h - review of welding

8/3/2019 H - Review of Welding

1/90

-1-

A REVIEW OF WELDING CAST STEELS AND ITSEFFECTS ON FATIGUE AND TOUGHNESS PROPERTIES

By: John F . Wallace*

TABLE OF CONTENTSPage

Outl ine of the Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2Summary of Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

I . INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

II . WELDABILITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Metal lurgical D iscontinuities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Welding Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Types of Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Low Carbon Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Medium Carbon Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22High Carbon Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Alloy Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

I II . M ECH AN IC AL P RO PE RT IE S O F W EL DS . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 8Fatigue Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 9

Weld Configurat ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Bas e Metal Strength Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Hig h and Low Hyd ro gen Elect rodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Weld Discontinuities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Welding Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41Residual Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 4Post-Weld Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Toughness of Welds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Types of Tests for Toughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Weldi ng Pr oces ses and Ele ctr odes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 9

S hi el de d m et al ar c m an ua ll y a pp li ed . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 9Flux cored electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1Submerged arc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Gas shielded arc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59Electroslag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Toughness in HAZ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1Effects of Various Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 4

Chemical composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64Individual e lements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66Microstructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69Heat inputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69Heat t reatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3

Welding Discontinuities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 8IV. APPLICATION OF FRACTURE TOUGHNESS CONCEPTS TO

M ECH AN IC AL B EH AV IO R O F W EL DS . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 0V. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 4

Types of Discontinuities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

by STEEL FOUNDERS' SOCIETY OF AMERICA. 197920611 Center Ridge RoadRocky River, Ohio 44116

Printed in the United States of America

*Professor of Metallurgy, Case Western Reserve University. Cleveland. Ohio


2/90

A REVIEW OF WELDING CAST STEELS AND ITSEFFECTS ON FATIGUE AND TOUGHNESS PROPERTIES

OUTLINE OF THE PROBLEM

Welding is extensively employed in the f inishing stage of s teelcasting production and in fabricating larger components by joini ngcastings or by joining castings to wrought s teel components . The easeof welding as well as the effects of welding on mechanical propertiesare thus very important considerations in the production and selectionof castings as engineering components . The purpose of this review istherefore to summarize and evaluate the published l i terature withrespect to weldabil i ty of carbon and low alloy steels and the effectsof welding processes on mechanical properties .

SUMMARY OF CONCLUSIONS

The technical l i terature has been reviewed and analyzed to pre-sent the s ignif icant factors that determine the weldabil i ty of caststeels and the effect of these welds on th e fatigue and toughnessbehavior . The effect of welding processes, the s ignif icant var iablesof these processes, the chemical composit ion, pre- and post-weldingtreatments and presence of discontinuit ies on the mechanical properti esare indicated with emphasis on the fatigue strength and toughness,

The weldabil i ty of carbon steels decreases with increasing carbon,alloy and sulfur contents .higher temperature preheats and greater use of postheating treatments

to obtain satisfactory welds. Fatigue properties are improved byremoval of weld reinforcements , the use of low hydrogen shielding,automatic compared to manual welding, full heat treatments and peening.The toughness of welds is optimized by: low carbon, oxygen, nitrogen,sulfur and phosphorus contents with a minimum alloy content for therequired strength level; low heat inputs and multipass welds; lowhydrogen arc shielding and basic f luxes; and f lat welding posit ionsand automatic welding processes.

The higher carbon and alloy steels require

The presence of discontinuit ies can lower both the fatigue andtoughness properties markedly.var ies with the s tress system under dynamic loading. Generally, the

most damaging discontinuit ies l is ted in order of decreasing severityare: cracks, undercuts , s lag inclusions, porosity and the p resenceof weld reinforcements .

The effect of these discontinuit ies


3/90

- 3-

PREFACE TO 1979 PUBLICATION

This publicat ion "A Review of Welding Cast Steels and I ts EffectsOn Fatigue and Toughness Propert ies" has been issued based on a reportissued in 1974 to members of the Steel Founders ' Society of America.This publicat ion has been prepared as a supplement to the new report"Repair Welding and Fabricat ion Welding of Steel Cast ings" issued bythe Steel Founders ' Society in 1979.

PREFACE TO 1974 REPORT

Advances in welding technology, in quali ty control of welds, andin performance of welded structures have led to increased customeracceptance of welding as a regular part of the foundry production pro-cess. Welding is now recognized as a procedure for upgrading cast ingquali ty during the course of manufacture through improvement of surfacecondit ions, or by el imination of shrinkage voids. I t is a lso acceptedas a method of producing large or complex assemblies where the sizeof the completed structure precludes production as a one-piece cast ing,or where total quali ty wil l be improved by dividing the structure intosimpler components which can la ter be welded into an integral assembly.

With the customer 's and cast ing user 's acceptance of the weldingprocess, welding has gained a posi t ion where the foundryman has to becognizant of a l l the welding processes, their capabil i t ies, precautionsnecessary for sat isfactory weldabil i ty , quali ty control and the effectsof welding on propert ies of the weld and the weld affected areas of thecast ing. Steel foundries have recognized these implicat ions quite early.A booklet , "Recommended Pract ice for Repair Welding and Fabricat ionWelding of Steel Cast ings", was therefore f irst published by the SteelFounders ' Society of America in 1957, with a revised and updated versionfollowing in 1969.

The above publicat ion on welding pract ices has been of greatbenefi t to steel foundries and their customers. Member foundries,however, have shown increased interest in the quest ion of weldabil i tywhich means the ease of welding or the precautions to be taken toassure a sat isfactory weld, depending on the welding process and al loyto be welded, as well as the effects of welding on propert ies andperformance of the weld and weld-affected area.

The Carbon and Low Alloy Technical Research Committee thereforerecommended an in-depth survey and evaluation of the published l i terature .This recommendation was approved by the Board of Directors in 1973 asResearch Project 95.

Professor Wallace, a t Case Western Reserve Universi ty in Cleveland,was asked to undertake this act ivi ty. The Research Committee complimentshis effort that yielded a unique report where the important aspects havebeen assembled and reviewed with outstanding competence and insight .Acknowledgements are a lso made to Walter E. Evans, Ralph D. Maier andJohn C. Rogers, graduate students a t Case Western Reserve Universi tywho assisted in the sect ions on weldabil i ty , fa t igue strength, andtoughness port ions, respectively.

PETER F. WIESERTechnical and Research Director

By direct ion of theTechnical Research Committee


4/90

-4-

A REVIEW OF WELDING CAST STEELS AND ITS

EFFECTS ON FATIGUE AND TOUGHNESS PROPERTIES

I. INTRODUCTION

This report reviews and analyzes the American and Brit ish

l i tera ture on the weldabi l i ty of carbon and a l loy s teel cas t ings fortota l a l loy contents up to 5%. The inf luence of var ious weldi ngprocesses and techniques including electrode selection, pre- and post-treatments on the susceptibil i ty to welding discontinuit ies and themechanical properties of the welded structures has been studied.Considerable emphasis has been placed on the dynamic properties ,fatigue and toughness because of the s ignif icance of these on servicebehavior . This report f irs t discusses the weldabil i ty of the var ious

types of s teel and the problems with welding discontinuit ies and thenthe effect of the welding processes and welding discontinuit ies on themechanical properties .

The weldabil i ty of s teel castings is of considerable s ignif icancebecause the welding process is employed extensively for the repair of discontinuit ies and for fabricating larger components f rom castings.Repair welding and fabrication practice for s teel castings was thesubject of an ear l ier publication of the Steel Founders ' Society of America (1) .other castings or wrought s teel to produce larger s tructures is anestablished commercial procedure (2) .

II . WELDABILITY

The use of welding procedures to join s teel castings to

No specif ic cr i ter ion of weldabil i ty has been generally accepted;the term is used to descr ibe the ease with which a metal can be weldedto produce a weldment of acceptable quali ty.usually judged from the standpoint of mechanical properties (3) :

The weld quali ty is

(1) The strength of the joint must be at least as great as thatof the parent metal;

The fracture ducti l i ty of the weld metal and heat affected

zone (HAZ) must be suff icient to ensure that the br i t t lefracture properties of the s tructure in service are notl imited by these factors alone;

The fatigue properties of the joint should not be impairedby the metallurgical condition of the weld metal or HAZ;

The metallurgical condition of the joint should not impairthe behavior of the s tructure during service as a resultof localized corrosion, etc.

(2)

(3)

(4)


5/90

-5-

The metal which can be welded to f i t the above cr i ter ia with nospecial precautions to prevent discontinuit ies or other diff icult iesis considered to have good weldabil i ty.

