impact properties of cast steel

7/28/2019 Impact Properties of Cast Steel

1/52

IMPACT PROPERTIES OF CAST STEEL

SECTIONS WITH SURFACE

DISCONTINUITIES

A RESEARCH PROJECT AT CASE INSTITUTEOF TECHNOLOGY

Sponsored by

STEEL FOUNDRY RESEARCH FOUNDATION

Charles W. BriggsDirector of Research

by STEEL FOUNDRY RESEARCH FOUNDATION

Rocky River, Ohio September, 1967

Published and Distributed by Steel Founders' Society of America

Westview Towers, 21010 Center Ridge Road Rocky River, Ohio 44116


2/52

TABLE OF CONTENTS

Page

Scope of the Research Report ............................................................................. 3

Summary of the Research Report Conclusions ................................................... 3

Preface ................................................................................................................. 4

Introduction ........................................................................................................... 6

Selection of Transition Temperature Criteria .........................................................8

Effect of Discontinuities on Properties ...................................................................9

Materials and Procedure ..................................................................................... 10

Mechanical Properties ......................................................................................... 13

Surface Discontinuities ........................................................................................ 15

Gas Cavities ......................................................................................................... 15

Surface Slag Inclusions ....................................................................................... 18

Hot Tears ............................................................................................................. 20

Welded Cast Steel ............................................................................................... 23

Metallography ...................................................................................................... 30

Ductile Fracture ................................................................................................... 33

Brittle Fracture ..................................................................................................... 35

Evaluation of Research ....................................................................................... 38

Conclusions ......................................................................................................... 40

References .......................................................................................................... 42

Appendix I ............................................................................................................ 43


3/52

Steel Foundry Research Foundation

IMPACT PROPERTIES OF CAST STEEL SECTIONS

WITH SURFACE DISCONTINUITIES

SCOPE OF THE RESEARCH REPORT

The ability of a steel casting to withstand impact loading without failureis determined by its strength, microstructure, composition, the magnitude of theimpact stress, the rate of loading, the ambient temperature, the presence of mul-tiaxial stresses and the presence of stress concentrations.

This investigation utilizes three different types of impact tests with differentstrain rates and two types of stresses (bending impact and tension impact) toevaluate the effect of different types of surface discontinuities that produce stressconcentrations or notches for a cast steel of two different strength levels and mi-crostructures.

The studies of this report are an extension of previous studies of a similarnature as to the determination of surface discontinuities on the static and dy-

namic (fatigue) properties. These studies were concerned with the determinationof surface discontinuities on the dynamic (impact) properties.

SUMMARY OF THE RESEARCH REPORT CONCLUSIONS

The results of tests made on cast steel sections made under dynamic loadingin bending impact and tension impact provide significant information concerningthe influence of sound sections and sections containing severe surface discontinu-ities on the impact properties of cast steel.

1. Tempered martensitic structure obtained by quenching and tempering alow alloy (8630) cast steel provided better impact resistance than the ferritic-pearlitic structure obtained by normalizing and tempering. This improvement intoughness was particularly marked in its effect on the transition temperaturesobtained with the three types of impact tests (two unnotched types and onenotched type).

2. Severe surface discontinuities in cast steel result in increased transi-tion temperatures and diminished fracture energies as compared to cast steelwithout the discontinuities. Impact ductility and fracture energy in bendingare dependent on both the location of the discontinuity and the stress concen-tration produced by the discontinuity. Location is not a major factor in ten-

sion impact.

3. Severe discontinuities in welds made in cast steel showed that the welddiscontinuities reduced the impact resistance of the cast steel. The order of de-creasing severity was undercut, slag inclusion and incomplete penetration.

4. Brittle fracture energies measured near the ductility transition temp-erature by impact for casting discontinuities appear related to the stress concen-tration of the discontinuity (proportional to Kt

-2). The Kt values estimated forhot tears are 9 and greater, whereas slag and gas cavity discontinuities werewithin the range of 1.5 to 2.6.

3


4/52

PREFACEto the

RESEARCH REPORT

What constitutes the effect of discontinuities at the surface of a structure on the ability of thestructure to perform its intended service? This has not been an easy question for design engineers toanswer. Most materials and design engineers have some very strongly held opinions on the impor-tance of discontinuities in structures as revealed by non-destructive testing but actual information on

this matter backed with factual data is very limited.It is a well known fact, however, that surface notches have a pronounced effect on the impact

resistance of metals and that surface discontinuities can act as notches. The surface discontinuitiesemployed in the research were gross and very severe ; so severe in fact that they exceed the designatedclasses of discontinuities as established by ASTM in the non-destructive testing standards ; namely, E 125Magnetic Particle Reference Photographs and E 71 Reference Radiographs. It was known before testingthat these severe surface notches would have an effect on the endurance strength of the cast steel.Likewise the V-notch in the Charpy impact specimen has a pronounced effect on the impact propertiesof a similar steel test specimen without the machined V-notch.

There seems to be considerable hesitancy by the engineering profession to draw comparisons be-tween the V-notch impact energy and transition temperature with those obtained by testing unnotched

test specimens in impact-bending and impact-tension containing severe discontinuities. The differencesobtained between the three types of tests are individually concerned with the location of the discon-tinuities with respect to the maximum stress, the microstructure of the steel at the base of thenotch, the severity of the stress concentration on notch acuity and the differences in the dimensionsof the three impact specimens. Some of these factors are not well established and any direct compari-sons or attempts to draw significant conclusions from a comparison of the three sets of impact testdata are not advisable. The only comparisons that can be made are observations concerning the dif-ferences between the transition temperature and impact energy of the cast steel specimens withoutdiscontinuities and those specimens containing the severe discontinuities. Estimates of what mightbe the effect ofminorormild types of discontinuities on the impact energy ortransition temperaturealso cannot apparently be made until factual data are available.

Thus, all that can be said is that severe discontinuities do have a pronounced deteriorating effect

on the impact energy and change the transition temperature of cast steels. Apparently other studiesmust be carried on by testing less severe discontinuities or testing entire steel casting with and with-out discontinuities.

A previous research report by the Foundation issued in August 1966 on The Effects of Surface Dis-continuities on the Fatigue Properties of Cast Steel Sections studied discontinuities in bending-fatigueand tension-fatigue test specimens. Such information could be correlated to the notched R.R. Mooretest normally employed by design engineers.

the degree of quality the more intense the processing requirements and the higher the end productcosts. Extra quality beyond normal commercial standards is in most cases not necessary or requiredbecause of the lack of information on the importance and value that is placed on the various degreesof discontinuities as observed by non-destructive testing. Severe surface discontinuities are not pres-ent in commercial steel castings hence values illustrated by the studies of this report would in all prob-ability not be encountered.

The steel foundry companies through research and the technology which they are generating willcontinue their progress in the production of quality castings but it may be advisable to point out thatunnecessary inspection requirements come high in terms of production costs.

It is the desire of the steel casting industry to produce a quality product. However, the higher

CHARLES W. BRIGGSDirector of ResearchSteel Foundry Research Foundation

4

September 1967


5/52

IMPACT PROPERTIES OF CAST STEEL

SECTIONS WITH SURFACE

DISCONTINUITIES

byE. S. Breznyak and J. F. Wallace

Research Assistant and Professor, Department of MetallurgyCase Institute of Technology

Steel Foundry Research Foundationin contract with

Case Institute of Technology


6/52

IMPACT PROPERTIES OF CAST STEEL SECTIONSWITH SURFACE DISCONTINUITIES

Introduction

One of the major design problems is the oc-casion of brittle fracture in steel structures. Themost damaging feature is that brittle fracturesoccur rapidly and without warning. Such failuresare not predicted because the nominal stress attime of failure may not be high but exists con-siderably below conventional design criteria ofultimate tensile strength or of yield strength. In-cidence of catastrophic brittle failures from cracksinitiating at weld-flaws have been reported forships and for storage tanks where the materialdid not appear to develop its normal strength andductility.(l. 2) The significance of these failuresis that simple uniaxial tension testing at slowstrain rates and at room temperature may not bethe criterion controlling the behavior of a com-ponent under stress. This idea is not surprising

since the tension test does not reflect the proper-ties of steel parts under situations of constraint,low temperature, high rates of loading, and local-ized stress concentration generated by the pres-ence of a flaw or a notch, or resulting from struc-tural design.

It is well-known that certain properties of car-bon and low alloy steels deteriorate with decreas-ing temperature. A temperature is defined as thenil ductility temperature (NDT) below which theability of the steel to absorb energy drops to a

very low value.(3) This point shifts to a highertemperature as stress increases and concentratesas in the presence of a notch. Below the NDT

point, failure is brittle and is characterized bynegligible ductility and occurs by cleavage.

