1–2/2002AvestaPolarit Corrosion Management and Application Engineering

acom

1–2/2002 1(24)

Fatigue behaviour of stainless steel welds Page 2

A new lean duplex stainless steel for construction purposes Page 17

Dear Reader,

There is in Swedish a rather stupid saying, which translated into English becomes

something like “Anybody waiting for something good can never wait too long”.

It may have had some bearing in the past when time was in surplus but it

surely doesn´t have in a modern society. The keywords today are “just in time”,

“time shearing” and similar and waiting for something is just frustrating.

Anyhow, we apologise for the time it has taken, we appreciate that you have had

patience, but finally acom is back on track again. The aim will be four issues

annually but due to the timing this year you will get two, but each will contain

two articles instead.

Enjoy the reading!

Jan Olsson

TECHNICAL EDITOR

1–2/2002 2(24)

acom

M. Liljas and C. EricssonAvestaPolarit AB, R&D, SE-774 80 Avesta, Sweden

Keywords: Stainless steel, Duplex, Fatigue strength, Fatigue design, Welds

Stainless steels are used more and more as a structural material due to their good mechanical properties combined with high corrosion resistance.One of the most frequent causes of failure in constructions, besides corrosion,is fatigue or corrosion fatigue. However, there are fewer data available on fatigue properties for stainless steels than for ordinary structural steels,which, to some extent could limit their use.

A standard austenitic stainless steel (316L) was compared with two typesof duplex steels (2205 and SAF 2304) using different welding methods and different material thicknesses. High cycle fatigue tests across the weldswere performed both with polished specimens and also with specimenshaving the weld reinforcement intact. The results confirmed that the duplexsteels have higher fatigue strength than austenitic stainless steels. The superior behaviour of the duplex parent material was to a great extentmaintained also in the weldments. Submerged arc welds (SAW) showedeven higher fatigue strengths than the base material. However, the duplexwelds produced with gas metal arc welding (GMAW) showed a slightreduction of the fatigue strength compared to the parent material. Corrosionfatigue tests in synthetic seawater environment (3% NaCl) at a frequencyof 1 Hz resulted in 25% reduction of the fatigue strength of 316L comparedto 10% reduction for the duplex grade 2205.

Results for butt welds (GTAW and GMAW) in 2205 show that the stressconcentration at the weld toe greatly influences the fatigue performance.The favourable geometry of the GTA weld profile gives exceptionally highfatigue strength. However, when stress raisers become more severe, e.g. load-carrying fillet welds, the fatigue strength drops significantly and becomes independent of the material’s static strength.

Nevertheless, stainless steels and especially duplex stainless steels showvery promising fatigue behaviour on the same level or better than ordinarystructural steel.

Fatigue behaviour of stainless steel welds

Background

Traditionally, stainless steels havebeen used mainly for corrosionprotection purposes in variousenvironments. For that reason thefocus has been to choose the optimal alloy from corrosionpoint of view. Corrosion data arenow readily available for a greatnumber of steel grades and corrosion environments. Also the effect of welding has beenthoroughly investigated and isreported in the literature.

One of the most frequent causes of failure in materials,including stainless steels, is (corrosion) fatigue. The failuresoften occur at stress raisers in ornear welds. Conventional stain-less steels have comparativelylow strength levels and notchsensitivity. For higher strengthgrades such as duplex stainlesssteels higher notch sensitivity isexpected needing more concernin design, particularly if thestrength shall be fully utilised.

For several reasons there is anincreased use of stainless steel as a construction material in lesssevere environments. The currentdesign rules for dynamic loadingof steels are based on data for C-Mn steels. It has been shownthat fatigue properties of stainlesssteels compare favourably withC-Mn steels resulting in a conservative design when using

1–2/2002 3(24)

acom

stainless steels. The objective with this paper is to shed somefurther light on the fatigue behaviour of stainless steels withspecial focus on welds and onduplex stainless steels. This paper will be limited to highcycle fatigue as such data aremore relevant for materialsdesign purposes than low cyclefatigue.

Introduction

FATIGUE PROPERTIES OF WELDSStainless steels have been subjected to extensive fatiguedocumentation during manydecades and there are plenty ofdata available [1–3]. As is the case with C-Mn steels the fatiguestrength for both austenitic andduplex stainless steels is relatedto the yield and tensile strengthlevels. The general rules valid forother materials can also be usedfor stainless steels. Thus a higherstrength material with higher fatigue strength will also haveincreased notch sensitivity. A smooth surface and freedomfrom defects (such as non-metallicinclusions) also have positiveeffect on the fatigue life of stain-less steels.

Many investigations haveshown that duplex steels, with a higher strength level than austenitic steels, have superiorfatigue strength [4–6]. In manycases the fatigue strength ofduplex steels is on level with thestatic yield strength. The effect of the duplex microstructure is to maintain the fineness of the

structure, giving a high strength.Several works have studied theinitiation stage mainly in straincontrolled fatigue tests. The initiation can occur in either or inboth phases depending on theconditions [7–8]. A concept ofmicrostructural barriers for shortcrack propagation can be used toexplain the role of the individualphases in duplex steel [9]. The general conclusion is thusthat the presence of two phasesseems to have a retarding effecton the initiation of fatigue cracks.

Concerning the fatigue behaviour of stainless welds thereis less published informationavailable. In cases where weldsactually have been tested there islittle or no information on thewelding process and weldingparameters. Such variables couldhave a large impact on propertiesincluding the fatigue behaviour.

For conventional austeniticstainless steels the weld metalgenerally has a slightly higherstrength than the base materialpartly because of a small contentof ferrite. Duplex steels show asimilar relationship when weldedwith the recommended fillermaterials with enhanced Ni content. However, other weldingprocedures can result in widelyvariable ferrite contents andstrengths. Depending on the welding method the oxygen leveland thereby the inclusion contentin the weld metal will vary considerably. For a high alloyaustenitic stainless steel weldedwith Ni-base filler, gas tungstenarc welds (GTAW) showed

markedly higher fatigue strengththan submerged arc welds (SAW)[10]. This result was attributed toa lower inclusion level in the former welds but possibly alsopresence of microfissures in the SAW. Different welding parameters can also give rise tovarying residual stresses andstrain in the weld area, whichshould have large effects on thefatigue results. Normally, fatigue testing of weldsis done with specimens removedtransverse to the weld seam. Thisapproach can be used to assessthe fatigue properties of the weldmetal as such if the specimen is machined to a smooth reducedsection. The limited informationgiven for such con-figurationsindicates that high quality welds show fatigue propertiescomparable with those of theparent metal.