The strength, ducti l i ty, and to a large extent, the susceptibil i tyof the weldment to cold cracking can be controlled by ensuring that theweld metal and HAZ have the proper metallurgical s tructures.s tructures are determined mostly by the chemical composit ion and coolingconditions. I t is therefore desirable to provide close control of bothvariables. The most common and effective means of controll ing coolingrates in the weld area is to preheat the casting to be repair weldedor fabricated to increase the base metal temperature and to decreasethermal gradients around the weld. This preheating may be localizedaround the weld area by using gas torches and various types of insula-t ion. However , a general preheat of t he entire casting is preferredwhen ecomomics and the welding operation permit . This general preheat

can also be conducted in var ious types of portable equipment composedof gas torches and insulating or protective hoods. Again, i t ispreferable to preheat in a permanent recirculating gas or oil f iredoven where the castings are placed in the oven and brought to temperaturebefore their removal. In any case, some means should be available of insuring that the castings are up to the required preheating temperaturebefore welding is init iated and maintained at that temperature through-out the welding operation. Surface reading thermocouples or crayonsthat melt at a specif ic temperature are commonly used for this purpose.

These

The specif ic preheat temperature is dependent on the s teelcomposit ion, section size, the degree of restraint at the joint , the

welding method and the ambient temperature. Preheating temperaturesmay vary from 100F to over 1000F, but most recommended temperaturesare 6 00F or less (4) . The benefits obtained from proper preheatinginclude:and residual s tresses and distor t ion. The recommended preheatingtemperatures are discussed for each of the var ious types of s teel inthe subsequent sections.

Types of Discontinuit ies

the prevention of cold cracks; the reduction of HAZ hardness;

One of the considerations of weldabi l i ty is f reedom from welddiscontinuit ies . Just as the casting process may result in discon-

tinuit ies that are f requently repaired by welding, the fusion weldingprocess is susceptible to cer tain ir regular i t ies that may requirecorrection for some applications. These weld discontinuit ies aregenerally repaired by addit ional welding (5) .processes and steels are more susceptible than others to these discon-tinuit ies; the weldabil i ty is considered to be better for these s teelsand methods where less diff iculi t ies are encountered.

Some types of welding

The major types of weld discontinuit ies in fusion welds arelis ted below and are shown schematically in Figure 1 (5) .


6/90

-6-


7/90

-7-

1 .

2. Inclusions: oxides, slag and tungsten.

3. Geometric imperfect ions: undercutt ing, underfi l l , excessivereinforcement, surface irregulari t ies, dropthrough and

mismatch.

Incomplete fusion and joint penetrat ion.

4. Metal lurgical Defects:a) Defects re lated to segregation:

hot cracking and microfissures;cold cracking, delayed cracking, porosi tyand subsur face shrinkage.

Imperfect ions induced by metal lurgical react ions:

embrit t lement;

metal lurgical notches.

b)

5. Other imperfect ions: arc str ikes, weld spatter (5) .

Incomplete fusion discontinuit ies are condit ions in a weldedjoint where adjacent layers of weld metal , base metal or weld metal andbase metal fa i l to melt together properly.lack of penetrat ion of the ful l base metal thickness; or beads fai lto intermelt ; or the presence of slag, oxide or other foreign materialsat interfaces prevent melt ing of adjacent materia ls. Incomplete-fusion discontinuit ies may be the result of faulty welding techniquesand improper welding condit ions. Insuffic ient welding current or

voltage, or excessive root thickness is a frequent cause. Misal ign-ment of the welding torch and t he l ine of welding can also producethem. Incomplete-fusion with previous beads often results when deepcrevices or undercuts are present in the previously deposi ted beadsor when slag removal is incomplete .discontinuit ies are common in welds, they can be el iminated entirelyby the use of proper welding condit ions (5) . Incomplete penetrat ionboth reduces the load bearing cross sect ion and acts as a stress concen-trator .

They usually result from:

Although incomplete-fusion

Inclusions in welds are undesirable foreign matter that wasintroduced during the welding operat ion. Oxide f i lms, slag from the

welding electrodes, and tungsten part ic les from gas tungsten-arcwelding electrodes are the inclusions usually found in fusion welds.These oxide f i lms can occur with improper shielding during welding.Slag inclusions are sol id non-metal l ic part ic les that us ually occur ascontinuous or intermit tent str ingers held between weld beads, orbetween weld beads and the base metal . Slag discontinuit ies occurmainly in welds made with the shielded metal-arc and submerged-arcprocesses that ut i l ize a welding f lux.techniques and/or improper welding techniques cause them.inclusions are part ic les of tungsten in weld metal deposi ted with thegas tungsten-arc welding process and are condidered to have the sameeffect as slag inclusions of equal size . L ike incomplete fusion,

Inadequate slag-removalTungsten


8/90

-8-

inc lusions can reduce weldment proper t ies and serv ice performance ,

depend ing on the i r cha rac te r i s t i c s , o r i en ta t ion and the se rv icecondit ions (5) .

Geomet r i c imper fec t ions a re we ld -shape fea tu re s such a s unde rcu t ,

excess ive re in fo rcement , poor p ro f i l e , excess ive we ld wid th , and o therstha t resul t f rom improper contour ing of the weld .f r o m i m p r o p e r w e l d i n g c o n d i t i o n s o r i n a d e q u a t e c o n t r o l o f we l d in gopera t ions . Undercut denotes depress ions a long the edges of a weld ,and paral le l to i ts length. They form as intermit tent or continuousgrooves which vary in depth and width. Undercut occurs when surfacetension forces draw the metal away from the sides of the weld groove.The weld reinforcement is that amount of deposi ted weld metal thatprojects beyond the surface of the base metal in the thickness direct ion.I t can result from high heat input ra tes. The most common irregulari t iesin the surface of welds are surface r ipples. Although they are diff icult

to avoid, r ipples can general ly be removed by grinding or machining (5).

Cracks are a very serious form of welding discontinuity andare classif ied as hot cracking, cold cracking and fissuring. Hot cracksare essentia l ly small hot tears and formed for the same reasons.cracking at the weld centerl ine is encountered with weld pools of atear drop shape and avoided by el l ipt ical shape weld pools.occur a t or near room temperature and can form a considerable periodafter the weld has been made. Cracks are also classif ied accordingto locat ion as i l lustrated in Figure 2 (6) .is re lated to the condit ions producing these cracks and crackingproblems are frequently referred to by these designations in this

paper.deposi ted weld and are produced by excessive hardness and bri t t lenessin the HAZ, stresses in the area or hydrogen. Toe cracks have similarcauses but can be observed at the surface of the weldment (5) .

Porosi ty consists of numerous gas pockets in the weld result ing

They usually result

Hot

Cold cracks

This type of c lassif icat ion

Underbead cracks are cold cracks occurring in the HAZ under the

from the evolution and entrapment of gas from molten metal duringsolidif icat ion.with bright , smooth walls, they vary in size and shape.may be present as i rregularly shaped gas pockets a long dendri teboundaries, or as tubular gas pockets; "pipe" or "wormhole" porosi t y.The gases that cause porosi ty may come from welding materials themselves,

from welding unclean or wet components, or from the use of improperwelding condit ions (5) .

Metal lurgical Discontinuit ies

Although the pockets usually are spherical in shapePorosi ty also

An understanding of metal lurgical discontinuit ies or structuralvariat ions in the welding of cast steel requires a knowledge of thetransformations that occur with the heating and cooling that accompaniesfusion welding.of the various thermal cycles on the phases obtained in the iron-carbon system is an important considerat ion.thermal f luctuat ions exert on these phases during welding of a steel

containing 0.3% carbon is i l lustrated by the simplif ied iron-iron

Steel is basical ly an iron-carbon al loy so the effect

The influence that these


9/90

-9-

carbide diagram in Figure 3 (6) . Five poi nts or locat ions have beenselected to i l lustrate the various str uctures and condit ions that occuras a result of the maximum temperature at ta ined at each locat ion.

As shown in Figure 3, some of the base metal in the HA Z (Points 1-4 )is heated into the austenit ic or high temperature (face-centered cubic)

form of i ron by the fusion welding operat ion. As these areas cooldown to room temperature , this austenite t ransforms back to the roomtemperature phases with their body-centered structures. This t rans-formation can occur to form several microstructural forms such asferri te , pearl i te , baini te or martensi te . These const i tuents havedifferent characterist ics and their hardness general ly increases forthe phases formed at the lowest temperatures. Martensi te , the phasethat is formed at the lowest temperature , is both very hard and bri t t leat any except the lowest carbon contents. For this reason, the formationof martensi te in the HAZ can lead to cold cracks, hydrogen-inducedcracks or high residual stresses e i ther during or subsequent to welding.Since the formation of such cracks and stresses reduces the mechanical

propert ies, welding condit ions are control led for the different steelsproduced to avoid the formation of this martensi te . Those steels thatrequire special precautions to avoid martensi te forming in the HAZ areconsidered to have inferior weldabil i ty to those steels that do notrequire special handling.