The principal factors which affect the ductile tobrittle transition in steel are temperature, rate ofload application, stress concentration as in thepresence of notch, specimen or part size, materialcomposition and microstructure.

The yield strength of steel increases markedlywith a decrease in temperature. Consequently,brittle fracture and failure can occur with a min-imum of ductility before yielding. Another phe-nomenon which appears to be related to the brittlefracture of mild steels is delayed yielding.(4)Steels loaded rapidly to a stress above the yieldstress exhibit a time delay before plastic yieldingoccurs. The time delay increases with decreasingtemperature at a constant stress. Since brittlefracture will occur when plastic yielding fails to

keep the stress below some critical value, it ap-

pears that a relation should exist between delayyielding and brittle fracture.(5)

The effect of rapid loading on the yield strengthof some steels is similar to that of temperaturebut not as pronounced. However, a relationshipexists between lower yield point and strain rate.(5)This increase in strength at higher strain ratesmay be accompanied by a loss in impact resistance.The ductility transition temperature can be shift-ed 140 F degrees downward by static loading ofCharpy specimens as compared to impact load-ing.(6) This represents approximately a 30 Fdegree rise for each tenfold increase in strain rate.The raising of the yield point is shown to be re-lated with the delay time to yield, since yield timeincreases at a higher strain rate even at normaltemperatures.(7) Thus, it is evidenced that theeffect of low temperatures, high strain rates, and

delayed yield times are interrelated and all pro-mote brittle fracture.

The origin of brittle failures can usually betraced to regions of high stress concentrationswhich are generated from structural stress con-centrations or material discontinuities. The loca-tion of these discontinuities are most always ex-ternal and extend to or originate at the surface ofa component. These are produced as the result ofthe casting process or design and include surfacediscontinuities of porosity, shrinkage, non-metal-lic inclusions and design notches of keyways, oil

holes, screw threads, sharp angles, rivet holes,weld discontinuities, etc.

The effects of a discontinuity or a notch in asteel section on brittle fracture can be surmisedby considering the geometry of a notched impactbar such as a Charpy V notch. As the longitudinalstress at the base of the notch increases, contrac-tion in the thickness direction is constrained bythe relatively unstressed material away from thenotch and a stress is produced in the thicknessdirection. This condition results in a state oftriaxial stress.(8) The significance of this state isthat the longitudinal stress necessary for yieldingis increased. Consequently, the fracture strengthmay be exceeded before plastic yielding can occurand flat, cleavage fracture will result. Finally, theentire flow curve of a notched specimen is raisedabove that of the unnotched specimen by anamount expressed by a plastic constraint factor.Elastic stress concentration at the base of thenotch can be extremely high as notch acuity in-

creases. However, the plastic constraint factor

6


7/52

cannot exceed a value of approximately 3 regard- fore cracking. The transition temperature of aless of how sharp the notch, Thus, it appears that steel is lowered as the microstructure is changedthe effects of a notch are to produce a high elastic from coarse pearlite through tempered bainite tostress at the notch tip, increase the strain rate, tempered martensite. Therefore, fully hardenedincrease the longitudinal stress necessary for and properly tempered steel results in better notchyielding, and raise the plastic flow curve. All of properties than normalized and tempered steelthese are factors that will enhance the probability and considerably better than annealed steel. The

for brittle fracture. high degree of notch toughness of martensite isunderstandable since the ferrite has an extremely

The effect of the size of a specimen is generallyfine grain size with the carbides uniformly dis-to increase the temperature at which brittle frac- persed throughout the ferrite.

ture occurs.(9,10) The toughness of structuralgrades of steel is influenced by section thickness The studies evaluating the influence of chemicalin two separate ways : metallurgical and geometri- composition of carbon steels on impact strength

cal (mass effect). (11) The transition temperature are sufficiently advanced to permit a qualitativeof a given steel usually decreases with increasing evaluation of the effect of individual elements.section thickness because of the coarser grain The effect of the elements on notch toughness ap-size produced by hot working or heat treatment pears to depend on whether the elements are pres-temperatures. This fact indicates that metallur- ent as substitutional or interstitial atoms in thegical structures of thick sections are less tough body centered cubic ferrite lattice.(19) Substitu-than thin sections; however, Charpy type impact tional solute atoms have a relatively small effect;tests on specimens of varying size but with iden- only nickel, molybdenum and manganese exhibit

tical metallurgical structure and geometrically beneficial effects while chromium, copper, vana-similar notches also show a size effect.(12,13) The dium, and aluminum produce mildly adverse ef-geometrical or mass effect explains this phenom- fects.(20) Interstitial elements of carbon, phos-enon in terms of stress distribution and stored phorus, nitrogen, and oxygen all act to reduceelastic energy. A thick section offers more con- notch toughness. For each increase of 0.1 per-straint to plastic flow than a thin section, particu- cent carbon, the 15 ft. lbs. transition temperaturelarly in the presence of a notch and therefore, criterion for Charpy V notch specimens is raisedcontains a more unfavorable state of stress. The about 25 F degrees ; the 10 ft. lbs. transition isthicker sections also provide a large reservoir for raised about 20 F degrees. Carbon also has astored elastic energy. The Griffith criteria for marked effect on maximum energy for fracture;

brittle fracture require that the elastic strain the higher the carbon content, the lower the max-energy must provide the necessary energy for the imum energy.(20) Phosphorus increases the tran-

formation of the fracture surface and conse- sition temperature even more rapidly than equalquently the greater the available stored energy, amounts of additional carbon; the 15 ft. lbs. tran-the greater is the ease with which a rapidly sition temperature is raised about 13 percent forspreading crack can be propagated.(5) each 0.01 percent increase in phosphorus con-

tent.(20) Sulfur, usually present in steels as non-

adverse effect on impact dependiong on the typetoughness of steel is its microstructure.(14, 15,16)

An increase of one ASTM grain size of the ferritesize, amount and distribution present in the alloy.

phase has been shown to raise the transitiontemperature by 25 F degrees. The size of the Welding is another source of discontinuitiesferrite grains is determined to some extent by the which can lead to brittle fracture. These discon-austenite grain size from which they form but it tinuities can be classified as both metallurgical andis more strongly influenced by the cooling rate geometrical. The local application of heat during

through the transformation range. Nickel and welding results in changes in the microstructuremanganese increase the toughness of ferrite, par- of the parent material in the heat affected zonetially by inducing finer grains.(17) Metallographic usually to the detriment of notch toughness. Thestudies have shown(18) that initial cracking occurs geometrical aspect involves stress raising dis-in the ferrite phase at the interface of the ferrite continuities represented by cracks, voids, lack ofand pearlite or martensite. The explanation put penetration, undercut or slag inclusions in theforth is that under applied stress the weaker fer- weld. These discontinuities will have an effectrite is forced to supply the plastic deformation similar to that of a notch, as previously discussed,while the stronger dispersed pearlite acts by con- and will therefore enhance the probability ofstraining the ferrite, thereby limiting its flow be- brittle fracture. Other factors which have been

7

Another important factor contro11ing the notchmetallic inclusions, reduces ductility and has an


8/52

shown to affect the notch toughness of weldmentsadversely are residual stresses, entrapped hydro-gen, and poor weld design.(21) Poor workmanshipis known to have been responsible for catastrophicfailures.

Several tests have been devised for measuringthe ability of mild steels to resist fracture. Thesetests such as the V notch Charpy impact, Drop

Weight, Tipper, Robertson Crack Arrest, and Le-high each have a different way of establishing atemperature below which brittle fracture oc-curs.(21) The width of the transition zone varieswith test conditions, the specimen geometry, andthe chemical composition of the steel. For a par-ticular steel under consideration, the results ofeach test cannot necessarily be correlated with theother tests. However, each type of test is effec-tive as a means of comparing the different mate-rials. The most commonly used test is the CharpyV notch and it is recommended that this test al-ways be made in addition to any other desired testfor the purposes of comparison.(22)

Selection of Transition Temperature Criteria

Numerous fracture tests have been devised andused to provide information where behavior ofmaterials under conditions of fracture is a prob-lem. These tests may be conducted at a particu-lar temperature of interest such as the minimumservice temperature, at room temperature, or overa range of temperatures which will include the en-tire ductile to brittle transition. Also useful in-formation is obtained from the impact energy re-quired for rupture, ductility measurements, andthe fracture appearance of the broken specimen.Confusion can arise in comparing the data fromseveral investigations even when the same testis used if the criterion employed to describe thetransition is not uniformly selected. In view ofthe complexity of this measurement, it is dis-cussed in some detail.