FATIGUE DESIGN OF WELDEDSTRUCTURES The design part will cover fatigueperformance of stainless steeljoints in as welded condition incomparison with smooth specimen data for parent andwelded materials. Design data onfatigue of welded structural steelshas gradually been built up andrecommendations have been established by The InternationalInstitute of Welding (IIW) [11],Eurocode 3 [12], etc. Differenttypes of standardised structureshave been tested to cover different types of load cases suchas butt welds, fillet welds etc.Depending on the structure’s

1–2/2002 4(24)

acom

susceptibility to fatigue failure,each structure has been given afatigue class i.e. FAT class in theIIW designation system. The FATclass gives the allowable stressrange at 2.106 cycles. This methodis often refereed as the nominalstress approach.

Stainless steel is increasinglyused for structural purposes withthe consequence that engineersare looking for fatigue designdata of stainless steel in the same way as they do when using structural steel. However, thedesign guidelines available, e.g.Eurocode or IIW, are built upondata of C-Mn steels only. To makeit simple, a first assumption is touse the guidelines for C-Mn steelswhen using stainless steels. The outcome of this approach is probably satisfactory in most cases,because the fatigue behaviour ofwelded joints is dominated by thejoint geometry and similar crackgrowth behaviour occurs in C-Mnand stainless steel. However, theintention behind these guidelinesis to provide structural integrityfor the used material and thelevels of the fatigue classes arederived from statistical evaluationof a great number of tested speci-mens from different laboratoriesand welding procedures etc. anda certain confidence against fracture has been established. Theweakness in using the guidelinesfor the nominal stress approach isthe question of whether or not thereal structure can be idealised forcomparisons with a standardisedgeometry e.g. butt weld, filletweld. Loading direction, welding

and will create scatter in the data.The weakest spot along the weld,even if it were just a single imper-fection, would most likely initiatethe crucial crack. Therefore, fatiguetesting at different laboratorieswith different materials and welding techniques etc. is requiredbefore it is meaningful to estab-lished fatigue design guidelinesfor stainless steels. IIW has norecommendations for stainlesssteel at present time. Eurocode 3has included stainless steels in thelast draft. Stainless steels areregarded as equivalent to C-Mnsteels.

There are not many studies carried out on fatigue performanceof welded stainless steels andespecially not for duplex stainlesssteels [13–17]. Some recent datawill be presented in this paper,however it is not a complete coverage of the published data inthe literature until today. It willanyway give indications on thefatigue levels for welded stainlesssteels.

Materials

One austenitic (316L or EN 1.4404)and two duplex (UNS S32304 orEN 1.4362 and UNS S32205 or1.4462) stainless steels were tested.SAF 2304 is a trademark ofSandvik Steel. For simplicity theduplex grades are listed as 2304and 2205 in this paper. The compositions are given in Table 1.The 316L material in 20 mm thickness had slightly higher Moand Ni contents than the thinner3 mm material. Commerciallyproduced plate and sheet material

procedures, residual stresses etc.have to be considered as well.Nevertheless, supplementarymethods to the nominal stressapproach have come into use toimprove the fatigue design processby the use of finite element analysis (FEA), the hot spot stressand the fracture mechanics approach. All these methods forfatigue design have their advantages and disadvantages.However, to get reliability in allthese methods it is essential tohave knowledge of the behaviourof the specific material. Stainlesssteel differs significantly from C-Mn steel on a metallurgicalscale, which in turn affects themechanical properties. Therefore,they should be tested in the same way as C-Mn steels to makeit possible to establish fatiguedesign recommendations, whichprovides structural integrity.Furthermore, stainless steel cannotbe looked upon as one singlematerial group. It is essential todivide stainless steel into at leastthree groups that differ significant-ly; ferritic steels, austenitic steelsand ferritic-austenitic (duplex)stainless steels.

Unfortunately, the limitednumber of fatigue data availableon welded stainless steel has ledto preconceived misunderstandingof the fatigue behaviour of stainless steels compared to C-Mnsteels. Individual sets of resultsshould be treated with caution.Variations in the quality of theparent material, and in testingprocedures as well as weldingprocedures influence the results

1–2/2002 5(24)

acom

in 20 and 3 mm thickness, respec-tively, were used. Tensile proper-ties transverse to rolling directionare listed in Table 2. The duplexsteels base materials contained 47 and 43 % ferrite, respectively.

In order to investigate theinfluence of weld type (with varying cleanliness) different welding methods were used(Figure 1). Gas metal arc welding(GMAW) and SAW were used for20 mm materials and GMAW and GTAW for 3 mm materials.For the fillet welds, flux cored arcwelding (FCAW) was used aswell. The welds were orientatedin the rolling direction. Type ER 316LSi-filler was used for the316L material, and Type ER 2209-filler for both duplex grades. Thewelding parameters are shown inTables 3a and 3b. Tensile testingwas performed both longitudinally(i.e. the weld metal) and transverseto the rolling direction of the welded hot rolled material. Theresults were used to relate to thefatigue data and are not displayedin this report. Welds and basematerials were thoroughly characterised concerning chemicalcomposition, microstructure andhardness, but all data will not bereported in this paper.