The carbon content , cri t ical cooling rate and cri t ical temperaturesas affected by composit ion and the grain size of the steel are theprimary factors that determine the tendency of each steel fo form marten-si te . Increasing the carbon content decreases the cri t ical coolingrate , lowers the A3 temperature and increases the hardness of thetransformed region (4). The A1 and A3 temperature may be predicted

with fair accuracy from the al loy content such as by the expressionsl isted below that are useful for low-alloy steels containing up to0.60% carbon :

A l (F) = 1333 - 25 (%Mn) + 40(%Si) - 26(%Ni) + 42(%Cr)

A3(F) = 1570 - 323(%C) - 25(%Mn) + 80(%Si) - 32(%Ni) - 3(%Cr) .

The higher the steel is heated above the A3 temperature , the larger theaustenit ic grain size becomes; this effect is significant in thehighest temperature region of the HAZ (Point 1 in Figure 3) (4) .

On cooling from the austenit ic region, each steel has a cri t icalcooling rate though the region in which austenite t ransforms to thelower temperature microconsti tuents.not very useful in predict ing structures obtained under rapid coolingcondit ions, continuous cooling diagrams (CCT) have been constructedfor a large number of steels. These diagrams al low a correlat ionbetween cooling rates and the result ing microconsti tuents formed.The coolingrates are favored by high heat input from the welding operat ion, thinnersect ions and high preheating temperatures.a l loy steel (AISI 8630) is shown in Figure 4 (5) .

Since the equil ibrium diagram is

rate is determined by welding condit ions; slow cooling

A CCT diagram for a low


10/90

-10-


11/90

-11-


12/90

-12-

I t becom es evident f rom a study of this diagram that the differentcooling rates shown by curves A, B, C and D provide different f inalmicrostructures with different hardness levels . The cr i t ical coolingrate as descr ibed above is def ined as the maximum rate that will justprovide a completely martensit ic s tructure. This cr i t ical rate islocated just to the r ight of curve A in Figure 4. From a we ldingstandpoint, par t ial ly hardened or par tial ly martensit ic s tructures areu n d e s i r a b l e e v e n t h o u g h s m a l l a m o u n t s o f m a r t e n s i t e c a n b e t o l e r a t e d i nsome cases.

The phases obtained during transformation of the austenite atvar ious coolingratesor the CCT diagram for each steel var ies with alloyand carbon content. These CCT diagrams are available for a large numberof steels and allow this correlation between the cooling rate andstructure. Steels with carbon contents under 0.10% have extremelyrapid cr i t ical cooling rates and therefore present no problem withhardening even though the M s temperature increases with decreasingcarbon.

For a given steel , the f inal s tructure can be reasonably approx-imated by means of microhardness tests taken in the HA Z if correlatingtables are available.welding variables, cooling rates and hardness obtained for var ioussteels .cooling t ime through the transformation temperature region and sectionsize for var ious s teels and welding processes (7) .

Welding Processes

A large amount of work has been done to relate

Data are available to show the relation between heat input,

The pr imary welding processes that are used for the repair andfabrication of steel castings are l is ted below and i l lustrated by thesketches in Figure 5 (8) .

A. Shielded Metal-Arc Welding (SMAW) involves the use of aconsumable electrode shielded by a gas or iginating fromthe electrode coating or in some cases, the inner coreof the electrode.

Gas Metal Arc Welding (GMAW) , also known as Metal-Iner tGas Welding (MIG), the consumable electrode is shieldedby an iner t gas from a second source; such gases are

helium, argon (sometimes mixed with small amounts of hydrogen) and CO2 for cer tain s teels ,

Gas Tungsten Arc Welding (GTAW), also known as Tungsten-Inser t Gas Welding (TIG), a gas shielded tungstenelectrode is used to deposit metal f rom a consumablewelding rod.

The Submerged Arc Welding (SAW) process has an arc main-tained between a continuously fed consumable bare wireelectrode and the work. Flux is dispensed over the areato be welded and the wire is fed into and melts some of

this f lux.

B.

C.

D.


13/90

-13-


14/90

-14-

Other welding processes of some interest are gas welding with an oxy-acetylene torch used as a source of heat with a f i l ler rod to depositmetal; electro-slag welding; and non-shielded arc.

The shielded metal arc uses a wide var iety of coated electrodesthat are classif ied by the s trength level of the undil uted weld metal

and type of coating. These different coatings vary the penetration of the electrode, welding posit i ons and the amount of hydrogen in the gasshield. A detailed l is t of these electrode coatings is furnished subse-quently.on the electrode does not permit these to be coiled and fed automatically.An adaptation of this process places the coating or f lux consti tuentsinside of a tubular electrode and permits automatic use but the var ietyof f luxes and result ing atomospheres available are much more l imited.Both the GMAW and GTAW processes can be operated automatically orsemi-automatically. The GMAW (or MIG) process is used on steels withhelium, argon or mixed gas shielding and also with CO2 shielding.

CO2 gas is considerably cheaper than the others and is employed onsteels when the mild oxidizing effect of this atmosphere can be tolerated.The GTAW (or TIG) process is s lower than the other four methods l is tedand expensive because of the shielding gas.use is restr icted to cases where i t has a technical advantage such asthin sections. These MIG and TIG processes have the advantages of being automatic with continuous feeding of the electrode into the weldand avoiding the fused f lux or s lag layer produced by the SMAW andsubmerged arc (SAW) processes. The submerged arc process is alsoautomatic and is capable of substantially higher rates of metal . deposi-t ion and of making single pass welds of considerable thickness comparedto the other three methods.

The SMAW process cannot be made automatic because the coating

The

For these reasons, i ts

Gas welding is useful for small repair jobs and permits closecontrol of the temperatures. Carbon steels can be welded wit hout af lux but the rates of metal deposit ion is so s low that the process isnot widely used in repair ing or assembling steel castings, except forcladding and special purposes. The electroslag process is well adaptedto joining thick sections (over 2 inches) . I t uti l izes f i l ler wiresfed into a molten slag pool contained between water cooled dams. I ts tar ts as an arc but when the f lux melts , the process depends on theheat generated by the f low of current f rom the electrode through thef lux to the work. The process generates considerable heat, result ingin a coarse grained weld deposit and HAZ so that subsequent heat treat-

ments are generally required. While not used extensively in castingrepair up to this t ime, i t is widely employed for t he cast-weldassemblies with heavy sections (2).of infer ior quali ty that is only used for lower grade products .

Non-shielded arc provides a weld

Following the joining operation, a postweld heat treatment maybe needed, depending on the base metal composit ion, welding method andservice conditions. This treatment can range from stress relief of theweld area to quenching and tempering the entire weldment. Some of thepurposes include: s tress relief , improved strength and toughness,diffusion of hydrogen from the HAZ or restor ing the corrosion resis tance.


15/90

-15-

Types of Steels

For the purposes of this paper, the carbon steels wil l be consi-dered in one category and al loy steels with a total a l loy content of upto 5% in other groups.

addit ions of the elements Ni, Cr, Si , MO and Mn. These added elementsimprove mechanical propert ies, heat t reatment response, e levated temper-ature propert ies and corrosion resistance.

The al loy steels discussed contain intentional

Plain carbon cast steels re ly primari ly on carbon for variat ionof their propert ies. The carbon content ranges between 0.10 and 1.00%,depending on the intended use.when the carbon is below about 0.30%,steels decreases as the carbon content is raised. The weldability oflow-alloy steels is good, but they require careful a t tention to weldingprocedures and f i ller metal choice (5). As previous ly discussed , thestrength, hardness and ducti l i ty of these steels varies widely depending

on their composit ion and the thermal t reatments. Both carbon and al loysteels are employed in the heat t reated condit ion; the treatments usedvary widely from annealing to normalizing to quenching and tempering,depending on the f inal use of the cast ing.

They can be welded without diff icultyThe weldabil i ty of plain carbon

Low-Carbon Steels --

The low-carbon steels are considered to be those not exceedingThese are the most easi ly welded by a large variety of 0.20% carbon.

methods. Hardenabil i ty is very low and even though the maximum hardnessfor a ful ly-hardened structure may be considerable as i l lustrated inFigure 6 ( 4) , the actual hardness in low-carbon HAZs rarely reaches

high values because the cri t ical cooling rates are not achieved underordinary condit ions. Preheating is not needed; several weldingmethods can be employed, and cooling is rarely encountered. However,the low-carbon steels are susceptible to porosi ty unless these aredeoxidized.steels, many of the wrought steels are r immed and this can lead toporosi ty from CO gas evolution unless e lectrodes with considerableamounts of a luminum, manganese and si l icon are employed.