A. Energy Transition

1. Energy at a particular temperature

2. Temperature for a selected impact en-ergy level such as 10 ft. lbs. or 15 ft. Ibs.

3. Temperature range for the transitionfrom low energy fracture to maximumenergy

4. Maximum energy at 100 percent fibrousfracture appearance

B. Fracture Transition

1. Temperature for which the fracturefibrosity or crystallinity is at some par-ticular level such as 50 or 75 percentcrystallinity of the fractured face

C. Ductility Transition

1. For bending loads, lateral contraction

measurements at the fracture face at thetension loaded side or lateral expansionat compression loaded side

2. For tension loaded specimens, reductionin area and elongation measurements

3. Temperature at which first indication ofductility is measured. A 1 percent later-al contraction under the notch has beenused for Charpy testing.

A typical energy temperature and crystallinitytemperature curve showing some of these pro-

posed criteria is given in Figure 1. For the pres-ent work three criteria were used to compare thetransition behavior of the steel :

A. Ductility Transition Temperature (DTT)-defined as the temperature below which noplastic deformation was detected as deter-mined by lateral contraction measurementsfor the bending tests and by percent reduc-tion of area and percent elongation for thetension tests.

B. 50 percent Fibrous Transition Temperature

(.5FTT)-defined as that temperature atwhich the fracture face appearance was 50percent fibrous and 50 percent crystalline.

8


9/52

C. Complete Fibrous Transition Temperature(CFTT)-defined as that temperature atwhich the fracture face was essentially com-pletely fibrous (90 to 100 percent). Thispoint coincided with the maximum energyabsorption at the upper range of the tran-sition.

Microstructural Examination . . . Knowledge

of the microstructure in the regions near the dis-continuities is helpful in determining transitiontemperatures. The microstructure, by using stand-ard light microscopy techniques in discontinuityareas, is compared to that of the basic materialto determine the extent of any structural dif-ferences.

Electron fractography techniques also are help-ful and used to study fracture morphology in thebrittle, ductile and intermediate fracture regions.

A two-step replicating technique was used in thestudies of this report. The subsequent viewing

and photography of the replicated surfaces wereperformed using a Philips Electron MicroscopeModel EM 100-B.

Effect of Discontinuities on Properties

A minimum of information is available whichdefinitely analyzes the effects of discontinuities onthe dynamic properties of various steels. Thebulk of information available, concerned with im-pact loading, is performed with specimens con-taining carefully machined notches. Most of theavailable information on the effects of discontinu-

ities has been obtained by slow room temperaturetension and bend tests. For example, a study ofthe effect of mass on the properties of 4330 caststeel having 6-inch-square and 8-inch-round sec-tions showed that both tensile strength and duc-tility were reduced away from the ingot sur-face.(23) The loss in properties was attributed tomass effect caused by the increased amount andgreater intensity of the alloy segregation andlarger dendritic arm spacing and greater concen-trations of non-metallic inclusions toward thecenter of the castings.

In a similar effort(24) the mass effect on the ten-sile and Charpy V notch properties of four heatsof various low alloy cast steels was investigated.At the centers of the castings where segregatedinter-dendritic areas and agglomerations of inclu-sions are prevalent, slight reductions were foundin tensile and yield strength while tensile ductilitywas noted to be greatly affected. However, CharpyV notch impact values measured in various loca-tions in the castings were found to be only slightly

affected. These results were surprising in view ofthe limited tensile ductilities observed for themore slowly solidified section centers.

Another investigation(25) was initiated to estab-lish a correlation between mechanical propertiesand imperfection size as determined by radio-graphic standards. Solidification imperfections ofboth porosity and shrinkage types with 4140 cast

steel heat treated to a strength level of 180,000 to200,000 psi were studied.A least squares statis-tical regression analysis showed a linear decreasein tensile strength with imperfection size in-creases. No marked differences between the twotypes of imperfections in reducing the tensilestrength was observed.

In similar investigations,(26, 27) attempts weremade to correlate solidification imperfections withtensile properties of cast stainless steel CA-15(12 Cr) plates. Radiographic techniques wereused to rate the porosity solidification imperfec-

tions on 0.1, 0.2, 0.3, and 0.6 inch thick plates heattreated to 180,000 to 200,000 psi ultimate strength.Results indicated that as the intensity or size ofthe imperfection increased, the tensile strengthand elongation decreased significantly. The yieldstrength showed no marked change with increas-ing intensity or size of imperfection for the rangeinvestigated. It should be pointed out that thediscontinuities were so prominent that they af-fected materially the cross section of the testspecimen.

Another investigation on the effect of centerline

and gross shrinkage porosity on 8630 cast steel(28)showed that severe centerline shrinkage porosity(Class 6 ASTM E-71) resulted in less than a 10percent reduction in tensile strength ; however,gross cavity shrinkage produced strength lossesof up to 50 percent.

Some investigations have been carried out tofind the effects of porosity on the mechanical prop-erties of welds. In one study(29) in the specimencross section of mild steel (35,000 psi yieldstrength) welds were reduced by porosity up to7 percent without significantly changing the me-chanical properties as determined by room tem-perature static tensile and Charpy V notch im-pact tests. The shape and distribution of theporosity did not cause significant differences inany of the tests. A similar study(30) was con-ducted to evaluate porosity effects in a higherstrength steel weld (100,000 psi yield strength),The experimental results showed that the crosssection of welds could be reduced up to 5 percentwithout appreciably changing the mechanical

9


10/52

properties measured. The effect of shape, loca-tion and distribution appeared negligible in theroom temperature tensile and Charpy V notchimpact tests. In bend tests, the porosity effectswere found to be more critical if located near theconvex surface. Above 5 percent weld porosity,all mechanical properties were noted to fall awaysharply to low levels.

From these data it was evident that a needexisted for information relative to the effect ofcasting and weldment discontinuities on the dy-namic properties of steel structures. Compre-hensive answers are needed to the followingquestions : What types and amounts of discontinu-ities can be contained in an economical and reli-able design ? If a discontinuity is present, shouldthe part be accepted, reworked or rejected? Ifthe engineer is to accept a part containing dis-continuities, he will need quantitative data as abasis for evaluating the effects of these discon-

tinuities. If the part containing the discontinuitycan be accepted, therefore it was produced moreeconomically than if it required reworking or hadbeen rejected. Most important, if the basis forevaluation is feasible, the part's reliability in serv-ice is assured.

A two-part investigation was undertaken tostudy the effects of discontinuities on the proper-ties of steel castings under dynamic loading. Thefirst part evaluated effects of cast steel sectionsunder fatigue loading of reverse torsion and re-verse bending and the results have been re-

ported.(31) Part two of this investigation as re-ported herein is concerned with adding to theknowledge of impact behavior of cast steel andcast steel weldments containing various surfacediscontinuities. For this study a popular Ni-Cr-Mo cast steel (8630) was selected. The specificdiscontinuities investigated in the designated castmaterial were surface discontinuities ; slag inclu-sions, hot tears and gas cavities. Welds wereproduced in the cast steel ; the impact behavior ofsound welds in the cast steel were compared towelds with undercut, internal slag inclusions and

incomplete penetration. Sound stock was cast toprovide test specimens for the purpose of controland comparison. Room temperature tensile datawere obtained for randomly selected heats. Twoprimary heat treatments were employed ; quenchand temper to a 130,000 psi to 150,000 psi leveland normalized and temper to a 80,000 to 90,000psi strength level. A few studies were also con-ducted with quench and temper and normalize andtemper steel specimens heat treated to the same110,000 to 120,000 psi tensile strength level.

Materials and Procedure

General Casting Practice . . . The steel pro-duced for this investigation was a cast Ni-Cr-Mo(8630) composition. Conventional basic induc-tion melting practice and aluminum deoxidationwere used to produce the test castings. Ninedifferent heats of steel were employed. Thecomplete compositions of these heats appear in

subsequent tables of impact data in Appendix I.In general, the percentage composition rangeswere :

The steel was cast into sand molds to producebars for tensile impact specimens and plates forbend impact specimens. Tensile specimens andCharpy V notch specimens were machined fromthese sections.

Production of Bending and Tensile ImpactSpecimens . . . The casting discontinuities in-vestigated included hot tears, slag inclusions andgas cavities. The procedures used for incorpo-rating these flaws into the castings have beenthoroughly detailed in the Steel Foundry Re-search Foundation's Research Report publishedin August 1966.(31) Briefly, hot tears were ob-tained by hindered contraction and formed per-pendicularly to the longitudinal axis of the barsand plates. Large surface slag inclusions wereproduced by embedding slag particles into thedrag surface of the mold cavity Gas cavitieswere obtained in the bend specimens by elimi-nating the aluminum as the final deoxidizer andusing only ferrosilicon. The size and distribu-tion of the gas cavities produced varied, there-fore an alternate method was employed for thetension impact specimen.A single hole 1/16inch in diameter was drilled normal to the longi-tudinal axis at the gage center of the (0.357 inch

diameter) tension impact specimens. The sever-ities of the hot tears, slag inclusions and gas cav-ities were rated according to ASTM referenceradiographs E71-64 and ASTM magnetic particlereference photographs E125-63.