316L, 20 mm316L, 3 mm2205, 20 mm2205, 3 mm2304, 20 mm2304, 3 mm

0.0210.0210.0200.0170.0210.015

C Si

0.60.50.40.40.40.5

0.81.11.61.61.51.6

Mn

0.0280.0310.0200.0220.0210.020

P

0.0010.0010.0010.0010.0010.001

S

17.717.622.122.222.922.8

Cr

12.611.25,85.84.74.9

Ni

2.62.13.03.00.20.3

Mb

0.040.060.180.160.100.09

N

Table 1: Chemical composition of tested alloys

Alloy, thickness

Table 2: Mechanical properties of tested parent materials, transverse direction

280279507607450537

545738353731

A 5%

143146227250208230

HV5

578582759835672729

323309582700508604

Alloy, thickness

316L, 20 mm316L, 3 mm2205, 20 mm2205, 3 mm2304, 20 mm2304, 3 mm

R p0,2

MPaR p1.0

MPaR mMPa

Stainless Steels

FerriticMartensitic

Ferritic-AusteniticDuplex

Austenitic

Cold Road Sheet(3mm)

FCAW

TransverseLoad-

carryingfillet welds

Hot Rolled Plate(20mm)

GTAW GMAW SAW

Butt welds

Figure 1. Test matrix showing the welded stainless steels used in this study

Fatique testingsmooth specimens

Welded smooth specimens (except FCAW)

Welded details of cold relled sheet

acom

1–2/2002 6(24)

Experimental

In this study both smooth and aswelded specimens according toFigures 1–2 were fatigue tested.Pulsating fatigue testing was carried out in servo hydraulic test-ing machines in air at a frequencyof 20 Hz and the stress ratio R=0.1.The test specimens were takenacross the rolling direction. Thetesting was performed in air, at20°C. Corrosion fatigue testswere performed at a frequency of1Hz and a stress ratio of R=0.1

in 3% NaCl solution. The fracturesurface of the test specimenswhich failed before 2.106 cycles, attheir respective stress levels, werestudied in Scanning ElectronMicroscope (SEM).

For the smooth specimens fatigue results were evaluatedusing the staircase method (about 30 specimens). The fatiguestrength (RD), standard deviation(s), and RD/Rm were calculated.The as welded specimens wereevaluated using SN-curves.

Results and discussion

WELD PROPERTIES Fatigue test results

Parent material: The ratio betweenmeasured fatigue and tensilestrength for all 20 mm parentmaterials was roughly 2/3(RD/Rm ratio in Tables 4–5. Thusthe duplex steels, having a higherstrength, show approximately 20–40% higher fatigue strengthcompared to the austenitic steel.The 3 mm duplex materials have higher tensile strengths also

Table 3b: Welding parameters for butt (B) and fillet (F) welds

Table 3a: Welding parameters for smooth butt welds

1–2/2002 7(24)

acom

resulting in increased fatiguestrength. However, the austeniticmaterial only shows a marginallyhigher tensile strength at 3 mmthicness, but a clear reduction in

fatigue strength compared to thethicker material.

Welded material: All welds hadtensile strengths across the weldsimilar to the parent material. The

thicker welded material exhibitedequal or higher fatigue strengthscompared to the parent material.On the other hand, for the thinnermaterial, the welded duplex

*Std. dev. calculated at maximum fatigue stress, RDmax. ”–” Indicates either invalid step length or too few specimens tested for relevant calculation of std. dev.

Table 5: Fatigue test results, CR sheet, 3mm

*Std. dev. calculated at maximum fatigue stress, RDmax. ”–” Indicates either invalid step length or too few specimens tested for relevant calculation of std. dev.

Table 4: Fatigue test results, HR plate, 20 mm

–

also the highest fatigue strength.The GMA welds (2205 and 316L)have higher hardness comparedto parent material but still thelowest fatigue limit values (lowerthan both parent material and SAwelded material). However, noconsideration was taken to whether or not the crack propa-gated through the weld or theparent material, when the fatiguestrength was calculated.

Corrosion fatigue test results

The results in Table 6 show thatthe fatigue strength decreases10% for the duplex 2205 gradeand almost 25% for the austenitic

materials showed reduction offatigue strength compared to theparent material, with GMA weldshaving the lowest fatiguestrengths. The austenitic steel hadcomparable fatigue strength forwelds and for parent material inboth gauges with the lowest values for the thinner material.

There is a good correlation between hardness and fatiguestrengths for the parent material.However, the fatigue data of welded 20 mm materials showonly a weak correlation to hard-ness measurements. The SAWwelds showed the highest hardness values and they have

Figure 2c: Geometry and dimensions of the butt welded fatigue specimen, GMAand GTA welded CR sheet, 3mm

Figure 2b: Geometry and dimensions of the smooth fatigue specimen, parent and welded CR sheet, 3mm

10±0.10.2

16

R15

32

90

10–12

Figure 2d: Geometry and dimensions of the load carrying fillet welded fatigue specimen, GMA, GTA and FCA welded CR sheet, 3mm

*Std. dev. calculated at maximum fatigue stress, RDmax. ”–” Indicates either invalid step length or too few specimens tested for relevant calculation of std. dev.

Table 6: Corrosion fatigue test results, HR plate, 20mm

1–2/2002 8(24)

acom

Figure 2a: Geometry and dimensions of the smooth fatigue specimen, parent and welded HR plate, 20mm

30

3

3

30

1–2/2002 9(24)

acom

316L grade in synthetic seawaterenvironment. In the case of 316Lindications of corrosion pits onthe fracture surfaces wereobservedin connection to surface defects asscratch marks or ferrite bands.Such pitting attacks were not asclearly visible on 2205, whichreflects the higher corrosion resistance of 2205. For SA weldedmaterials of 316L and 2205 thefatigue strength is higher than forthe parent materials.

Discussion

Fatigue tests on smooth specimenswith the weld reinforcementremoved and no notches presentgive information mainly on theinfluence of metallurgical factors.The fatigue data obtained in thiswork are in essence in accordancewith literature data [3]. Both steeltypes exhibit fatigue strengthlevels close to the yield strength.The fatigue strength obtained for3 mm duplex parent material ishigher than for 20 mm material.This result was expected becauseyield and tensile strengths arehigher for the thinner material,which is the result of a finermicrostructure. The lamella spacing is approximately 5mm forthe 3 mm material, compared toabout 25 mm for the 20 mm plate.

For the austenitic steel the thinner gauge material showedlower fatigue strength in spite ofa higher tensile strength. Alsohere the 3 mm material had a finermicrostructure; the grain size was18 mm compared to about 30 mmfor the thicker material. One reason for the unexpectedly low

fatigue strength could be connect-ed to Al2O3 particles from thespecimen preparation detected onthe fracture surface at the initia-tion site. This effect could well bemore pronounced for the softeraustenitic material.

