetalThe all-weld-metal tensile properties of the electroslag weld

joint met the requirements of FES70-EWT-G category perAASHTO 01.S and AWS 01.1, which specifies 345 Mra (SO ksi)minimum yield strength and 490--650 Mra (70--95 ksil tensilestrength. The actual tensile properties are given in Table 3. The

Table 3-Average Tensile Properties of CS Grade Base Metal andWeld Metal Deposited by ESW Using EWT-G Metal Cored FillerMetal

Weld Metal Base Metal25 k)/mm ABS Gr. CS

Yield strengthMPa 452 283(ksi) (65.5) (41.1)

Tensile strengthMPa 647 454(ksi) (93.9) (65.8)

% Elongation 27 30%R.A. 51 54Hardness, HRB 90 71

CVN Impact Toughnessof Weld Metal and HAZ

The results of eVN toughness testing of the weld metal andHAZ, which included tests 1 and 2 mm from the weld interface,showed thaI the requirements for 20 Jat-18°e (1 5 ft·lbs at 0°F)in the weld (Fig. 5) and 27 J at-lOOe (20 ft-Ibs at 14°fj in theHAl (Table 4) were met. Furthermore, decreasing the heat inputfrom 35 kl/mm to 25 kJlmm resulted in increasing eVN tough­ness for both the weld melal and the HAl. As shown in Table4, the eVN toughness of Ihe EWT-G weld metal deposited at 25kJlmm exceeded that oithe unaffected base metal. Although theHAZ exhibited the lowest CVN toughness values, these valueswere still well above that required by the ASS Grade es spec­ification.

The CVN transition curves for the weld metal (shown in Fig.5) illustrate the superior toughness of the Mn-Ni-Mo steel tubu­lar filler metal (EWT-GI over the conventional mild steel (E70S­3) solid filler metal. 5ince dilution of the weld metal by the AS5Grade CS base metal approached 50%, the Mn-Ni-Mo fillermetal wire provided the means to improve CVN toughness ofthe weld metal. The reason for the excellent CVN toughness isthe low heat input and the favorable microstructure.

50 I AUGUST 1997

30

60 .a~:

90

100

(2.4mm)

I

:: -_ .........•

(3.2mm.~7""

/~

...~.........

-'00 -50 a 50Temperature, ·C

EWT-G

·15<)

"""Il"2."3J<.~:IWeb:I.x14' _ 17'

Range' I'x 1 3/.·

~

o

160

um{irij

.,~ 120..cW

~ 80

o11< 40Zr;

Fig. 4 - Fiftiglle test setup to evaluate the web-to-flange electros/agweld T-;oint.

Fig. 5 - Charpy V-notch toughness transition curves for weld metal(centerline location) deposited on ABS Grade CS steel by ESW at 25kj/mm.

FiB. 6 - Etched cross-section of Tojoinr deposited by ESW.

I_______J

PlateRolling Direction

I ..

--...,.-- ~,______ .J

,.-----....r- ----,HAZ1,.. j

,..--. i--..r---.,.,~ Distance ~ ~

Weld Centerline [

. . ,..----..r---.,Fusion line 1,.. ~

- -----I-- ~----

Fig. J - Locations o;ChiffPY V-notch toughness specimens of the weldmetal and HAZ extracted from the mid-thickness ofelectros/ag weldedT-joints.

Fatigue PropertiesNine full size I-Beams were fabricated out of the lA-m-long

electroslag welded T-joint as shown in Fig. 4. In this figure, theelectroslag-welded T-joint is Ihe web-ta-flange portion of Ihe 1­beam. Beams were tested at three stress range levels: 138, 207,and 276 MPa (20, 30, and 40 ksi). For example, using a stressrange of 138 MPa 120 ksi), the maximum stress was +10 ksi (69MPa) in tension to -10 ksi (-69 MPa) in compression for a max­imum of 2 million Hz. In all cases, the beams containing thef1ange-to-web T-joints passed the Navy's stringent requirementsfor fatigue. These requirements include passing the Category Bcrileria of both AW5 D1.1 Structural Welding Code - Steel(Ref. 6) and M5HTO/AW5 Dl.5 Bridge Welding Code (Ref. 7)

Resistance to Solidification CrackingDespite Ihe fast welding speeds of 76 mm/min (3 inJmin)

and low heat input of 25 kVmm (635 kjJin.), the weld metal ex­hibited excellent resistance to solidification cracking. The fac­tors promoting the resistance to cracking were 1) filler metalcomposition (containing very low carbon with Mn-Ni-Mo al­loying) to produce an eqlliaxed grain structure at the weld cen­terline, and 2) a tubular filler metal to reduce the depth!widlhratio of the weld pool. Ultrasonic testing and metalJographicsections of weld joints showed no evidence of cracking. A weldjoint cross,.section is shown in Fig. 6.