Three severities of shrinkage were producedby improper risering techniques in the bendingimpact specimens. However, only two severitiescould be successfully established within thesmaller gage section of the tension impact spec-

10


11/52

imens. The severities of the shrinkage discon-tinuities were determined by ASTM referenceradiographs E71-64 and ASTM magnetic particlereference photographs E125-63.

Bending and tension impact specimens free ofdiscontinuities were also produced for controland comparison purposes ; also, tensile andCharpy V notch specimens were machined fromcoupons to establish the base mechanical and im-pact properties of each heat.

Welding Procedure . . . The bending and ten-sion impact specimens were welded by a com-mercial welding company. The types of welddiscontinuities investigated were : 1) sound weld,2) sound weld with the reinforcing head removedby machining, 3) weld undercut after machining,4) weld slag inclusions after machining, and5) weld incomplete penetration after machining.

The low hydrogen welding electrodes usedwere Hobart LH 918M to obtain notching me-chanical properties after heat treatment. A pre-heat temperature of 300 degrees F was employedand a post heat temperature of 1100 degrees Ffor one hour was applied to the specimens im-mediately after welding to retard cooling andprevent the formation of underbead cracks.

A symmetrical, double vee butt joint was em-ployed and is shown with the order of weldingpasses in Figure 2a. The tension impact specimenweld joint design is shown in Figure 2b and themachined specimen is shown in 2c.A 0.045 inch

diameter welding electrode of designation Na-tional Standard 2 1/4 Cr, 1 Mo was selected toprevent specimen overheating and to permit mul-tiple weld passes in the small weld section.The weld material composition was similar tothe LH9018 electrode used for the bending spec-imens. The weld metal was deposited usingHeliarc tungsten inert gas technique (TIG) .Since the smaller weld rod was not coated, anargon cover (a number 8 cup with a flow rate of12 cubic ft./hr.) was used during the welding.

Machining . . . The bending impact specimens

were machined to the test bar dimensions of7 1/2 inches long by 1 inch wide by 1/2 inch thickfrom as-cast plates 8 inches long by 1 1/4 incheswide by 5/8 inch thick. The plates were groundon a rotary vertical spindle : a Blanchard surfacegrinder using a selective grinding procedure to as-sure equal stock removal from opposite surfaces.This technique was closely monitored when grind-ing the bars containing shrinkage in order tomaintain the shrinkage centrally with respect to

the surfaces of the finished machined bar. Slowtable and spindle speeds of 9 and 900 rpm, respec-tively, with a minimal automatic feed of 0.004inch per minute were used. These slow speedsand feeds were used in conjunction with a resinoidbonded, cool cutting wheel to minimize generationof heat and residual tensile stresses in the bend-ing specimens.

The tension impact specimens were made byinitial rough turning and finished by contouredwheel grinding. The finished machined dimen-sions of the tension impact specimens are shownin Figure 2c. Charpy V notch specimens weremade to dimensions according to ASTM specifica-tion. The 0.010 inch radius at the base of the 45degree included angle V notch was lapped to re-move marks of machining. The tensile specimensused to determine the static strengths of thevarious heats had a diameter of 0.357 inch and agage length of 1.40 inches.

11


12/52

Heat Treatment.. . The tensile strength levelsselected were 80,000 to 90,000 psi and 120,000 to145,000 psi with some limited number of spec-imens at 110,000 to 120,000 psi tensile strength.These levels were obtained by two heat treat-ments: (1) normalize and temper and (2) waterquench and temper. Specimens were waterquenched from the tempering terperature. Thespecimens tested with the weld bead in place were

heat treated undera protective neutral atmos-phere to avoid decarburization. The other weldspecimens were heat treated in gas fired furnaces.

All the other specimens containing gas cavities,hot tears, surface slag inclusions and quenchcracks were austenitized in salt pots prior toquenching ; the tempering was conducted in elec-tric resistance heated furnaces.

Mechanical Tests.. . The machined specimenswere subject to various mechanical tests which in-cluded standard tensile, impact Charpy V notch,

impact tensile and impact bend testing. Tensiletesting was performed as a control to ascertainthe static strength and ductility response of thesteel to the thermal treatments. The data ob-tained with the Charpy V notch testing served asa control and provided a baseline for comparisonwith results of the impact testing with the scaledup bending specimens and with the tension impacttesting. The Charpy testing was performed on aWidermann-Baldwin Pendulum Type Impact

Tester with a 240 ft. Ib. capacity and a pendulumvelocity at impact of 204 in./sec.

Bend testing of the scaled up specimens wascompleted at Watertown Arsenal, Watertown,Massachusetts, on one of the three largest pen-dulum impact machines in the world. The 2200ft. lb. energy capability of this testing apparatusallowed for a scale up of the bending specimens todimensions as previously described. The arm

length is 80 inches and swings through an angleof161 degrees as measured from the rest position.The velocity generated at the pendulum strikeranvil just prior to impact is 346 in./sec. Thespan length of the specimen holding fixtures was4 7/8 inches. Those bending specimens (specimenlength equals 7.5 inches) which did not fracturewere required to bend through an angle of 112degrees to clear the holding fixture span.

A technique for tension impact testing wasdeveloped to measure and record drop weight

velocities prior to and after fracture. From thesedata, the energies required to affect specimen frac-ture were calculated. The system which wasused, shown in Figures 3a and 3b, consists of aPhillips tube counter, a chronograph adapter gat-ing circuit, and silicon photovoltaic light sensors.Two light sources and the light sensors are housedin opposite legs of a specially constructed horse-shoe shaped bracket mounted to the right of theimpact fixture shown in Figure 2b. A light sensor

12


13/52

is located behind each of two slits which weremilled 0.020 inch wide by 1/2 inch long at exactlyone inch separation. As the 80-pound drop weightdescends from a fixed height of 10.5 feet, a pro-truding shield passes over the first slit, blockingthe incident light on the sensor. The sensor thensends a signal to the gating circuit which acts asa high speed electronic switch and activates thecounter. The counter records counts at a rate of

100,000 counts per second until the shield passesover the second slit, at which time the signal fromthe second light sensor to the gating circuit stopsthe counter.

The number of total counts for the drop weightto traverse one inch is recorded on the tubecounter and thus the velocity is computed. Theupper light sensor was positioned 3/4 inch belowthe initial drop weight contact with the lowerspecimen crossbar holding fixture in order to re-cord the drop weight velocity (f) after specimenfracture. The initial velocity (o) was determined

by unobstructed free fall and was recorded at 26ft./sec. (312 in./sec.) at the position of impact.This free fall velocity (o) was exactly that calcu-lated from the known height (h) of the drop[o= (2gh)

1/2], indicating a negligible friction lossfrom the drop weight guide rails. From a knowl-edge of the initial velocity (o) and the final veloc-ity (f) and the mass of the drop weight (m1) andthe mass of the lower erossbar specimen holder(m2), the energy (E) for fracture was computedas follows:

The degree of uncertainty with this system is10 microseconds while the time duration for frac-ture was in the order of 1 millisecond, well withinrecording capability of this system. From theenergy relation above, it is implied that no energylosses in the system, such as heat, occur and thatthe final velocity of the specimen crossbar holderis the same as that of the drop weight. This latterassumption was also investigated by high speedphotography. From the high speed photographsa comparison was made with the counter tech-nique of the drop weight velocities after fracture.

Comparison of the velocities computed from thecounter readings to those taken from the photog-raphy data shows the former to be higher in allcases. However, the maximum deviation fromthe high speed photography data was less than 5percent for these tests, indicating a high degreeof correlation between the two testing techniques.

A third method was used to compare fractureenergies with those obtained using the countertechnique. Strain gages were mounted on the

lower impact holding fixture to provide force-timeoscilloscope traces. The force-time curves wereconverted to force-elongation curves. The areaunder the force-elongation curves provided thefracture energy values for comparison with theother techniques used. In all three cases higherimpact values were recorded for the counter tech-nique than for the converted force-time traces.The maximum deviation from the counter tech-

nique is 9.0 percent which is considered to be goodin view of the arbitrary method used in tracingthe best mean curve through the vibrational fluc-tuation of the force-time trace and the approxima-tions used in the subsequent conversions and cal-culations.