The welded 20 mm materialsshowed fatigue strengths close toor above those of the parent mate-rials even though the defect levelin these welds seemed compara-tively high. No harmful effect ofmulti-pass welding could thus beobserved. The fracture analysis ofthe fatigue specimens showedthat cracks did not run preferenti-ally in the weld metal of SAwelds. For GMA welds in 2205,however, most cracks initiated inthe weld metal while 2304 showedthe opposite behaviour. The fatiguestrength for the welded 3 mmmaterials is lower than those ofthe parent materials especially for the duplex grades. The defectlevels in these welds were not larger than in the thicker material.One reason for the lower fatiguevalues could be differences inmicrostructure. Compared to thebase material these welds containclearly higher ferrite levels andhave a coarser microstructure. Forthe duplex steels the GMA weldhad slightly lower hardness thanthe base material. However, this difference did not result inpreferential cracking in the welds.For all 3 mm materials the surfacesof the specimens were coarselypolished and most of the fatiguecracks were initiated at scratchmarks in the HAZ, which couldbe more sensitive.

To summarise the results ontests of smooth welded stainlesssteel specimens, the fatigue lifereached the level of the thickerbase material irrespective of welding method. The duplexsteels have clearly higher fatiguestrength than the austenitic steels.The thinner duplex base materialshad higher strengths resulting inhigher fatigue values. Howeverthe welds did not meet these higher values. In fact, the thinnerwelds had even lower fatiguestrengths than the 20 mm welds.This could be due to different testsamples with a less optimal surface preparation. In the case ofthicker material, round-machinedbars were used. For the thinnermaterial, flat bars with fairlysharp edges were tested. Judgingfrom the thicker gauge materials,the present study did not showany substantial difference in per-formance of the welding methodsused. If welded with adequatecontrol they are fully comparableas far as fatigue is concerned.

In the corrosive environmentthe fatigue strength at 2.106 cyclesis lowered to a greater extent forthe 316L than for the 2205 whichis expected because of the bettercorrosion properties of the 2205.The fatigue strengths for the SAWwelded material of 316L and 2205are higher than for the correspon-ding parent materials.

Fatigue design

In this paper data of both austeni-tic and duplex stainless steels arepresented. Nevertheless, the focuswill be on duplex steels because

1–2/2002 10(24)

acom

very few data have been presentedin literature and fatigue propertiesof welded high strength steel isparticularly crucial in the designwork. Both heavy structuresmade from hot rolled plate (10–20 mm) and thinner structuresfrom cold rolled sheet (3 mm) will be covered for duplex steels.For austenitic steels only selectedresults from hot rolled plates (10–20 mm) will be shown. Thejoints considered (Figure 3) are:

• Transverse butt welds • Transverse load carrying fillet

welds• Transverse non-load carrying

fillet welds ref. data only [13-17]

• Longitudinal fillet welded attachments ref. data only [13-17] The different welding methods used in this comparison were, SAW, GMAW, GTAW and SMAW for the heavy gauges and GTAW, GMAW and FCAW for the sheets.

Transverse butt-welds

X-ray examination of the GTAand GMA welded specimens(Figure 2c) showed no traces ofpores, inclusions etc. in the weldmetal and the welds were alsochecked to confirm full penetra-tion. The results of the fatigue testing are shown in Figure 5, 26 specimens of GTA welds and16 specimens of GMA welds.Fractography revealed that themajority of he specimens hadtheir crack initiation in the middleof the specimen at the weld toe, confirming that the specimenedges were ground satisfactory to avoid crack initiation at theedges. SEM/EDS-analysis showedno traces of any slag inclusions at the initiation sites.

The fatigue results are surprisingly high for the GTAwelded specimens. The stressrange was about 0.85 of the smoothwelded specimens fatigue strengthat (2. 106) cycles (Figure 5). Thereare of course obvious reasons forthat. Firstly, the stress concentra-tion Kt factor is just 1.54 and the

Figure 3c: Geometry and application of load, fatigue design specimen, longitudinal fillet weld attachment (not used in this investigation, ref. data only)

Figure 3d: Geometry and application of load, fatigue design specimen, transverse non-load carrying fillet weld (not used in this investigation, ref. data only)

Figure 4a: Schematic illustrations on cross sections of the GTA (on top) and GMA welded butt joints, Q indicates weld toe angle

Figure 4c: Schematic illustrations on cross sections of the GMA welded load carrying fillet joint, B indicates root opening

Figure 4b: Schematic illustrations on crosssections of the GTA welded load carrying fillet joint, A indicates incomplete fusion

Figure 4d: Schematic illustrations on cross sections of the FCA welded load carrying fillet joint, A indicates incomplete fusion

Figure 3b: Geometry and application ofload, fatigue design specimen, transverse load carrying fillet weld

Figure 3a: Geometry and application ofload, fatigue design specimen, butt weld

Q

A

A

▲

▲

▲

B

1–2/2002 11(24)

acom

weld toe angle about q=160°(Figure 4a), calculated by FEAwith the actual weld toe radiusobtained from cross section micro-graphs and imported to the FEAprogram. Secondly, no pores, slaginclusions or crack-like defectscould be found at the initiationsites. It is well known that GTAWgives smooth weld profiles andthe method is used for dressingthe weld toe to increase the fatigue strength of heavy weldedstructures. For the GMA weldedbutt joints there is a significantdrop in fatigue strength com-pared to the GTA welds. The Ktfactor is about 2.37 and q=132°(Figure 4a) and even though noimperfections are found at the initiation sites, the weld profile ismuch more irregular along the

weld pass than for the GTA-weld.The higher irregular stress con-centration along the weld pass isalso shown by the fact that theGMA welded specimens havemultiple crack initiation sites. For GTA welded specimens therewas usually a single or a double initiation site.

It is interesting to note that if the fatigue notch factor Kf is calculated (not shown in thispaper) and compared with the Kt factor, the geometry-inducedstress concentration totally controls the fatigue strengthreduction compared to smoothspecimen for both the GTAand GMA welded butt joints.