While i t is usual pract ice to deoxidize low-carbon cast

The low-carbon steels are welded commonly by the shielded metalarc , submerged arc and GMAW or MIG processes using CO 2 gas shield (9) .Addit ional deoxidizers are usually added to the electrode in this la t tercase. The choice of e lectrodes for welding low-carbon steel is basedon the desired f inal mechanical propert ies of the weld.the AWS classif icat ion and weld deposi t composit ions of typicalelectrodes used in the SMAW process for carbon and low-alloy steels ( 4) .The last two digi ts refer to the type of coating, while the precedingnumbers l ist the minimum tensi le strengths of the undiluted f i l lermetal .in Table II.

Table I l ists

Detai ls on the characterist ics of the various coatings are l isted

Electrodes of the EXXl0 and EXXll c lasses are sui table for low-carbon steels because of their deep penetrat ion characterist ic ,minimal slag production. Low-hydrogen coatings are general ly unnecessary


16/90

-16-


17/90

-17-


18/90

-18-


19/90

-19-


20/90

-20-


21/90

-21-


22/90

-22-

unless suff icient residual al loy elements are present to raise thecomposit ion to low allo y levels . The low-hydrogen coatings require care-ful handling and may pick up moisture when exposed to air for any signi-f icant length of t ime. I f these become moist , they should be oven dr iedand taken hot f rom the oven for usage. Therefore, the electrodes mostof ten used are the cellulosic electrodes at s trength levels of 60 ksi .

These electrodes give a good combination of mechanical properties andease of welding at al l posit ions (9) .

Most problems with low carbon steels are caused by impurityelements and non-homogeneous structures. Sulfur is the most commonand damaging impurity element. Sulfur levels at 0.05% can produceliquid sulf ide f i lms at the grain bo undaries (10) that cause solidembrit t l ing fi lms and possible hot cracking, al though this is a muchmore ser ious problem in the higher carbon and alloy grades.sence of s trong sulf ide formers, such as manganese, reduce this pro-blem but may not completely eliminate i t ( l l) .manganese sulf ide inclusions also reduces problems with hydrogen

embrit t lement by providing sinks for this element (12, 13) .more stable sulf ide formers, such as rare ear ths, can be more effectivein this respect. Phosphorus has a s imilar embrit t l ing effect to sulfur(4, 14) and is more diff icult to remove.

The pre-

The presence of

Other

These problems with hot cracking caused by low melting point f i lmsand embrit t l ing elements apply to the entire range of carbon and alloysteels . In fact , sulfur and phosphorus levels are much more cr i t icalin higher performance steels , and maximum allowable contents aregenerally much lower than in l ow-carbon steels . The higher toleranceof low-carbon steels for impurity elements exists because the segregationof sulfur and phosphorus are increased by higher carbon contents and

certain alloy elements . The only low-carbon steels which require someprecautions are the free-machining steels , where sulfur levels mayreach 0.30%.steels , welding with cellulosic electrodes or other conditions producinghydrogen may result in porosity, due to the formation of H 2S gas in theweld metal.used to avoid this problem.

Even though addit ional manganese is specif ied in these

Electrodes of the EXX15, EXX1 6 and EXX18 classes may be

Low-carbon steels are not highly susceptible to cold crackingeven under fair ly severe conditions; the lack of toughness character is t icof higher-carbon martensites is not found in these cast s teels under0.20% C.

Medium Carbon Steels - -

Medium carbon steels are generally considered to be plain carbonsteels with 0.20-0.50% carbon.is required is generally not necessary below 0.30% carbon.suggested composit ion l imits above which preheating is required are0.28% carbon and 1% Mn according to one source (14) . As the carbonincreases above 0.30%, preheating and the use of low-hydrogen electrodesbecome more of a necessity to avoid cracking problems. Towards thetop of this range, the HAZ often attains full hardness wit hout addi-

Hardening of the HAZ so that preheatingThe


23/90

-23-

t ional a l loying and reaches the maximum hardness shown in Figure 6.This high hardness with a low ducti l i ty or bri t t leness make the hard-ened structures very susceptible to both cold-cracking from thermalstresses and to hydrogen-induced cracking. Figure 7 (4) shows under-bead crack sensi t ivi ty, underbead hardness and bend angle at maximumload as a function of carbon equivalent . This carbon equivalent has

been formulated to al low for the influence of the manganese and si l iconcontents on the hardening behavior of these medium carbon steels.is evident that both cracking susceptibi l i ty and hardness increase rapid-ly with carbon equivalent and that the bend angle at fa i lure decreases.All these trends denote a marked loss in toughness.

I t

Close control of cooling rates in the HAZ must be maintained.The preheat temperature required depends on the carbon content , sect ionthickness and number of weld passes.for thin sect ions at 0.35% C to 400F for 0.45% C with 4 inch sect ionsand single pass welds. The preheat requirements for mult ipass weldsare substantia l ly less i f the welding operat ions are conducted effi-

c iently. Preheating temperatures of 400F are needed with 0.50% Csteels welded under other than low-hydrogen condit ions and temperaturesof 500F are used in the upper part of the 0.30-0.50% C range forheavy sect ions, single pass welds and other than low-hydrogen condi-t ions (1,4). Interpass temperatures a t least equaling the preheattemperature should be maintained.

These temperatures vary from 100F

High Carbon Steels - -

High carbon steels are c lassif ied as containing between 0.50 and1.00% carbon.weld area from the high hardness and bri t t leness that can resul t from

martensi te formation. Precise determination of the necessary preheatand interpass temperatures is diff icult for steels with carbon contentsover about 0.50% C because of their sensi t ivi ty to other factors.Attempts have been made to correlate hardness and crack sensi t ivi tydirect ly, but the results indicate that such a relat ionship cannot beisolated (4). The fol lowing welding procedure is recommended: low-hydrogen shielding when available in desired strength levels, awelding technique which minimizes di lut ion and the resultant hardeningof the weld metal and a postweld heat t reatment consist ing of a tleast a stress re l ief and possibly a ful l anneal (1) . The choice of afi l ler metal depends on the desired f inal strength of t he joint .ments used in the quenched-and-tempered condit ion require E90XX or El00XX

electrodes; lower strength f i l ler metals may be used in other cases.

Preheating should always be used, the exact temperature being

They are diff icult to weld because of cracking in the

Weld-

determined by cooling condit ions and welding method. Temperatures of 600F or even higher are sometimes employed, but such temperaturescause diff icult ies with oxidation and welder discomfort . The preheattemperature must be adjusted in such cases to a l low the diffus iblehydrogen to leave the hardened area and prevent addit ional embri t t le-ment. In addit ion to the preheat , welding techniques that provideslow cooling should be used including a high heat input , mult ipasswelds and the use of insulat ion to restr ic t heat f low.


24/90

-24-


25/90

-25-

Alloy Steels - -

Classif ication of the weldabil i ty of al loy steels by composit ionis diff icult s ince many di fferent combinations and purposes of al loyingare employed.

toughness and corrosion resis tance. The inf luence of alloying onweldabil i ty has been equated to carbon content by a term known ascarbon equivalent.carbon steel in Figure 7. A fair ly widely used relation is shownbelow(8) .

Alloys are added to s teels to improve the s trength,

This approach was used for the other elements in

Welding Carbon Equivalent = %C + % Mn + % Ni + (% Cr +%Mo + %V )20 15 10

As this value increases, the susceptibll i ty of the s teel to cold-cracking

____ ___ _____________

does l ikewise. The carbon equivalent is designed to be employed alongwith a knowledge of the electrode size and weld dimensions as a means

of selecting preheating temperatures. While this approach has theadvantage of s implicity, i t i s l imited in range and is an oversimpli-f ication of the s i tuation.

Even within the l imits of 5% total al loy content employed inthis paper , a considerable var iety of al loy steel castings are included.These have a wide range of weldabil i ty and are uti l ized for many pur-poses. Considerable detail on their repair welding is presented inreference 1, and i t is not the purpose of this publication to duplicatethat coverage. A summary of the electrodes used, preheat and post-heating temperatures recommended are included in Table III (15) .table provides l imited data on the carbon steels already discussed,

as well as the var ious categories of low and medium alloy steels .