Mechanical Properties

The strength levels and ductilities for the var-ious heats as measured by the standard roomtemperature tensile testing are reported in Table1. The heats are also identified in terms of heat

treatment, types of impact tests and discontinu-ities. The reported hardness values shown inTable 1 are an average of at least six readingsfrom three impact specimens from each heat.The Brinell hardness numbers when converted toapproximate ultimate tensile strength levels arecomparable to the values obtained for the tensiontested keel block stock.

Charpy V notch impact values plotted in theform of ductile to brittle transition curves areshown in Figure 4 for three heats of quenchedand tempered, and normalized and tempered mate-

rials. These are heats that were selected ran-domly to provide a control test or a baseline forcomparison with subsequent impact testing. Theimpact energy data for fibrous fracture conditionand lateral expansion of these steels are listedin the tables in Appendix I. It will be observedfrom Figure 4 that the transition occurs more

13


14/52

14


15/52

sharply and at lower temperatures for thequenched and tempered heat treated cast steelscompared to the normalized and tempered caststeels. Above the completely fibrous transitiontemperature, the fracture energy for both micro-structures are at about the same level in spite ofthe differences in strength.

Figure 5 shows the ductile to brittle transitioncurves for two cast steels of relatively the samestrength level. One of the steels has been quenchedand tempered and the other has been normalizedand tempered. It can be seen that the quenchedand tempered steel gives the best impact proper-ties. The impact energy data for fibrous fracturecondition and lateral expansion of these steels arelisted in the tables in Appendix I.

The reproducibility of the data is indicated inTable 2 where the transition temperature data are

summarized for the Ni-Cr-Mo cast steels. Theductility transition temperature of - 200 degreesF for the quenched and tempered materials isshown to be 100 F degrees below that observed forthe normalized and tempered. When the 50 per-cent fibrous (.5FTT) and completely fibrous tran-sition temperature (CFTT) are compared for eachof the two property levels, a greater divergenceis noted, that is, approximately 125 F degrees and170 F degrees, respectively. The scatter observedin the impact levels within each structure condi-tion are normally expected and can be attributed

in part to variations during experimentation andto slight variations in the strength levels of thesteels.

Surface Discontinuities

Impact energy-temperature transition curvesfor the surface discontinuities of gas cavity, slaginclusions, hot tear and welding discontinuitieshave been constructed from the test data. In eachcase the various discontinuities are compared to

15

sound steel test specimen data. Ductility, fibros-ity and impact energy data are listed for the testtemperature range and the values are listed in thetables in Appendix I. The degree of severity ofthe various discontinuities rated according to

ASTM specifications are shown in Table 3. Thetransition data are summarized in Tables 4 and 5.

Gas Cavities

Gas cavity discontinuities appear as surface de-fects but can extend in depth into the castings.The appearance of gas cavities at the surface andin the fractured sections are shown with bendingimpact energy-temperature curves in Figure 6.

Although the gas cavity size and distributionvaried to a slight degree among the cast bars, the


16/52

data are reasonably consistent. The area occupiedby the gas cavity discontinuities ranged from 12to 21 percent. The severity of the transition shiftto higher temperatures of cavity containing barsas compared to sound bars was second only to thehot tears. The CFTT was increased by 100 F

degrees over the sound materials for both struc-ture conditions. Above the CFTT the fractureenergy diminished to approximately 25 percentof the sound material for the quenched and tem-pered and 35 percent for the normalized and tem-pered bars, respectively.

16


17/52

17


18/52

The tension impact energies observed with thegas cavity discontinuities also were affected. Be-

cause of the small gage section of the tension bars,cavities could not be produced during the castingprocess with sufficient reproducibility and there-fore were simulated by drilling a 1/16-inch diam-eter hole in the test bars prior to heat treatment.The cavity produced occupied an area of 23 per-cent of the gage section; this constitutes a pro-nounced portion of the cross section. A photo-graph of the resulting cavity in the tension barand the fracture surfaces obtained are shown

with the fracture energy curves in Figure 7. Thetransition temperatures are comparable to those

obtained with the bending impact tests. The frac-ture energies above CFTT were also markedly re-duced to approximately 25 to 30 percent of thatof sound cast steel for both structure conditions.

Surface Slag Inclusions

The slag inclusions of the surface discontinu-ities least affected the transition temperatures inthe bending tests as compared to sound material.

The appearance of surface slag and representative

18


19/52

examples of the fracture faces are shown with theimpact energy curve for the bending tests in Fig-ure 8. The total areas of the slag projected intothe cross section of the bending specimens wereless than 5 percent. The fracture energy above

the CFTT as compared to sound material wasdiminished to approximately 50 percent for thequenched and tempered and 40 percent for thenormalized and tempered bars, respectively. Thetransition data for tension impact bars containingslag were similar to those obtained with the bend-ing tests except for the ductility transition whichoccurred approximately 50 F degrees lower for thetension impact tests with both heat treatmentconditions. These data are compared in Table 4.

The measured areas occupied by the slag discon-tinuities in the gage cross section were found torange between 6 and 14 percent. The fractureappearance with slag discontinuities is shownwith the fracture energy-temperature curve in

Figure 9. The tensile fracture energies abovethe CFTT were decreased to 40 percent that ofthe sound steel material for the normalized andtempered condition as also was observed for thebending tests. The fracture energies were moresharply reduced for the specimens with slag in-clusions in the quenched and tempered conditionas tension tested; these fracture energies werelowered to approximately 30 percent of that ob-tained with the sound material.

19


20/52

Hot Tears apparent. The magnitude of the fracture energyThe fracture energies for the bending test bars values obtained at a constant temperature are

above the CFTT dropped sharply to about 25 to functions of both defect size and position of the30 percent that of the sound material. Since hot tear with respect to the striker anvil. The

neither position nor degree of severity of the hot higher impact values were invariably obtained fortears could be controlled during the casting proc- the specimens with the hot tear discontinuities atess, approximately 70 specimens for each struc- positions away from the area of the striker im-ture condition were tested to establish trends. The pact. This result is to be anticipated since bothhot tear areas were measured on the ruptured the tensile stresses and the rate ofload applicationfaces and were found to occupy 5 to 40 percent of diminishes as the distance from the striker in-the specimen section. An example of brittle and creases. The results show, therefore, that whenductile fracture appearance for bending impact is located at the positions of high stress, hot tearsshown in Figure 10 for both heat treatments. The are highly detrimental. The ductility and frac-scatter obtained with the fracture energies is ture transitions were observed to occur at sub-

20


21/52

stantially higher temperatures than recorded for 0.030 to 0.040 inch from the surface while thethe sound bars as is shown in Tables 4 and 5. area occupied by the discontinuity in the gageHowever, the transition temperatures did not diameter was within the range of 4 to 10 percent.exceed that of the tests with the Charpy V notch Tensile stock with larger hot tear discontinuitiesbars. fractured during the specimen machining opera-

The impact data obtained for the hot tear tion. The magnitude of the stresses generatedtension impact specimens shown in Figure 11 as a result of the grinding operation was ap-appear more reproducible. Measurements made parently sufficiently high to induce failure ofof hot tear discontinuities on the fractured faces those bars containing hot tears greater than somerevealed the initial crack depths to be within critical size. The machining operation conse-

21


22/52

quently provided an indirect screening of themagnitude of the hot tears which were tested.The transition from brittle to ductile fracture for

tension impact also shifted to higher temperaturesfor both heat treatments as compared to thesound steel, as shown in Figure 11. The fractureenergy above the CFTT was reduced to 20 percentof that of the sound steel. The CFTT was ob-served to occur for both the tensile and bendingimpact tests at approximately 50 F degrees abovethat of sound steel for the quenched and temperedspecimens and 200 F degrees above the soundsteel for the normalized and tempered condition.

The hot tear discontinuities had the most del-eterious effects on the impact properties of thecasting discontinuities investigated. Since surface

notch acuity is a primary factor in controllingcrack formation and propagation, this effect isnot surprising." However, it might also be antic-ipated that the transition range would be higherthan observed for the Charpy V notch testing.This was not the case and the reason could beattributed to crack depth and possibly other

* Note-Entire cross section is uniformly stressed inter-nally on tension impact testing.

22


23/52

metallurgical variables, such as the decarburiza-tion at the surface of the hot tear.

Welded Cast Steel

Typical microstructures obtained for the fullyheat treated weldments are shown in Figure 12.The as-welded structure for cast steels is readilydiscernible and normally consists of three regions:the weld metal, a coarse grained heat affected zone

and the structure of the base metal cast steel.The beneficial effect of the heat treatment is ahomogenization of the structure in which thesezones are no longer as clearly discernible. Theweld material appears slightly coarser grainedthan the cast steel for both the quenched andtempered and normalized and tempered heat treat-ments. Some trace of the heat affected zone isshown for the quenched and tempered bendingspecimen weldment.