If these results are comparedwith the IIW guideline, buttwelds q<30°, the FAT class is 100

and both GMA and GTA weldedspecimens are well above thislevel. However, our results arefrom thin sheets and IIW data arebuilt upon heavy plate thicknessaround 10 mm. Razmjoo [2] indicated for austenitic and duplexstainless steels fatigue performancecomparable to that of C-Mnsteels, however, the data are scarce. Results reported for butt-welded joints in austeniticsteel [14] are shown in Figure 9.Kosimäki-Niemi [13] have shownthat the fatigue strength for buttwelds in a duplex stainless steel(similar to 2304) is better thanthat for an austenitic steel andwell above the FAT100 level(Figure 8).

It is also interesting to note that when using GTAW for butt

◆◆

◆◆

◆

◆ ◆ ◆ ◆

10000000

Number of cycles (N)

Stre

ss r

ang

e [M

pa]

1000000

Duplex 2205, CR sheet Smooth, parent materialSmooth weld, GTAWSmooth weld, GMAWButt weld, GTAWButt weld, GMAW

100000

1000

100

◆

◆

IIW FAT 100

◆

◆

◆◆ ◆ ◆

Figure 5: Fatigue results for parent and welded duplex 2205, CR sheet, 3mm

acom

1–2/2002 12(24)

◆ ◆ ◆

◆ ◆ ◆

◆ ▲

◆▲

◆◆◆◆

◆ ◆

Tr, load carrying fillet weld, GTAWTr, load carrying fillet weld, GMAWTr, load carrying fillet weld, FCAW

▲

◆

10

▲

◆◆

10000000

Number of cycles [N]

Stre

ss r

ang

e [M

pa]

1000000

Duplex 2205, CR sheet

100000

1000

IIW FAT 80

IIW FAT 45

▲▲

▲▲▲▲

▲

▲

▲▲

▲◆

◆◆

▲

▲

joining of duplex stainless steels,these results indicates that thehigher static tensile strength in aduplex steel compared to a regular austenitic steel actuallycontributes to a higher fatiguestrength. The common understand-ing for high strength steel is thatthe fatigue strength in a weldedjoint is independent of the staticstrength. Further testing has to be done to confirm if it is differentfor duplex stainless steel whenthe stress concentrations aremoderate.

Transverse load carrying

fillet welds

The SEM-analyses showed thatthe GTA, GMA and FCA filletwelded cruciform joints in 3 mmduplex 2205 (Figure 2d) had

incomplete fusion and root opening as a result from incorrectwelding data. Figures 4b–d illustrate these imperfections. Theopenings were most pronouncedfor GTA and FCA welded speci-mens. Nevertheless, the fatigueresults are in general very goodfor these joints (Figure 6) andthey are well above the compa-rable FAT classes 45–80 (leveldepending on welding procedureand quality control). The SEM-analyses did not revealed anyslag inclusions at the crack initiation sites for any of the welded joints.

The GTA welds cracked fromthe weld root through the throatat the high stresses and from theweld toe at the low stresses. Thetoe-initiated specimens had many

initiation sites along the fusionline, preferably at the edges of theGTA weld pulses. All the GMAwelded joints showed failurefrom the weld toe. Due to theconvex profile of the GMA weld(q=101°), the toe radius was smaller than the toe radius for theconcave weld profile of the GTAwelds (q=141°). The consequenceof this is that the stress concentration at the weld toe ofthe GMA weld is more severe and one would expect the fatigue strength to be lower thanfor the GTA weld. However, the result is not seen here. Theappearance of the GTA weld ismisleading. It looks smootherthan the GMA weld, however ona microscopic scale at the weldtoe close to a weld pulse edge

Figure 6: Fatigue results for load carrying fillet welds in duplex 2205, CR sheet, 3 mm

1–2/2002 13(24)

acom

10000000

Number of cycles [N]

Stre

ss r

ang

e [M

pa]

100000100000

Duplex 2205, HR plate

Smooth, parent materialSmooth weld, SAW and GMAWTr. non-load carrying fillet weld [15]Tr. non-load carrying fillet weld [15]Long, fillet weld attatchment [13]Long, fillet weld attatchment [15]Tr. load carrying fillet weld [15]

10000

1000

10

IIW FAT 45

▲

✕

✕ ✕

✕✕

✕✕

✕✕

✕ ✕

◆

▲ ▲ ▲ ▲

▲ ▲▲

▲ ▲▲

▲

◆◆

◆◆◆ ◆

◆◆

◆◆

◆

▲

IIW FAT 80

the radius is relatively sharp. Furthermore, the FCA welded

joints showed that although theopenings at the weld root, thecrack initiated in the weld toe.The FCA weld profile is concave(q=146°), resulting in the highestfatigue strength at a high numberof cycles among the comparedwelding methods. However, at alow number of cycles the FCAwelds are among the worst. Oneexplanation could be that theFCA welds were welded in twosets giving slightly different geometry of the weld profile andbesides that, they should havebeen welded with the same procedure. Some unknown factorhas influenced the fatigue resis-tance of the joint. This variationshows the danger in looking at

single set of data. In this case it istherefore chosen to illustrate alljoints in the same diagram andjudge them as one single group.All specimens of this group ofload carrying fillet welds werebetter than FAT80. It should also be noted that theIIW guideline is built up from testing of heavy plate about 10 mm.

Maddox et al [15] has recentlypresented results on cruciformjoints in duplex 2205 (Figure 7)where he recommended fatigueclass 36 for a joint failing in theweld throat. An IIW recommeda-tion for C-Mn is 45 for a cruciformjoint with weld root crack. Theresults are within the scatter bandfor C-Mn. Thus, a fairly conserva-tive recommendation is given

presumably due to scarce dataavailable on this type of joint.

Raszmjoo [2] showed that datafor cruciform joints of austeniticstainless steels fall within thescatter band for C-Mn steels andMaddox [15] confirmed theseresults (Figure 9).

Transverse non-load carrying

fillet welds

Lihavainen-Niemi-Viherma [16]reported data on non-load carry-ing fillet welds in duplex 2205(Figure 9). Maddox [15] has alsoreported data and a recommendation for using fatigueclass 80, same recommendation as for C-Mn steels in as welded condition.