Microstructure of WeldThe micrOSlruclure of weld metal deposited on Grade C5

sleel wilh E705-3 solid wire (Fig. 71 contained very Iiltle acicu­lar ferrile. Thus. the CVN toughness values for the E705-3 weldmetal were less than those for the EWT-G weld metal (Fig. 5),which contained a mixture of acicular ferrite and low-carbon

.;

WELDING JOURNAL I Sl

-,,- '. '-, . '-

,~

" .fig. 7 - Microstructure of weld mefc11 deposited by fSW ,1t 25 kj/mmusing E10S-3 filler metal. 400X.

bainite as shown in Fig. 8. The EWT-C wire was designed withMn. Ni and Mo and very low carbon to produce electroslagweld metal containing a mixture of tough acicular ferrite andbainite. Furthermore, the addition of Mn. Ni and Mo in thetubular filler metal virtually eliminated the OCCurrence of grainboundary ferrite - Fig. 8. The use of Ihe EWT·G wire appearsto be ideally suiled for eleclroslag welding of mild sleels.

Economy of Electroslag WeldingElectroslag welding should be of interest to commercial

shipbuilders as well as heavy-section steel fabricators becauseESW represents a substantial cost savings over conventional arcwelding processes. For example. similar welds deposited in aT-joinl configuration by FCAW and SAW would require at least18 passes and 7 passes (Ref 141. respectively, compared to asingle full joint penetration pass by ESW. The deposition ralefor ESWof Ihe T·joinl was approximately 50 Ibfh, while the ralesfor FCAW and SAW were 7 and 15 IbIh. respeclively (Ref 141.Since the entire ESW syslem (siding shoe, flux feeder, powerinpul, etc.) are all electrically controlled, the process can befully automated with little additional effort. With respect to theshop environment, the ESW process is noiseless and smokeless.

ConclusionWeld melal deposiled by nonconsumable guide ESW of 25­

mm-Ihick ABS Grade A and Grade CS sleel plales in a T-joinlconfiguration meets all the requirements of ABS, AASHTO andAWS. Electroslag welding is a viable and cost-effective weldingprocess for construction of double-hull structures.

Acknowledgments

The authors are grateful 10 the Navy Joining Center for sup­porting this project and to Ray Juers of Naval Surface WarfareCenter for his guidance. The authors are also grateful to StoodyCo. for providing tubular filler metals.

References

1. Tankers playing by uncle's rules. Aug. 1992. Marine iog. pp.50-55.

2. Greening the world tanker fleet. Sept. 1991. Marine Log, pp37-43.

3. Beach, J. E. 1991. Advanced surface ship hull technology - clus­ter B. Nava/ Engineers/oumal, pp. 27-37.

4. Frankel, E. G. 1996. Economics and management ofAmericanshipbuilding and the potential for commercial competitiveness. JournalofShip Production. Vol. 12 (1), pp. 1-10.

S. Daidola. J. c., Parente, Land W. H. Robinson. 1996. Pro­ducibiJity of double hull tankers. Journal of Ship Production, 12(1):20-38.

6_ ANSVAWS 01.1-96, Structural Welding Code -Steel, AmericanWelding Society, Miami, Fla.

7. AWSIMSHTO 01.5-95, Bridge Welding Code, American Weld­ing Society.

8_ Parrot, R. S.• Ward, S. W_. and Uttrachi, G. D. 1974. Electroslagwelding speeds shipbuilding. Welding Journal, 53(4): 218-222.

9. Howser, B. C. 1985. Applications and trends in electroslag weld­ing in the United States. Proceedings ofElectms/ag Processing for Ma­rineApplications, Report EW-20-85. United States Naval Academy. pp.4.1-4.11.

10. Yu, D. W., Ann H. S.• Devletian, 1. H., and Wood, W. E. 1986.Solidification study of narrow gap electroslag welds. Welding Research:The State ofthe Art. E. F_ Nippes and D. J. Ball (eds.), ASM International.Materials Park, Ohio, pp. 21-32.

11. Yu, D. W.• Ann, H. 5., Devlelian, ). H., and Wood, W. E. 1987.Improvements in microstructure and properties of steel eleclroslagweldments using low alloy lubular filler metal. Welding Metallurgy ofStructuralSteels, J. Y. Koo (ed.), The Metallurgical Societyo(AIME, War·rendale, Pa., pp. 233-254. •

12. Wood, W. E.• and Devletian, J. H. 1987. Improved FractureToughness and Fatigue Characteristics of Electroslag Welds, PublicationNo. FHWA/RD871026, U.S. Department of Transportation FederalHighway Administration, Turner-Fairbank Highway Research Center,Mclean, Va., p. 244.

13. Turpin, R. B., Scholl, M .• Devlelian, )., and Wood. W. 1983.Field welding of rail by the electroslag process. Proceedings ofSecondInternational Heavy Haul Conference, Colorado Springs, Colo., Paper­82-HH-16, pp. 142-157.

14. Juers. R. H., and McConnell. F. 1996. Electrogas welding of ad­vanced double-hull subassemblies. Welding Journal, 75(1): 45-53.

(