23


24/52

The appearance of the weld in the tension barspecimen is somewhat more uniform than in thebend tests. It was noted earlier that differentwelding techniques were used for the tensile speci-

mens than for the bending specimens. The formerwere welded with uncoated electrodes under anargon atmosphere using the tungsten inert gas(TIG) process while for the latter the specimenswere metal arc welded in air with coated elec-trodes. The filler weld rod material was ofslightly different composition for the two weldingtechniques. For normalized and tempered speci-mens, the weldments appear similar, exhibiting atransition which occurs from the weld zone ofpredominately ferrite to the base metal that con-tains an increased amount of pearlite.

The energy-temperature curves for impact test-ing with the sound welded specimens are shownin Figure 13. The weld discontinuities are com-pared to machined sound weldments in each case.Ductility, fracture appearance, and impact energydata are given in the tables in Appendix I. Asummary of transition data is shown in Table 5.The severity of the various weldment discontinui-ties are shown rated according to ASTM specifica-

tions in Table 3.

Sound, Machined Welds. . . Sound weldmentswere machined to remove the weld reinforcementto the uniform thickness of the cast steel plates.The bend test results indicated the quenched and

tempered sound machined weldments developedimpact energies greater than observed for thesound cast steel. The impact energy transitioncurves are compared in Figure 13. This behavioris attributed to the response to heat treatment ofthe weldments. Hardness measurements travers-ing the weld section showed an average BHN of225; the equivalent tensile strength is approxi-mately 120,000 psi. The cast steel hardness wasmeasured to be 285 BHN; the ultimate strengthis 132,000 psi as recorded in Table 1. Because ofthe lower strength level of the quenched andtempered weldments as compared to the castmaterial, the increase in impact toughness wouldbe an expected occurrence. The transition datawere similar; the ductility transition is slightlylower for the cast steel while the CFTT wasapproximately 50 F degrees lower for the welds.

The transition temperatures for the cast steeland the welded cast steel in the normalized andtempered condition were also similar; the cast

steel exhibited a slightly lower ductility transition

24


25/52

by 50 F degrees and the welds had a lower CFTTby 50 F degrees. The impact fracture energies forthe welds and cast steel were equivalent.

for the welded cast steel, however, were within80 percent of the sound cast steel. Similar re-sponse of the welded cast steel to heat treatmentas compared to the cast steel was observed as

The welded, machined tensile impact specimens indicated by hardness measurements. The transi-in neither the quenched and tempered nor the tion temperatures for the normalized and tem-normalized and tempered condition developed duc- pered welded cast steel varied only slightly fromtile fracture energies as high as recorded for the that of the cast steel. By some criteria, the castsound cast steel. The curves are compared in steel provided lower transition temperature andFigure 14. Fracture energies above the CFTT by other criteria, the transition occurred at a

25


26/52

lower temperature for the welded and machined tween the as-welded and as-machined weld bend-cast steel. The fractures for both the bending ing impact results. Similarly for the tension im-and tension test bars occurred through the weld pact testing, no variation in transition tempera-metal and not at the heat affected zone. tures was observed from the as-weld surface con-

dition. Fracture energies measured for the as-Sound, As-Welded..... The sound, as-welded welded cast steel, as shown in Figure 16, displayed

cast steel test bars contained the reinforcing welds a decrease for both heat treatment conditionsin place during testing. This condition resulted in because of the mild notch caused by the presence

of the weld bead. The weld fracture for the ten-a position of stress concentration at the juncture

sion bars was also observed to occur at the weldof the weld and the cast steel. The sound, as-

welded bending test bars for the quenched and deposit-cast steel interface at the location of thestress eoncentration resulting from the mild notch.

tion temperatures as compared to sound machinedwelded cast steel. However, slightly lower frac- Weld-Slag Inclusions . . . A sharp decrease inture energies above CFTT were recorded as shown the bending impact properties resulted from thein Figure 15. The fractures were observed to presence of slag inclusions. The impact energiesinitiate at the surface juncture of the weld bead for the quenched and tempered bending bars wereand the base metal, suggesting a mild notch effect reduced to approximately 35 percent that of thefrom the weld reinforcement. However, for the sound welds as shown in Figure 17. The entirenormalized and tempered condition, no significant transition range was shifted to higher tempera-changes in either the transition temperatures or tures by about 125 F degrees. The location ofimpact energies above CFTT were observed be- the discontinuities in the fracture faces, as shown

tempered condition displayed no change in transi-

26


27/52

in Figure 17, are approximately midway between

the neutral axis and the maximum stressed outersurface fibers. The area of the slag inclusions inthe specimen cross section ranged between 12 and18 percent. Normalized and tempered welds con-taining slag were not as severely affected as thehigher strength cast steel. The ductility transi-tion was not affected; however, the .5FTT andCFTT shifted upward by about 100 F degrees.The fracture energies above CFTT were reducedto about half of that of the sound welds.

Slag weld inclusions in welded cast steel were

found to be detrimental when loaded in tensileimpact. The fracture energies dropped to onethird of sound weld values (QT) and the entiretransition was shifted upward by approximately150 F degrees. The fracture energy transitioncurves are shown in Figure 18. The shift intransition for the normalized and tempered, heattreated cast steel was 100 F degrees or somewhatless. The fracture energies above the CFTT werereduced to about half of that of cast steel with a

27


28/52


29/52

for both structure conditions. A linear shift up- steel plate appeared to be the least severe weld

ward of 50 F degrees was observed for all transi- discontinuity for both the bending and tensiontion temperatures for the quenched and tempered impact testing. Nevertheless, incomplete penetra-condition. For the normalized and tempered welds tion was detrimental to the impact properties ofthe CFTT was unchanged. However, a 50 F de- the steel, as indicated by the energy transitiongree shift to higher temperatures was observed curve for the bending tests shown in Figure 21.for the ductility transition and the .5FTT. The The fracture energy above CFTT was reducedarea occupied by the undercut was measured to to less than half of that of sound welds for thebe 6 to 15 percent. quenched and tempered heat treatment while for

Weld-Incomplete Penetration . . . Incomplete the normalized and tempered incomplete weldpenetration between the weld metal and the cast penetration test, the reduction was less severe

29


30/52

but was still only about 60 percent of that of thesound welds. The upward shift in transitionwas nearly 50 F degrees for both heat treatments.The area of incomplete penetration, as measuredin the fracture face, was 9 to 15 percent. For thetension impact bars the incomplete penetrationarea in the gage diameter was 4 to 6 percent. Thereduction in fracture impact energies, as shownin Figure 22, is not very severe. The fractureenergy developed above the CFTT is approxi-mately 75 percent that of sound welds for bothstructure conditions. An upward shift of ap-

proximately 50 F degrees of the transition for thequenched and tempered cast steel is apparent.Incomplete weld penetration discontinuity did notaffect the CFTT for the normalized and temperedsteel ; the ductility transition and .5FTT, however,did shift upward about 50 F degrees.

Metallography

The microstructure of the cast steels afternormalizing and tempering is a mixture of ferrite

and pearlite and tempered martensite in the caseof the quenched and tempered steels. The micro-structures adjacent to the discontinuities wereexamined to ascertain whether any structuraldifferences were discernible. Evidences of de-carburization to varying degrees were found forthe casting discontinuities of gas cavities, slaginclusions and hot tears. The depth of decar-burization around the gas cavities varies from 5to 10 mils for the sample investigated. The ex-ample shown in Figure 23 shows a decarburizedlayer of approximately 5 mils adjacent to the

pinhole. The slag inclusion in Figure 23 showsa carbon depletion of approximately 10 mils.Cracks are shown emanating from the slag in-clusion into the ferrite region with one crackterminating at the ferrite in a region containingpearlite. The depth of decarburization found withhot tears varied for the specimens examined from3 to 7 mils for the bending impact bars and 1 to3 mils for the tension impact bars. The amountof decarburization increased with crack width.

30


31/52

Decarburization was also detected in the weld

discontinuities of undercut, incomplete penetra-tion and slag. The decarburized layer for the weldundercut was observed to be approximately 2 mils.Decarburization depth for the incomplete penetra-tion was measured to be 3 to 5 mils, as shown bythe example in Figure 24. Regions around theweld slag show decarburization to a lesser ex-tent. Thin regions of ferrite are shown outliningthe slag inclusions in the illustrations of Figure25. A crack is indicated originating at the slag

stringer through a decarburized region into the

surrounding tempered martensitic microstructure.Decarburization is found in only the disconti-

nuities that have access to the surface of thespecimen with the decarburization taking placeduring heat treatment. Extensive decarburiza-tion has been observed to affect mechanical prop-erties of steel. The ultimate tensile strength ofsteel is reduced because the areas depleted ofcarbon do not possess the strength of the un-affected regions. Only limited information is

31


32/52

available on the effect of decarburization on im-pact resistance. One study (32) on several alloysteels showed that the pressure of a decarburized

layer in an edged notch sheet specimen tested intension provided an increase in notch tensilestrength compared to steel without decarburiza-tion. This increase was attributed to the influenceof decarburization on crack initiation require-ments. The percent of shear in the fracture in-creased only slightly indicating that crack prop-agation was little affected. Notch strength wasfound to increase with greater depths of decar-burization. A decarburized zone adjacent to a

stress raiser has a similar effect during impacttesting.