Rasmjoo [2] suggests that thejoint classfication for C-Mn is

Figure 7: Fatigue results for parent and welded duplex 2205, HR plate, 10–20 mm

acom

1–2/2002 14(24)

10000000

Number of cycles [N]

Stre

ss r

ang

e [M

pa]

100000100000

Austenitic 304/316, HR plate

10000

1000

10

IIW FAT 80Smooth, parent material (316)Tr. load carrying fillet weld [15]Tr. non-load carrying fillet weld [15]Tr. non-load carrying fillet weld [16]Long, fillet weld attachment [16]Long, fillet weld attachment [15]Smooth, parent material [304] [14]Butt weld [14]

▲

▲▲▲

▲▲

▲▲ ▲ ▲

▲

▲▲▲▲

✕

✕ ✕✕

✕✕

✕ ✕

✕ ✕✕

✕

◆

◆◆◆ ◆◆ ◆

◆ ◆

◆

▲▲

✕

Figure 9: Fatigue results for parent and welded austenitic steel 304/316, HR plate, 10–20 mm

Figure 8: Fatigue results for duplex 2304 and a similar (23Cr–4Ni) material, HR plate, 10–20 mm

10000000

Number of cycles [N]

Stre

ss r

ang

e [M

pa]

100000100000

Duplex SAF 2304, HR plate

10000

1000

10

IIW FAT 100

IIW FAT 63

Smooth, parent material Butt weld [13]Long, fillet weld attachment [13]▲

▲ ▲▲

▲ ▲ ▲

1–2/2002 15(24)

acom

appropriate to use for austeniticsteel based on the limited dataavailable. Data for austeniticsteels are shown in Figure 9.

Longitudinal fillet weld

attachments

Data reported by Koskimäki-Niemi [13] for duplex steel (simi-lar to 2304) indicate god resultswell above the recommendationsfor C-Mn, FAT63, for attachmentlength <300 mm. Maddox [15]and Viherma [16] reported datafor duplex 2205 and they recommend the FAT71, i.e. thefatigue class when attachmentlength <150 mm for C-Mn steel.

Rasmjoo [2] concluded fromreported data that austeniticsteels give somewhat lower performance than C-Mn steel.However, data reported byMaddox [15] and Viherma [16] do not confirm this conclusion(Figure 9). Data are well abovethe FAT71 class.

Concluding remarks on fatiguedesign.

Until there is enough data toestablish fatigue design guide-lines for stainless steels therecommendations for C-Mnshould be used. Even though, theresults considered in this papershow superior fatigue strength to the fatigue classification forC–Mn steels (e.g. IIW, Eurocode3), scatter in data have to be considered, as for C-Mn steels,which lower the recommendedstress range. For further work it isalso important to consider fatiguelife improvement techniques e.g.TIG-dressing, influence of residual

stresses and testing in corrosiveenvironment.

As a comparison to the nominalstress approach results presentedin this paper, Partanen and Niemi[17] reviewed hot spot stressapproach results of welded detailsof both C–Mn and stainless steels.They concluded that fatigue classFAT100 or higher can be used asthe design hot spot fatiguestrength for toe failure of weldedjoints i.e. butt welds, transversefillet welds, the ends of longitud-inal fillet weld attachments ofmoderate thickness (<10 mm).

Conclusions

SMOOTH SPECIMENS• Fatigue and corrosion fatigue

properties have been assessed on smooth fatigue specimens for parent materials and welds of austenitic and duplex stainless steels.

• In air, the fatigue strength resultsshowed a marked correlation with the tensile strength with a factor 2/3, i.e. close to the yield strength of the materials.

• Welds had a small effect on the fatigue strength for all steels in one set of tests (20 mm material).Multi-pass welds (GMAW, SAW)were at least on par with the base material for fatigue strength.

• For the thinner material (3 mm) the austenitic steel showed lower fatigue strength in spite of a higher tensile strength. The duplex base materials exhibited high fatigue levels. However welding with GTA, and in

particular GMA, resulted in lower fatigue strength level than the base material.

• The welding methods tested appear to give comparable fatigue performance.

• In seawater environment (3% NaCl), the fatigue strength decreases 10% for the duplex 2205 grade and nearly 25 % for the austenitic 316L grade. The SAW welded materials were more resistant to corrosion fatigue than the parent materials.

FATIGUE DESIGN• Results for butt welds and fillet

welds have been included in design SN-curves showing the relation to C-Mn steels.

• Fatigue tests of welded structure show that stainless steels have superior fatigue life compared to the fatigue classification for C-Mn steels

• GTA butt welds of duplex stain-less steels show surprisingly high fatigue strength

• The weld profile appears to have less than expected influ-ence on fatigue performance forload carrying fillet weld of thin material (3 mm).

Acknowledgement

Parts of this research work(smooth welds) received financialsupport from European Coal SteelCommission (ECSC). Diplomaworkers, Magnus Piva and Ann-Jeanette Merilä, carriedthrough the testing of as weldedspecimens.

References

[1] Y Bergengren, Fatigue properties of stainless steels in air at room temperature a literature survey, Swedish Institute for Metals Research, IM-2900, 1992

[2] R Razmjoo, Design guidance on fatigue of welded stainless steel joints, TWI Document, 1994

[3] ASM Handbook, Vol. 19, ”Fatigue and Fracture”, ASM International, 1996

[4] W Wessling, H E Bock, Stainless Steel 77, 1977, p. 217

[5] J E Truman, G A Honeyman, Proc. 11th ICC, Vol. 5, Florence, 1990. P. 553

[6] L Coudreuse, J Charles, 6th World 2000 Duplex Conference, Venezia, 17-20 Oct., 2000, p. 629

[7] T Magnin et al, Low cycle fatigue, ASTM STP 942, 1988, p. 812

[8] S Degallaix et al, Fatigue Fracture Engineering Mater. Struct. Vol. 18, No1, 1995, p. 65

[9], J Stolarz, J Foct, 6th World 2000 Duplex Conference, Venezia, 17-20 Oct., 2000, p. 639

[10] R Visnu et al, Stainless Steel World 99 Conference, The Hague, 16-18 Nov. 1999, p. 131