This

The majority of the alloying elements employed increase thehardenabil i ty of the s teel and reduce the cr i t ical cool ing rate.the more highly-alloyed steels , full hardening is possible even inthick sections and at modest cooling rates . With this increase inhardenabil i ty, the carbon content becomes increasingly cr i t icalbecause i t controls the hardness of the martensite and affects i tstoughness. The relation between maximum hardness and carbon content,as shown in Figure 6, becomes a realis t ic estimate of the hardness of martensit ic al loy steels because fully-hardened structures can readilybe obtained. A carbon content of 0.20% in a plain carbon steel might

produce a hardness of R c 30 under relatively rapid cooling conditions,whereas the same carbon level in a high-alloy steel at the same coolingrate might result in a hardness of R c 40-45.markedly increase the susceptibil i ty to cold-cracking and generallyrequire a tempering treatment before use in service.

In addit ion to increasing hardenabil i ty, most al loy elements

In

Such hardness levels

except for s i l icon, cobalt and aluminum, lower the M s temperature of the s teel . Sil icon has no effect and aluminum and cobalt raise the M s .An expression for calculating the M s temperature for a s teel of knowncomposit ion as shown below (4):

Ms (F) = 1042 - 853(%C) - 60(%Mn) - 30(%Ni) - 30(%Cr) - 38(%Mo)


26/90

-26-


27/90

-27-

The hardenabil i ty is affected by grain size and segregation as well ascomposit ion and requires special calculat ions that are not presentedin this paper.

Alloying elements increase the susceptibi l i ty of steels towelding discontinuit ies, part icularly cracking and the metal lurgicaltypes. The susceptibi l i ty to hot cracking is st i l l largely control ledby the sulfur content of the steel but a l loy steels are more sensi t iveto a given sulfur level . While a 0.05% S level is permissible in manycarbon steels, a much lower sulfur content can cause serious crackingproblems in an al loy steel . The effects of a l loys on hot crackingsusceptibi l i ty depend primari ly on their influence on the behavior of

by reducing the solubil i ty of sulfur a lso promote hot-cracking. Suchelements include carbon, phosphorus, nickel , a luminum and large amountsof sili con (over 1% and only in the heat-treatable alloys). As men-t ioned above, manganese tends to form fair ly stable sulf ides which

resist grain boundary segregation and therefore decrease hot-crackingsusceptibi l i ty .are chromium, vanadium and molybdenum. These elements exert theirinfluence indirect ly by forming stable carbides and thus decreasing theeffect of carbon. While the general influence of a l l these elementsis known individually, i t has not been possible to predict their com-bined effects. For this reason, the cracking tendency of each part icularsteel is evaluated by means of standard weldabil i ty tests.

sulfur. Those that promote sulf ide segregation to grain boundaries

Other e lements which reduce hot-cracking tendencies

The presence of significant amounts of a l loys increases the sus-ceptibi l i ty of the steels to both cold-cracking and hydrogen-inducedcracking.

which showless toughness in the

as-weldedcondit ion than carbon steels

of similar carbon content . The increased cracking results from anumber of factors:in increased stored elast ic energy; the lower M s and M f temperatures

atures which cannot act to temper the martensi te , stress re l ieve, orproduce significant diffusion of hydrogen from the hardened area; andthe more complete t ransformation to martensi te , which has a lowersolubil i ty for hydrogen than austenite or ferri te , results in a greaterdegree of supersaturat ion of hydrogen. I t is evident that great caremust be observed in welding al loy steels to prevent cold-cracking andonly low-hydrogen processes be used when possible .

The higher hardenabil i ty in a l loy steels results in structures

higher yield strengths obtained by al loying result

of a l loy steels mean that t ransformation often takes place at temper-

The choice of f i l ler metal for a l loy steels is more cri t ical inal loy than carbon steels. Whereas low or medium carbon steels couldbe welded with a number of filler materials, alloy steels frequentlyhave propert ies that can only be obtained by a l imited number of weldmetals. The select ion is further complicated by the fact that matchingof composit ions is often impossible or undesirable . The f i l ler metalrequirements must be examined from the standpoint of weld metal propert iesin the f inal structure rather than composit ion and propert ies of f i l lermetal a lone. Increased al loy content of f i l ler metal produces a pro-blem with oxidation losses and the weld metal composition is greatlyaffected by pickup of a l loy elements f rom the base metal .


28/90

-28-

In addit ion to the general coverage provided by Table III andthe above discussions, some individual considerations are the heattreated condition in which the casting i s received and the subsequentheat treatment that is to be conducted. In cases where f ield weldabil i tywithout preheating or postheating is desired, the alloy ste els have alow carbon content (generally below 0.20%) and depend on multiple Mn-Cr-Mo-Ni Alloys for their s trength. The purpose is to maintain a lowcarbon equivalent value. The low carbon level provides a much tougher ,crack-resis tant martensite when this is formed.level and the section thickness, the amount of al loying and heat treat-ment employed pr ior to welding is adjusted. These steels are welded inthe normalized, normalized and tempered, and quenched and temperedconditions. When postheating procedures are not feasible, i t is therecommended welding practice to weld with low hydrogen electrodes andto provide fast cooling rates in the HAZ during welding . This techniquedeposits small weld beads at low heat input in a s tr inger fashion withno weaving of the electrode.

self- tempering effect . The strength levels of these f ield welded steelsare, of course, l imited because the amount of hardening elements isrestr icted by the requirements for a relatively low carbon equivalent.

Depending on the s trength

The multiple weld deposits provide a

When the application requires higher s trength or involves heaviersections, the castings are quenched and tempered af ter welding. However ,in cases where the quenched and tempered weld metal fails to reach therequired strength, only tempering can be used. These castings are fre-quently annealed pr ior to welding and require preheating temperaturesas high as 600F. Since these s teels are to be heat treated af terwelding, high heat inputs during welding are used to retard martensiteformation.

cases to provide the needed properties in the f inished component.Cracking of this deposited weld then becomes a more severe problemthan the HAZ of the base casting making the use of the higher preheatingtemperatures a necessity (8) .

III. MECHANICAL PROPERTIES OF WELDS

The deposited weld also is usually high strength in t hese

In addit ion to the s trength and ducti l i ty requirements , thedynamic properties of repair and fabricated welds in s teel castingsare of major s ignif icance in the service behavior of the f inishedcomponents . Because welds present an altered dimensional and metallur-gical s tructure, as discussed in the previous section, the properties

of welds are frequently considerably different than those of the base,heat treated casting.metallurgical discontinuit ies that have been descr ibed since the weldshave both a different s tructure and composit ion compared to the basemetal . The presence of the other types of weld discontinui t ies(Figure 1) also exert considerable inf luence on mechanical behavior .This par t in the paper discusses the effect of numerous welding pro-cesses, var iables and discontinuit ies on the fatigue and toughnessproperties of welds in s teel castings.

Many of these effects are produced by the


29/90

-29-

Fatigue Behavior

Weld Configurat ion --

The weld geometry or configurat ion is a major considerat ion in

weldments joining wrought sect ions and shapes but these considerat ionsdo not apply for the most part to welds in steel cast ings. When repairor fabricat ion welds are made on steel cast ings and ground flush to thesmooth contour of the part , the effect of weld geometry is removedunless undercuts have occurred during welding.welds are employed to make "L", "T ", " X" or box sect ions, and the weldsare not ground smooth, the weld geometry can i nfluence fat igue strengthmarkedly.of the fat igue propert ies of cast "L" and box sect ions compared tothose shapes on welded joints in medium carbon steels (16).blem of the fat igue behavior of weld configurat ions in f at igue and thedesirabil i ty of using steel cast ings or forgings for some contours has

been recognized in other texts (17).

However, when fi l le t

A report has been published by SFSA s howing the superiori ty

The pro-

The effect of not grinding the weld reinforcement or ra isedweld deposi t from a butt weld is i l lustrated in Figure 8 (17). Thisfigure also indicates that i f the weld reinforcement is removed com-pletely by proper grinding or machining, the fat igue strength of thejoint starts to approach the propert ies for the as-received base plate .Fatigue tests conducted on simulated butt-welded specimens ofmildcarbon structural wrought steel with the simulated weld reinforcementmachined from base metal showed similar results (18).variables are the f lank angle , 8 , and the radius at the toe of theweld, R as shown in Figure 9a.

the radius wil l increase fat igue strength as i l lustrated in Figure 9b (18) .Other weld configurat ions, such as f i l le t welds, are a lso influencedby the shape of the weld reinforcement, the overal l dimensions of theweld and the presence of small undercuts a t the toe of the weld (16,17).The fat igue propert ies of welds and lap joints are even more affectedby geometrical considerat ions but these shapes usually do not apply tosteel cast ings.