The impact test is conducted under conditionsin which the maximum load applied is sufficientto produce failure. The existence of a decarbur-ized steel layer at the point at which the failurewould normally be initiated will result in an in-crease in the amount of energy absorbed duringthis failure. Such a condition would exist if thesteel at a discontinuity were decarburized or ifthe surface of a bending impact test were de-carburized. The increase in the amount of energy

32


33/52

absorbed during the fracture process would result

because of the greater plastic deformation thatoccurred prior to initiation of cracking.

The existence of the decarburized layer at dis-continuities can increase the degree of fibrosityin the fracture surface, thereby resulting in alowering of the fracture transition temperature.This lowering occurs because the plastic deforma-tion at the discontinuity reduces the notch acuity.

Also a relaxation of the degree of stress triaxiality

occurs when compared to a similar test section

where no decarburization was present and themetal at the discontinuity was at full strength.This effect would explain in part the lower frac-ture transition of the hot tear discontinuities testsas compared to the Charpy V notch specimens.

Ductile Fracture

Measured ductilities at temperatures abovewhere the fracture appeared (CFTT) were ob-served to vary in several ways. First, the ductility

33


34/52

34


35/52

measurements for various heat treatments showedthat the ductilities of steels in the normalized andtempered condition were substantially higher thanfor the quenched and tempered condition. How-ever, there exists a greater amount of plastic de-formation in the lower strength normalized caststeel at test conditions above the ductility transi-tion.

A comparison of the ductilities for the various

casting discontinuities and the weld discontinui-ties in cast steel showed a variation of ductilityin a regular manner with the energy required forrupture. The bending impact ductilities weremeasured by maximum lateral contraction on thetensile side of the test specimens; while for thetension impact test, elongation in the 1.4-inchlong gage and reduction in area were measured.

The tensile failure above the ductility transitionoccurred by the shear mode and produced a neckedregion with the typical cup-cone fracture appear-ance. The effect of discontinuities was a localiza-

tion of the necked region at the position of thecontrolling discontinuity. This effect is similarto machining a necked region or reduced area intothe gage section prior to testing. It follows thatduring the testing, the plastic deformation islocalized and necking will occur early; conse-quently, the accompanying longitudinal elongationis not uniform along the gage length but is re-stricted primarily to the cross section in thevicinity of necking. The radius of curvature ofthe necked region results in a stress concentrationin the necked region leading to a relatively dimin-

ished ductility and a lower energy fracture. (33)This fact implies that the energy absorbed duringductile failure is related to the effective volumeundergoing plastic deformation.

A similar study showed the energy absorbedwas related to the gage volume. (34) Specimensof the same material were tested with 1/2 inchdiameters but of two gage lengths: 2.5 inchesand 5.0 inches. The shorter gage absorbed 1300ft. lbs. of energy and broke; the longer specimen

absorbed 2100 ft. Ibs. of energy and did notbreak.

This trend of the effective volume undergoingplastic deformation as the primary control of theenergy absorbed for ductile fracture is reflectedin the ductilities recorded for both the tensionand bending specimens. The data show that lowerductilities were invariably associated with thelower fracture energies while with higher ductili-ties, the fracture energies increased. The highestenergies and corresponding highest ductilitieswere recorded for the discontinuity free unnotchedtest specimens.

Brittle Fracture

The fracture energy (E) at any temperatureon the transition curve consists of the sum of the

energy to initiate a crack (Ei) and the energy topropagate the crack (Ep):

E = Ei+ Ep

Several investigators have shown that verylittle energy is required to propagate a crack bycleavage and the elastic strain energy availablein the vicinity of the crack tip is sufficient tosustain fracture. It has been determined (35) thatthe plastic work contribution in the brittle frac-

ture of a ship steel broken at room temperature

35


36/52

was of the order of 5 ft. lbs. per square inch orapproximately 0.5 ft. lb. in the case of the ten-sion impact specimens, if the fracture made isactually cleavage. An analysis of the energiesrecorded in the brittle region for the tests ofthis report shows that a negligible portion of thetotal energy is absorbed to provide work of plasticdeformation and that nearly all the energy is ex-pended for fracture initiation, or E ~ Ei near the

ductility transition temperature (DTT) . Sincenegligible gross plastic deformation occurs duringcleavage fracture, the nominal stress for fracturecan be calculated by dividing the maximum stressresulting from stress concentration by the stressconcentration factor.

The maximum normal stress theory which hasbeen found applicable as a criterion for brittlefracture (36) states that when one of the principalstresses reaches a critical value, brittle failurewill occur. The energy necessary for brittle frac-ture is inversely proportional to the square of

the stress concentration factor. A knowledge ofthe stress concentration factor (Kt) was there-fore necessary to determine the energy requiredfor brittle fracture. The various shapes of thediscontinuities were measured and to a large ex-tent idealized to obtain an approximation ofKtfrom compiled stress concentration design data.(36, 37) Values of Kt were taken directly fromstress concentration curves or were obtained byestimates based on the curves and shapes of dis-continuities. The manner of estimation is shownschematically together with the source fromwhich the K

tvalues were selected in Table 6.

The stress concentration produced by the cavitymachined in the tension impact tests to simulategas cavities was readily obtained, since the 1/16inch diameter hole was artificially produced bymachining and was therefore a constant dimensionand location. The Kt value obtained for the ten-sion impact cavity was 1.93. Determination ofthe stress concentration value for cavities in bend-ing impact was more involved since the size, shapeand distribution were all variables. A circularshape for the cavities was assumed and from a

series of measurements from various test speci-mens, a mean gas cavity diameter of 1/16 inchwith spacings of 1/4 inch was selected. Thismethod resulted in a stress concentration valueof2.3. If the gas cavity diameter and spacingincrease proportionately, the Kt remains constant.However, if the spacing is increased for a constantcavity diameter, Kt will increase up to a maximumvalue of3. (36) This example shows how dis-

continuity spacing can result in a substantial de-

crease in Kt and that the limiting value for stressconcentration as produced by gas cavities is 3.

Stress concentration values were obtained forslag inclusions by assuming that the stress con-centration produced by the continuous slag parti-cles could be simulated by a hyperbolic groove asshown in Table 6. The stress concentrations for

36


37/52

slag inclusions are based on measured diametersand depths ranged from 2.1 to 2.6 for bendingimpact and 1.58 to 2.2 for tension impact. It isreasonable to assume that these ranges of Kt areconservative since the interaction of adjacent in-clusions was not considered. The Kt value willincrease sharply and will theoretically approachinfinity as the inclusion particle size decreases orthe section size increases. However, if the sec-

tion size becomes smaller or the inclusion sizebecomes large, then the Kt value diminishes to alimiting value of approximately 1.4 for both bend-ing impact and tension impact tests. (36) Theserelationships point out the strong influence sec-tion size and particle size may have on the frac-ture energy and why correlation between labora-tory test data and actual failures of large com-ponents is different in many cases.

Hot tear discontinuities can be said to haveessentially a zero radius of curvature at its ex-tremity. Theoretically, as the radius approaches

zero, the stress concentration approaches infinity.(38) As an approximation, K

tfor hot tears was

estimated at 10.

The fracture energies recorded for the variousdiscontinuities at a constant temperature near theDTT are listed with the Kt values in Table 7.The temperature selected for the quenched andtempered cast steel was 200 degrees F and -150degrees F for the normalized and tempered caststeel. The energy was plotted for particular dis-continuities as a function of Kt

-2 and is shown inFigure 26.