[11] A Hobbacher (Ed), Fatigue design of welded joints and components, International Institute of Welding, 1996

[12] Eurocode 3: Design of Steel Structures, Part 1, General Rules for Buildings, Commission of the European Communities,1990

[13] M Koskimäki, E Niemi, Applications of Stainless Steels ’92, 9-11 June, 1992, Stockholm, Vol. 2, p. 889

[14] Data sheet on fatigue properties for butt welded joints of SUS304-HP (18Cr-8Ni) hot rolled stainless steel plate, NRIM Fatigue data sheet No.53, 1998

[15] S Maddox, C Branco, C Sonsino, Fatigue design of welded stainless steels, ECSC Information day, Stainless steels – New products and developments, Seville, Spain, October 1998

[16] V Lithavainen, E Niemi, R Viherma, ECSC Project – Development ofthe use of stainless steel in construction, Contract no. 7210 SA/904, 2000

[17] T Partanen, E Niemi, Hotspot stress approach to fatigue strength analysis of welded components: Fatigue test data for steel plate thickness up to 10 mm, Fatigue Fract. Engng Mater. Struct. Vol.19. No.6, p. 709–722, 1996

acom

1–2/2002 16(24)

1–2/2002 17(24)

acom

Pelle Johansson and Mats LiljasAvestaPolarit AB

Introduction

Stainless steels are used predomi-nantly for their corrosion resistancein moderate to highly aggressiveenvironments. For constructionpurposes carbon steel normally hasbeen the main choice of materialdue to low cost, long experience,applicable design rules and alarge variety of strength classes.However, different stainless steeltypes can also provide a very widerange of mechanical propertiesand they have the apparentadvantage of no need for surfaceprotection. Examples are lamp-posts, handrails, storage tanks,vessels, etc.

Duplex stainless steels in par-ticular, with twice the mechanicalstrength of conventional austeniticand ferritic stainless steels, have apotential for use in constructions.Since the introduction of thisfamily of steels about seventyyears ago these steels have beenused in many structural applica-tions. However, for various reasons, the primary motive forthe selection a duplex grade hasbeen towards the higher end ofits excellent corrosion resistancein combination with its highmechanical strength. This is notsurprising since they all contain a high level of chromium withinherently a good corrosion resistance.

The duplex grades have beenimproved during recent years (1).In the early 1980’s a ”secondgeneration” of duplex steels wasintroduced with improved weld-ment properties mainly throughnitrogen alloying (2,3). The mostcommon duplex grade today isEN 1.4462 (UNS S31803/S32205)and this steel is used in a greatnumber of applications in a widevariety of product forms. Parallelwith the development towardshigher alloy duplex grades forcorrosive conditions there hasrecently been a great interest inleaner compositions for construc-tion purposes. The best knowncommercial leaner duplex steel isEN 1.4362 (UNS S32304) withabout 23% Cr and 4% Ni. As anexample, this steel has been successfully used as a constructionmaterial of blast and firewalls onoffshore platforms and in bridges(4). An apparent and frequentlyused way to further reduce thecost is to reduce the nickel contentand compensate with manganeseand nitrogen additions. Low nickel content will also reduce theinfluence of price fluctuations ofraw materials. Such conceptswere described already in 1970’sbut to our knowledge no commercial steel was produced atthat time (5). In former SovietUnion several duplex steels with

low nickel and high manganesewere listed in 1970’s but therewas no information regardingtheir use. In 1989 an Armcopatent was published on leanduplex steel thermally stableagainst transformation to marten-site and with excellent as-castproperties. Thin-walled castingsfor automotive applications werementioned (6). A commercial steelaccording to this invention,Nitronic 19D (UNS S32001), con-taining essentially 20% Cr, 5% Mn,1.1% Ni and 0.13% N, was laterproposed as modular frame material in a new automotiveconcept (7). Nitronic 19D has alsobeen used for longitudinally welded umbilicals (8). Much workwith low nickel duplex gradeshas been performed in SouthAfrica. High manganese variantswith some molybdenum and copper additions showed bettercorrosion resistance than EN 1.4462 in sulphuric acid (9). A spun-cast low nickel grade(RelyNite) with 21% Cr, 7% Mnand 0.35% N was developed forpit props in underground mining(10). Lean duplex alloys with lownickel content and manganeseaddition having a metastableaustenite can be produced withinteresting mechanical propertiesafter deformation (11,12).However, as with temper rolled

A new lean duplex stainless steel for construction purposes

austenitic grades there is a risk ofreduced strength of the weldedarea. Work with nickel-free, highmanganese, high nitrogen duplexsteels for structural engineeringapplications showed promisingproperties for a steel compositionwith 22% Cr, 10% Mn and 0.3% N(13).

As shown above there havebeen several low nickel, manganese-alloyed duplex stain-less steels presented during thelast decade. Depending on theproperty requirements, differentalloying philosophies have beenused. One important feature isthe stability of the austenite.Transformation to martensite canresult in very high mechanicalstrength but can also producesensitivity to cracking under certain conditions. The objectiveof the present work was to deve-lop a thermally stable, low nickel,general-purpose duplex gradewith a corrosion resistance comparable to that of 1.4301 withundiminished weldment proper-ties in the as-welded condition. In this paper a new grade,AvestaPolarit LDX 2101, meetingthose requirements, will be presented.

Experimental

Pre-study of lean duplex alloysstarted with thermodynamic calculations for construction ofphase diagrams and comparisonby examinations of microstructurein small 0.5 kg ingots. The pre-study indicated that the mostinteresting ranges of main alloyelements in a lean duplex stain-

less steel were, C 0.03–0.05%, Mn 4–6%, Cr 21–22%, Ni 1.0–1.5%, Cu 0–1% and N 0.20–0.25%. Trials withthis range of compositions wereperformed on 30 kg laboratoryheats to further optimise the properties.