The significant

Decreasing the f lank angle or increasing

Base Metal Strength Level --

The unnotched fat igue strength of carbon, low and medium alloysteel cast ings increases with tensi le strength, a t least up to a tensi le

strength of about 200 ksi .to level off a t a lower strength (160-180 ksi U.T.S.) .is i l lustrated in Figure 10 (15,19). The notched fat igue test resultsshow more scat ter than the unnotched fat igue s trengths and are usuallyconsidered to be more applicable to the service performance of mostcomponents.

The notched fat igue strength, however, beginsThis behavior

Welding may reduce the fat igue strength of the higher strengthsteels even more than the presence of mechanical notches.the welding l i terature (20-23) show that :

Data in


30/90

-30-


31/90

-31-


32/90

-32-

(1) for steels with weld strengths of 55 to 110 ksi , weldfat igue strength increases sl ightly with increasingtensi le strength;

the fat igue strength of higher strength steels levelsoff because of the increasing notch sensi t ivi ty at higher

tensi le strengths;considerable scat ter occurs in the data result ing inreports that the fat igue strength is not affected bytensi le strength.

This effect of the tensi le strength of the base material welded

In this and in subsequent plots the R value refers

(2)

(3)

on the fat igue strength of t ransverse butt steel welds is i l lustratedin Figure 11 (20).to the algebraic rat io of the smallest and largest stress ( tension +,compression -) during fat igue test ing. In this case, the test ingcycle was zero stress-tension.

An explanation has been presented for the fact that the fat iguestrength of weldments does not show the same degree of dependence ontensi le strength exhibited by the notched fat igue strength of thebase metal (24). I t was reported that fa t igue cracks which form in theregion of stress concentrat ion at the toe of the weld actual ly ini t ia teat non-metal l ic inclusions which further concentrate the stress. Thishigh stress concentrat ion makes the stress a t the inclusion t ip veryhigh and the strain independent of materia l strength. Since fat igueis stra in dependent , i t too becomes independent of materia l strength.

High and Low Hydrogen Electrodes - -

The earl ier invest igators of this subject concluded that the useof low-hydrogen electrodes resulted in no advantage in fat igue strengthover that for high hydrogen rut i le e lectrodes.was sta ted that no significant difference exists between the strengthof butt welds made with rut i le e lectrodes and welds made with low-hydrogen electrodes provided that they both have the s ame reinforcementshape.

In one case (25), i t

However, when the weld reinforcement is removed, the advantagesof low hydrogen welds become evident .the fat igue crack ini t ia tes a t the toe of the weld in the HAZ of the

base plate . When the reinforcement is removed, fa t igue cracksini t ia te a t discontinuit ies in the weld and the propert ies of the weldmaterial become significant .fa t igue specimens machined from welds indicate the superiori ty of lowhydrogen weld metal (26).reinforcement was machined off , and a hole was dri l led through theweld to produce a stress concentrat ion within the weld metal (27).When these specimens were tested in pulsat ing tension, welds made withlow hydrogen electrodes displayed significantly higher fat igue strengths.Similar results are a lso available for machined butt welds with notches0.79 inches deep and a 0. l0 inch radius machined in the edge of a

With the reinforcement intact ,

Fat igue tests made on al l weld metal

In addit ional tests , t he weld overfi l l or


33/90

-33-

3 inch wide, 24 inch long specimen as i l lustrated in Figure 12 (26).Other data on the s uperiori ty of welds made with low hydrogen electrodesin the presence of weld discontinuit ies wil l be discussed in thefollowing sect ion.

Weld Discontinuit ies --

An extensive invest igat ion (28) has been undertaken by the SFSAto determine the influence of welding discontinuit ies on the fat iguestrength of welded cast steel in bending and torsion. The tests wereal l performed on low al loy (8630) steel that was normalized andtempered to a tensi le strength level of about 88 ksi and quenched andtempered to approximately 122 psi . The dimensions of the fat iguespecimens employed are shown in Figure 13 (28).

These specimens were welded in the as-cast condit ion, and thenormalizing and tempering and water quench and tempering treatments

were performed (austenit iz ing in sal t pots) after welding was completed.Accordingly, these results show the effects of the weld discontinuit iesonly, and are not influenced by the heat effect that occurs duringwelding.using the shielded metal arc welding process and commercial E9018-B3,low hydrogen type electrodes.to produce slow cooling and post-heat of l l00F for one hour immediatelyafter welding to prevent cold cracking.

Double vee butt welds were employed on each type of specimen

A preheat temperature of 300F was used

The weld joint , number and order of weld bead deposi t ion andthe types of discontinuit ies tested are shown in Figure 14 (28).c lassif icat ion of the welding discontinuit ies that were tested are

l isted in Table IV (28).summarized for bending in Table Va and for torsion in Table Vb. Ineach case, the test results are compared with a sound cast st eel of similar heat t reatment and strength level .

The

The results of the fat igue tests are

This invest igat ion (28) conducted complete S-N curves on thesesteels, so the values of K f* were obtained for the various types of discontinuit ies on the normalized and tempered and quenched andtempered steels for both bending and torsion fat igue tests.values have been l isted in Table VIa for bending and Table VIb (28)for torsion at l0 5 , l 0 6 , and l07 cycles,

These K f

The test results indicate that undercuts have the most markedeffects on fat igue life , fol lowed by slag inclusions and incompletepenetrat ion with the least effect from the sound and sound weld testsmachined f lush to the fat igue bar surface.the approximate percentage loss in endurance l imit from the variousdiscontinuit ies for bending and torsion fat igue (28, 28a).

____________

The figures below indicate

*(K f) = Endurance Limit of Unnotched SpecimensEndurance Limit of Notched Specimens______________________________________


34/90

-34-


35/90

-35-


36/90

-36-


37/90

-37-


38/90

-38-

For Bending Fatigue - At Two Tensile Strength Levels

DISCONTINUITYAPPROX. % LOSS IN ENDURANCE LIMIT

Q & T-122 ksi N & T-88 ksi

Weld - Undercut

Weld - Incomplete Penetrat ionWeld - Slag InclusionsWeld - Sound - Not MachinedWeld - Sound Machined

For Torsion Fatigue - At Two Tensile Strength Levels

31

22212019

22

613

52

DISCONTINUITYAPPROX. % LOSS IN ENDURANCE LIMIT

Q & T-122 ksi N & T-88 ksi

Weld - UndercutWeld - Slag Inclusions, Severe

Weld - Sound - Not MachinedWeld - Lack of Penetrat ionWeld - Sound Machined

3630

201515

1513

810

3

The loss in fat igue propert ies from the s ound weld occursbecause of the different composit ions and the geometric effect of theweld reinforcement previously discussed. The sound-machined specimenloses fat igue strength only because of the difference in weld and basecast ing propert ies and composit ion. These specimens were heat treatedafter welding so any heat effects would be removed, The effect of thedifference in composit ion is shown by the traverse of hardness readingstaken across the cast steel weld deposi t* area in Figures 15a and

15b (28).

Extensive experimental work has been performed on the effectthat slag inclusions in welds have on weld fat igue strength (26, 29-31)Transverse butt welds were made on mild steel plate with both rut i leand low hydrogen electrodes containing small discrete slag defects.The steel plate had a tensi le strength of about 64.3 ksi . The decreasein fat igue strength with increasing totaldefect length and depth have been held constant , and the higher fat iguestrength in low hydrogen welds in 1/2 inch thick plates are presentedin Figures 16a and 16b (26, 29-31).on two pass but welds machined f lush with the surface before test ing.

inclusion length, where the

These fat igue tests were conducted

Addit ional tests were conducted with 1-1/2 inch thick butt welds.

The influence of a 300F preheat andThe heavier sect ion resulted in significant residual str esses andproblems with hydrogen (7, 31).a 1200F stress re l ief for 1-1/2 hours were studied. The tests wereconducted on a mild steel plate with a tensi le strength of 63 .4 ksi .The fol lowing conclusions resulted from these tests (31).

_______________________

*Cast 8630 type steel , typical ly .30 C, .80 Mn, .30 Si , .45 Cr, .55 Ni,.20 Mo, SMAW, low hydrogen electrode E9018-B3 with typical . l0 C, .90 Mn,.80 Si , .15 Cr, 1 .60 Ni, .35 Mo, .05 V


39/90

-39-


40/90

-40-

(1) Discrete s lag inclusions are more deleter ious near therolled plate surface because the surface is in residualtension.

When the effect of discontinuit ies central in thicknesswere blanketed by compressive residual s tress , large

and small s lag par ticles produced similar s trengths.When these compressive stresses were relieved pr ior totesting, the larger inclusions resulted in a lower s trength.