These data are only approximate values sincethe impact curves from which the energy near

the ductility transition was obtained as well asthe inherent approximations involved in the Ktvalues are a source of scatter. However, fromthese data some generalizations can be drawn:

First, it can be seen that since the impact energyis inversely proportional to the square of thestress concentration factor (Kt

-2) , the stress con-centration exerts a very strong influence on thebrittle fracture energy. Second, it would be

anticipated from this relationship that for ahigher Kt, the energy for brittle fracture would bereduced to a negligibly low or zero value and there-fore, the straight lines drawn through the datapoints shown in Figure 26 should extrapolate tothe origin. That this is not the case can be at-tributed partially to the mechanism of fractureand to the impact test itself. When a crack ispropagating in a brittle manner some small finiteenergy must be supplied since, during the fractureprocess new surface is being created at the ad-vancing crack tip. Further, since sliding contact

is made between the bending specimen and thesupport fixture, work is expended during theimpact test. The most significant factor is thatkinetic energy is imparted to the tension impact

37


38/52

and bending impact specimens and they are pro-pelled at high velocities away from their initialrest position. For these reasons the straight linesdrawn through the data points in Figure 26 inter-cept at some value along the energy axis. Theintercept values for the bending impact tests aresubstantially higher than for the tension impacttests indicating the larger amount of energy ex-pended in these various factors.

The data indicate that the higher stress con-centrations produced by hot tears are more dam-aging to the impact properties than those producedby the gas cavity and slag discontinuities. WhenKt values were estimated for the surface slag in-clusions, the presence of the particle was notconsidered. Since the slag particles are weakand possess a low interfacial strength, the datasuggest that the inclusion is an inert particle withrespect to the fracture process. In other words,it is the shape, size and distribution of the surfacevoids produced by the slag particles and not theparticle itself which determines the stress con-centration value (Kt).

Lastly, the slopes of the curves in Figure 26should provide information relative to cast steelfracture strength for two strength levels (heattreatment and microstructures) and impact tests.However, the slopes of the curves as drawn inFigure 26 show no significant differences.

Evaluation of the Research

The ability of a steel casting to withstand im-

pact loading without failure is determined by anumber of factors. These factors include themagnitude of the impact stress, the rate of load-ing, the ambient temperature, the presence ofmultiaxial stresses and the presence of stressconcentrations. Other metallurgical factors suchas the composition of the steel, its microstructureand strength level will influence the ability ofcast steel to withstand impact loading.

This investigation utilizes three different typesof impact tests to evaluate the effects of somemetallurgical variables and the presence of dif-ferent types of discontinuities that produce stressconcentrations or notches. The tests investigatedinclude the conventional Charpy V-notch, an un-notched bending impact test and an unnotchedtension impact test. These tests vary with re-spect to the strain rate, the presence of a notch,the stress distribution and the amount of metalthat is fractured. The strain rate of the CharpyV-notch impact test is approximately 650 inchesper inch per second compared to 220 for the un-

notched bending impact test and only 45 for thebending impact test. The cross sectional area ofthe steel fractured is 0.124 sq. in. for the CharpyV-notch, 0.10 sq. in. for the tension impact testand 0.50 sq. in. for the unnotched bend test. Boththe Charpy V-notch and the unnotched bendingimpact provide bending stresses from a simplebeam or three point loading arrangement; thetension impact test provides uniaxial tension with-

in the gage length.

Two primary factors are considered in evaluat-ing the results of the impact tests on the soundmetal and determining the influence of the variousdiscontinuities. The first is the temperature ofthe transition of the fracture from ductile tobrittle as measured by several methods. Thistemperature is influenced by the type of test, butfor a given test, the lower the temperature oftransition, the greater the impact resistance ofthe steel specimen. Secondly, the energy requiredto fracture the impact test specimen also hasconsiderable significance, particularly the fractureenergy level in the ductile fracture region.

The higher the energy for a cast steel testspecimen, the greater the resistance of the steelspecimen to failure under impact load. An engi-neering evaluation of the influence of various met-

allurgical factors and types of discontinuities isbased on the effect of the discontinuity on thetransition temperature and fracture energy, tem-pered by the different conditions involved in eachtest. It is evident that the higher strain rate

and presence of the notch in the Charpy V-notchtest produces a higher transition temperature forthe same steel. In addition, the energy requiredto fracture test specimens is generally greatestfor the large cross section of the unnotched bend-ing impact test, intermediate for the tension testbecause of the intermediate volume of metalstressed equally and smallest for the Charpy V-notched test which is fractured with a bendingstress.

The results obtained with the impact tests onsound cast steel serve to substantiate the influence

of metallurgical factors that have been well estab-lished. The improved toughness obtained with atempered martensitic structure compared to aferrite-pearlite structure is evident in all threetests. This condition is primarily indicated bythe lower transition temperature of the steels withthe tempered martensitic structure. The advan-tages of the liquid quenching and tempering treat-ments over normalizing and tempering cycles is

thereby illustrated.

38


39/52

The research of this investigation shows thatthe levels of ductile fracture energy tend to behigher in some cases for the normalized andtempered steels than for the quenched and tem-pered because of the higher strength level ob-tained with the quenching and tempering treat-ment. However, it is generally established thatthe level of ductile fracture energy decreases withincreasing strengths. When the tempering tem-

peratures are adjusted to produce the samestrength level, the fracture energy of thequenched and tempered cast steel specimens forall tests are always higher.

A severe loss in toughness resulted from thepresence of discontinuities on cast steel surfaces.This lower toughness occurred for all tests and atboth strength levels. It should be remembered inevaluating these studies that the discontinuitieswere all greater than those permitted by the

ASTM reference radiographs ofE71 and E99even for the most severe class. The preparationof discontinuities in small test specimens is amost difficult undertaking and hence the discon-tinuities are gross in character. It is practicallyimpossible to prepare specimens with gradationsin severity similar to those shown in the ASTMreference radiographs. This does not mean thatdiscontinuities of less severity than those studiedin this research would show higher impact energiesbut it does illustrate what is possible under mostsevere conditions. Also the test section is re-duced because of the presence of the discontinui-ties and this condition itself results in a re-

duction in the impact energy. This change indimensions may also change the transition tem-perature.

Surface discontinuities as compared to internal

discontinuities are more severe in lowering impact

energy of bending specimens not only because of

stress concentration at the surface discontinuity

but also because of the severity of the notch

resulting from the nature of the discontinuity.

Hot tears which exist at the surface of cast steel

and extend inward are responsible for a severe

loss in impact energy. Considerable scatter waspresent depending on the location and severity of

the hot tears but the over-all results are conclu-

sive. The presence of hot tears of the consider-

able severity tested in this investigation is un-

acceptable in steel castings that are to be utilizedin impact applications. This is in accordance with

the present ASTM nondestructive tests reference

standards.

The presence of severe gas cavities on the sur-face of cast steel also produced a considerable in-crease in transition temperature and a decreasein the level of fibrous impact energy for both theunnotched bend and tensile impact tests on caststeel. This loss occurred with both heat treat-ments and strength levels. It is indicated that

the presence of severe gas cavities in the highlystressed areas subject to impact loading would

render steel castings unsuitable for moderate tocritically stressed impact service.

The increase in transition temperature anddecrease in the fibrous level of ductile fractureenergy that occurs with severe slag inclusionson the surface of cast steel impact specimens wassomewhat less than experienced with hot tearsand gas cavities. It was evident, however, thatan appreciable loss in impact resistance could re-sult depending upon the location and severity ofthe discontinuity. The utilization of steel cast-ings with surface slag inclusions in critically

stressed areas for impact service does not appearadvisable but the casting may well be suitablefor uses involving mild impact loads.

Comparison of the impact test data obtainedwith the sound cast and sound welded cast steelpermits an evaluation of the effect of welding onimpact resistance. It is necessary to allow forthe variations in strength level that resulted insome of the specimens because of the difficulty inmatching the exact heat treatment response ofthe cast steel with weld deposit in order to obtaina comparable strength. The impact resistance of

sound cast steel was similar to the sound machinedwelded cast steel.

The presence of the complete weld deposit inthe sound as-welded but not machined weldedspecimen reduces the impact resistance to somedegree compared to the welded machined speci-men. This loss in impact resistance is not largebut occurs consistently and is particularly evidentin decreasing the level of ductile fracture energy.It apparently results from the stress concentra-tion at the junction of the raised weld depositand the smooth contour of the specimen. Theseresults indicate the necessity of matching theheat treatment response of the weld deposit withthe casting as closely as possible to obtain equiv-alent strength and machining the weld deposit tofollow the uniform contour of the part for goodimpact resistance.

The influence of severe welding discontinuitiesvaries to some extent depending upon the type of

impact test. The loss in impact resistance, as

39


40/52

measured by both the transition temperature andlevel of ductile impact resistance, is less for in-complete penetration than for undercuts or slaginclusions with both tests. However, the decreasein impact resistance from incomplete penetrationis significant for both bending and tension impact,indicating the undesirability of utilizing weldedsteel with this discontinuity for the more severe

impact applications. Severe undercut discontinui-ties produce the most serious loss in impact re-sistance, as measured by both criteria, when thebending stresses occur, since this is a surfacediscontinuity. Severe slag inclusions result

impact properties of cast steel

Documents