For discontinuit ies central in t hickness, s tress relief produces an increase in s trength where the s lag is discrete,but a decrease with a continuous slag l ine. The reductionin strength, result ing from stress relief for a contin uousdefect central in thickness, occurs because the compressivestresses at the defect are removed by the treatment. Theimprovement in s trength for discrete defects is that s tressrelief removes hydrogen.

Stress relieving raises the fatigue strength of ruti leweld specimens with discrete s lag inclusions at theircenter or at their surface to a value approximately thesame as that for low hydrogen weld specimens with centrallylocated discrete defects because of hydrogen removal. How-ever , preheat had l i t t le effect on this low strength, mildcarbon steel .

The fatigue strength increased with R, the ratio of theminimum stress to the maximum stress in each stress cycle.

(2)

(3)

(4)

(5)

Lack of penetration is another weld discontinuity that reducesfatigue properties .on transverse butt welds with a lack of central penetration (32) .specimens were made from a 1/2 inch thick, 4 inch wide hot rolled andnormalized mild s teel plate with the analysis and mechanical propertiesshown below. Rutile electrodes were used and none of the specimenswere pre- or post-heated. The results indicate that the fatiguestrength at 2 x l0 6 cycles endurance decreases with longer lack of penetration.the lack of penetration (% area defective) and the reduction of fatiguestrength (%).23 ksi with good penetration to 5,000 psi with 60% lack of penetration(l7) .

Other work (33, 34) indicates even higher losses in fatigue strengthfrom incomplete penetration. Reductions in endurance l imit of up to40% can occur with a 15% lack of penetration, and this can increase toover a 50% loss in fatigue strength with 30% incomplete penetration.

Pulsating tension fatigue tests were performedThe

An approximate direct or l inear relation exists between

The fatigue strength at 2 x l0 6 cycles was reduced from

The large fatigue strength loss in the transverse butt weldsoccurs because the applied stress is transverse (normal) to the par tialpenetration.maximum principal s tress , however , i t has a negligible effect onfatigue strength. Tests on f i l let welds indicate l i t t le difference infatigue resis tance between full and partial penetration longitudinalf i l let welds.

When the incomplete penetration l ies parallel to the


41/90

-41-

Gas porosi ty in welds can also lower fat igue strength; theeffect on low cycle fat igue strength is minor but a t higher cycles,much greater reductions in fat igue strength occurs (35). The data in di-cate that 4-1/2% porosi ty in a weld can lower the endurance l imit from25 to 45%.at the surface in bending and torsion.

porosi ty on the fat igue strength at 2 x l0 6 cycles is shown fromseveral sources (17) in Figure 17.

The more severe losses occur when the porosi ty is presentA summary of the effects of

The marked effect of undercuts on the fat igue strength observedfor welds in cast steel (28) has also been noted in other weldments(33, 36).The fat igue fracture of sound specimens involving this type of defectwas always observed to originate a t the root of the undercut . The

undercut increases.58.3 psi , a reduction of the fat igue l imit was observed from 26.3 ksito 12.8 ksi when the depth of the undercut increased from 0 to 0.35inches, i .e . , a reduction rat io of about 51% (36).

fat igue strength at 2 x 106 cycles decreases when the depth of theFor a steel plate with a tensi le strength of

The effect of weld cracks on the fat igue strength of weldmentscan be expected to be marked because of the sharp stress concentrat ionthat results from this type of discontinuity. One invest igat ionreport ing on the influence of cracks was conducted on transverse buttwelds containing cracks paral le l to the weld direct ion and transverseto the applied stress.sect ional area of the specimen produced considerable scat ter of results;however, on the average the fat igue strength at 2 x l0 6 cycles wasreduced from about 14,000 psi without cracks to only 35 to 45% of thatvalue with cracks (17).

Cracks penetrat ing about 10% of the cross

A broad-based invest igat ion on the influence of a number of welding discontinuit ies on the fat igue behavior of welds employed byboth Wohler and program tests* was conducted (37).the comparat ive effect of a number of discontinuit ies. These are shownbelow according to the relat ive severi ty with the more severe l is tedfirst :and restart .importance.

This work i l lustrated

cracks, undercut , lack of fusion or penetrat ion, slag inclusionsThe last discontinuity is only considered to be of minor

Results from conventional fa t igue tests and from program tests,which employ a sta t ist ical ly varying load, are displayed in tabularform as Table VII and graphical form in Figure 18 (37). This study

also invest igated the effect of voids and found these to be of minorsignificance.incomplete data are presented in this la t ter work (37).

This is a t odds with the other data on porosi ty and only

Welding Processes --

The effect of welding processes on the fat igue strength of weldscan be influenced considerably by the shape of the weld reinforcement

*fat igue tests involving stress ampli tudes that vary in accordance withstat ist ical ly determined service condit ions


42/90

-42-


43/90

-43-


44/90

-44-

produced by that process.welding processes (17) conducted on transverse butt welds tested inpulsat ing tension concluded that automatic welds provided considerablypoorer fa t igue propert ies than those made manually.difference resulted from the unfavorable reinforcement shape, as dis-cussed under weld configurat ion, ra ther than any metal lurgical difference.When the weld reinforcement was removed from automatic welds made by the

submerged arc process, the fat igue strength was similar to manual weldsmade by the shielded metal arc method (17).rel ieved to el iminate the effect of residual stress.

One invest igat ion of the influence of

However, this

These tests were stress

Another invest igat ion (20) concluded, based on a study oftwenty-five references, that submerged arc and electroslag weldingconducted automatical ly provided superior fa t igue strength than manualwelds. This improved fat igue strength was at tr ibuted to the fewerinternal discontinuit ies and smoother weld surface obtained with theautomatic processes compared to manual methods. This assumes that anunfavorable weld shape is not obtained in the automatic method. Thisimproved fat igue strength of the automatic processes was also obs erved

for the semi-automatic MIG process with CO2 shielding and flux coredarc welding to a somewhat reduced extent . The electroslag processoffers the potentia l of a considerable r eduction in discontinuit ies withthe result ing bet ter fa t igue behavior (17, 20).discontinuit ies can vary so widely, depending on the welding condit ions,that sta t ing specif ic values of fa t igue strengths with the variousprocesses is not feasible .

The differences in

Residua l St ress - -

When a weld cools, i t is restrained from contract ing by therelat ively cool base metal .

tension and the underlying cast ing to be in residual compression. Astress rel ief t reatment re l ieves the residual tensi le stress es a t thetoe of the weld and would be expected to increase the fat igue strength.However, i t has been reported that residual stress, and hence a stressrel ief , has l i t t le effect on fat igue strength.mild steel butt welded plate was fat igue tested in pulsat ing tension withand without a stress re l ieving treatment a t 1200F. The data indicatedthat the stress re l ieving had no effect on the fat igue behavior.However, the notched fat igue strength of both a mild and medium carbonsteel was reduced by the presence of a residual tensi le stress whenthe fat igue tests were performed with a completely reversed stresscycle (39).

This causes the weld to be in residual

In one study (38),

The relat ive insensi t ivi ty of weld fat igue strength to residualtension when the weld is tested in pulsat ing tension and the loss of fa t igue strength because of r esidual tension when the weld is testedin a reversed stress cycle has been i l lustrated in other invest igat ions(40) .The influence of residual tension on weld fat igue strength is dependenton the stress ra t io. The stress re l ief t reatment improves the fat iguestrength at 2 x l0 6 cycles as the rat io R becomes more negative or asthe value of the compression stress increases. I t is a general rulethat no increase in fat igue strength is obtained by stress re l ievingstructures that are subject to purely tensi le loads but an improvementin fat igue strength can be obtained if the loads are al ternat ing


45/90

-45-

tension and compression.

Post Weld Treatments - -

Several techniques are ut i l ized to increase the fat igue strengthThese include spot heat ing, local compression, local heat ingof welds.

fol lowed by rapid quenching, prior overloading and peening (40).Grinding of the weld to reduce the stress concentrat ion of weld con-figurat ion is a lso a very effect ive method. The applicabil i ty of mostof these methods to repair or even fabricate welds in cast ings isl imited. Spot heat ing and local compression are l imited to weldsproducing local ized notches such as spot welds or longitudinal weldends (40) and are not useful for t ransverse welds. Local heat ingfollowed by quenching and prior overloading also are techniquesl imited to small weld areas and requiring special equipment; these arehardly feasible for use with cast ings except in unusual cases (41-43).

Peening of the weld surface with an air hammer is a technique thatis adaptable to the repair and fabricated welds in cast ings. Theimprovement in the fat igue behavior obtained by peening is i l lustratedin Figure 19 (41).15.68 ksi and 26.88 ksi for the as-welded and peened specimens, sohammer peening increased the fat igue strength 70%.after peening removes most of the beneficial effect of peening.work shows similar improvements in fat igue strength (44) .

At 2 x l0 6 cycles, the fat igue strengths are

Stress rel ievingOther

Grinding the weld reinforcement f lush to the weld surfaceimproves the fat igue strength of welds, as discussed under weld con-figurat ions. When welds in longitudinal gussets were ful ly ground, thefat igue strength of mild steel f i l le t welds was improved from 50 to100% (44). The effects of peening and grinding on the fat igue strengthof t ransverse butt welds in a low carbon al loy steel with a tensi lestrength of72 ksi as welded is i l lustrated below in Figure 20 (45) .Grinding increased the endurance l imit of the welded specimen by 20%,peening by 51%, and grinding plus peening by 56%. Grinding did notcause a greater increase in the endurance limit because the ini t ia langle between the weld reinforcement and the base plate was not verysharp original ly. However, the ground specimens st i l l display agreater fa t igue strength at shorter l i f e than do the peened specimens(45).

Toughness of Welds

The toughness or impact resistance of welds in steel cast ings isof considerable significance for many dynamically loaded applicat ions.The relat ive toughness of the cast steel or base metal is affected bymany variables including: the chemical analysis, microstructuralconst i tuents, strength or hardness level and grain size . Optimumtoughness is a t ta ined in cast steels (a l though not necessari ly inwelds) by a f ine grained ful ly hardened and tempered structure witha low sulfur and phosphorus content and a carbon content below 0.36%.Considerable scat ter occurs a t any strength level because of thevariat ions in these factors, but the toughness is general ly reducedby higher strengths, as indicated by the Charpy V-notch values for


46/90

-46-


47/90

-47-

tests a t 70 and -40F in Figure 21 (19). Even though these are al lquenched and tempered al loy steel cast ings, the scat ter in values atany strength level is very significant ,

Welding exerts a significant influence on the toughness, part i-

cularly in cases where the cast ings are not completely reheat t reatedafter welding. The variat ions in temperature cycles and propert iesdemonstrated in Figure 3 produces a wide range of toughness values inthe HAZ.rel ieving and further reduced by reheat t reatment.

The influence of these variat ions is reduced by stress

However, the fol lowing factors a lso affect toughness: thedifferent welding processes such as:(manual-metal arc) , metal- inert gas welding, f lux-cored wire welding,submerged arc welding and electroslag welding; and the different featuresof each process including:posi t ion, heat t reatment and i ts effect on HAZ microstructure , restraint ,

discontinuit ies and composit ion.

shielded-metal arc welding

type of e lectrode, heat input , welding

Four possible paths exist in a welded structure by which abri t t le fracture could progagate . These paths shown in Figure 22 ( 46)include the parent plate , heat-affected zone, fusion l ine and the welddeposit .e l iminated as a possibi l i ty . Each region has i ts own characterist icfracture progagation temperature . Therefore, once fracture has beeninit ia ted, propagation occurs in the zone of lowest notch ducti l i ty .Fracture can ini t ia te a t the various types of welding discontinuit iesdiscussed in the previous sect ion of this paper.

In a steel of high notch ducti l i ty , the plate path is

Types of Tests for Toughness - -

Several types of tests may be employed to measure the toughnessof weldments. These include: Charpy impact tests , general ly V-notch;drop weight tests; dynamic tear tests; and fracture toughness testswith the associated crack opening dis placement (COD) measurements.The Charpy V-notch bar has been employed for longer periods and more gen-eral ly than the other types, so most of the data wil l refer to this.The Charpy test measurements are usually conducted over a range oftemperatures to establish a bri t t le-ducti le t ransi t ion. The resul tsare measured in terms of:(ducti le or f ibrous to bri t t le or crystal l ine) and the transi t ion

temperature from the bri t t le to ducti le fracture .require a minimum fracture energy, such as 15 f t . lbs at a giventemperature .

fracture energy, fracture appearance

Many specif icat ions

The Charpy test does have some drawbacks. The test resultsare primari ly quali ta t ive and used for re lat ive measurements of tough-ness rather than being direct ly applicable to service condit ions. Thetest fa i ls to differentia te between fracture ini t ia t ion and propagation,al though this drawback has been overcome in the instrumented Charpytest . The test is small and convenient to use but may not be applicableto larger sect ions.


48/90

-48-


49/90

-49-

The dynamic tear test is a drop weight or pendulum test con-ducted on t he width of a rectangular bar . This tes t s imulates theworst condition anticipated in a s tructure and measures the resis tanceto crack propagation.br it t le-ducti le transit ion when conducted over a range of temperatures

than the Charpy V-notch test (47) .

I t provides a much sharper indication of

The drop weight test measures the abil i ty of the s teel todeform in the presence of a very sharp notch.abrasive saw cut made in a br i t t le crack star ter weld, the results areaffected by the weld metal hardness and toughness of the HAZ under thestar ter weld.to obtain the nil ducti l i ty temperature (NDT) for the s teel .the dynamic tear and drop weight test measure the relative rather thanabsolute toughness of the s teel .

Since the test uses an

This test is also conducted over a range of temperaturesBoth

The fracture mechanics based tests are more recent and have not

been employed very widely to measure the toughness of s teel welds,These tests develop a plane-strain s tress intensity factor(K I c) thatcan be measured by calculating the s tress s tate at the t ip of a f lawin the test specimen. This property has the advantage of being a fun-damental property of the mater ial directly related to i ts tendency topropagate a br i t t le crack under s tress . Accordingly, these values canbe used directly in calculating mater ial behavior under service condi-t ions. The crack opening displacement (COD) testing extends fracturemechanics into elastic-plastic behavior . Since fracture propagationat high strain rates is much easier than fracture init iat ion in manyof these s teels under quasi-static conditions, the prevention of f rac-ture init iat ion insures the avoidance of failure.

measurements of mater ial behavior when appreciable yielding occurs atnotches or f laws before fracture. The K I c and COD measurements allowthe location of the notch or crack at a specif ic location in the welddeposit or HAZ to measure the toughness at this location. However , thetesting requirements are considerably more complex than for t he Charpy,drop weight or tear tests .

The COD test permits

Welding Processes and Electrodes --

While the level of notch toughness of the deposited weld metalwill vary with the composit ion of the deposited metal , the cooling-conditions and the s tructure, a maximum attainable toughness exists

for each strength level (48) . This effect of tensile s trength ontoughness was i l lustrated in Figure 21 for cast s teels and the samesituation applies to deposited weld metal .

a . Shielded Metal Arc - Manually Applied -- Considerable infor-mation is available on the toughness of weld metal deposited fromvarious types of s tandard shielded electrodes. Figure 23 (46) containsCharpy V-notch transit ion curves for a number of electrodes that werewelded into var ious rolled plate s teels . The class of electrode isshown for each curve with the type of s teel plate in parenthesis af tereach one. Each curve also has a thicker black section showing the nilducti l i ty temperature (NDT) for that s teel in drop weight testing.


50/90

-50-


51/90

-51-

Some general ranges or bands of Charpy V-notch transit ioncurves for some classes of electrodes have been established asil lustrated in Figure 23 (46) .of the EXXl5 and EXX16 low hydrogen electrodes.electrodes including type EXX18 as well as EXX15 and EXX16 are s tronglyrecommended for welds that require high toughness.

be sure that these electrodes s t i l l have a low moisture content whenemployed; rebaking and using warm from the oven is recommended.all three types o f electrodes can be used, the EXX1 8 clas ses are favoredbecause of their good welding character is t ics (19) .tempered castings, E11018-G for a minimum yield s trength of 100 ksiand E11018-G or E12018-GF for a minimum yield s trength of 110 ksi arepreferred. The typical Charpy-V-notch transit ion curves for E11016and E11018 weld deposits shown in Figure 25 (19) i l lustrate this tough-ness behavior . The data below this curve shows the composit ion andtensile properties of the deposited weld metal .

This plot shows the superior toughnessThe low hydrogen

It is necessary to

While

For quenched and

The reasons that low hydrogen electrodes are about the only

type used for welding high-strength notch-tough steels are:diff iculty with underbead cracking that could occur on the alloy gradesof s teel used for toughness; better impact properties , as shown inFigures 24 and 25; and the basic mineral coating makes i t feasible toadd carbon and other alloying elements to produce weld metals of varyingcomposit ions and strengths (48) .weld metal deposited from low hydrogen electrodes have also been shownin other work (49) .deposits with a Ni-Cr-Mo-V composit ion and tensile properties of about140 ksi yield s trength, 149 ksi tensile s trength, 16% elongation and60% reduction in area are shown for both Char

h - review of welding

Documents