LABORATORY HEATSEvaluations of hot and cold rolledplates were performed on materialproduced from the 30 kg ingots.The aim was to find a suitablecomposition that combinedmechanical properties, corrosionresistance, structural stability andreforming capacity of austeniteafter thermal cycling, e.g. weldingoperations. The optimised com-position based on the laboratorytrials is shown in Table 1. Thesteel designation is AvestaPolaritLDX 2101

INDUSTRIAL HEATSSeveral full-scale heats of LDX2101 were produced, both innearly square (bloom) sectionaimed for evaluation of long

products as rod and bar and invarious slab sections for manufac-turing flat products such as plate,coil and sheet. All heats were conventionally produced, i.e.,scrap re-melting in an electric arc furnace, refined in AOD/CLUconverters and continuously cast.The weight of the casts was 75 to90 tonnes. The casts made forlong products were processed viaa CLU (a proprietary processusing steam-oxygen decarburi-sation) vessel.

The final products within thetrials of long products includedbar, rod and reinforcement bar indifferent dimensions. Most ofthose products were not subjectedto solution annealing. In thispaper, therefore, presentation ofproduct properties is confined tovarious flat products. Annealingtemperature was in the range1020 to 1100°C followed by rapidcooling.

Different types of final products from the trials of coilsand plate manufacturing are presented in Table 2.

acom

1–2/2002 18(24)

Table 1: Typical chemical composition for LDX 2101 in w-%. Fe bal.

Table 2: Flat product forms in the trials

acom

1–2/2002 19(24)

Table 3: Typical mechanical properties at 20°C for different flat products ofLDX 2101

band (WHB) and cold rolledsheet (CR). Brinell Hardness measurements are also included. The materials are in a solution-annealed condition.

CORROSION TESTSIntergranular corrosion testsaccording to EN/ISO 3651-2 method A (Strauss) and methodC (Streicher) were made on severalLDX 2101 products. None of thesamples failed in the tests. Theseresults are as expected for duplexsteels, which are less susceptibleto this kind of corrosion thanaustenitic stainless steels.

All flat products were also tested regarding chloride pittingcorrosion resistance by using the

Industrial trials of manufacturingof longitudinally welded pipesfrom 9 and 16 mm thick hot rolledplate were also performed, as wellas rectangular hollow sectionsmade from coil with a thicknessof 3 mm.

Results

MICROSTRUCTURE A balanced chemical compositionin LDX 2101 results in a micro-structure containing approximatelyequal amounts of ferrite andaustenite. This is obtained afterannealing at a temperature ofabout 1050°C. Due to its relativelylow alloying content of substitu-tion elements, this lean grade isless sensitive to segregationduring solidification than standardduplex and superduplex grades.The relatively low chromium andmolybdenum contents make precipitation of intermetallic phases more sluggish than in con-ventional duplex steels. This canbe shown by the influence of isothermal heat treatment on theimpact toughness. In Figure 1, a Time–Temperature-Charpy V(T-T-ChV) diagram for twoduplex grades are shown. Themost sensitive temperature rangefor LDX 2101 is 600 to 750°C. At the nose temperature, ~650°C,about 10 hours isothermal annealing is needed to reduce theimpact strength to 50J and almost100 hours to 27J. For 1.4462 theembrittlement transition is muchfaster with a nose temperature of about 850°C and only about 15 minutes to reduce the impacttoughness to 27J.

The position of the nose, regularlyassociated with intermetallicphase precipitation, is moved to a lower temperature compared tothat for regular duplex steels.Lower temperature means reduceddiffusion and thereby a retardingeffect on precipitation. Figure 1shows that LDX 2101 has a farbetter structural stability than thestandard duplex stainless steelssuch as 1.4462.

MECHANICAL TESTAll products were tensile testedin accordance with the standardprocedures in EN and ASTM. InTable 3 typical values for differentflat products are presented; hot rolled plate (HRP), white hot

Figure 1: GT-T-ChV-diagram for 1.4462 and LDX 2101 plate materials (15 mm).

Avesta cell (ASTM G150). Theoutcome in the determinations ofcritical pitting temperature (CPT)varied between 15 and 22°C,depending on the surface condi-tion. This level is higher thanwhat is normally obtained withEN 1.4301 and more on par withthe results for EN 1.4401.

A number of different testmethods were used to assess thesusceptibility of LDX 2101 tostress corrosion cracking (SCC).Specimens were stressed accordingto the four-point loading practicein ASTM G39, and exposed to a3.6 M calcium chloride solutionas well as a 3 M magnesium chloride solution. In both casesthe tests were done at 100°C for500 hours. The specimens wereloaded to 60 and 90% of the proofstrength at 100°C. U-bend tests in a 3 M magnesium chloridesolution were also performedwith the bending parallel andperpendicular to the rolling direction. The temperature andduration were the same as for thefour-point tests. The LDX 2101-material passed all the SCC testswithout failure due to cracking.However, many samples in thetests got some kind of superficialuniform corrosion attack.

Uniform corrosion tests in dif-ferent concentrations of sulphuricacid and temperatures were made.The result of the test is presentedin an isocorrosion diagram,Figure 2, where the curves arerepresenting a corrosion rate of0.1 mm/year. At temperaturesabove the curves the corrosionrate is greater than 0.1 mm/year.

For reference the austenitic stain-less steel grades 1.4301 and 1.4436plus the duplex grade 1.4362 areincluded in the figure.

WELDING TRIALSManufacturing trials of longitudi-nally welded pipes and tubeswere made. Pipes with the sizeOD 273 x 9 mm and OD 273 x 16mm and length of 6,000 mm weremade. Welding methods for thelongitudinal pipe welds wereplasma arc welding (PAW), gastungsten arc welding (GTAW).For the thicker wall, submergedarc welding (SAW) was used tofill up the joint. All weld beadswere performed with filler of2209-wire except for the first key-hole PAW-pass. The pipes werepost-weld solution annealed at1050–1070°C and pickled inmixed acid. Tensile samples takentransverse to the weld joints weretested. All ruptures on the tensilesamples occurred in the parentmaterial. The ultimate tensile

strength was 30 to 40 MPa higherthan for the parent plate material.1.4362 are included in the figure.

Impact toughness testing,Charpy-V, was performed trans-verse to the welds and with thenotches placed in the weld metaland in the parent material next tothe weld metal. The test tempera-ture was –50°C. The impact resultfor the pipe with 16 mm wallthickness, welded with SAW was47 J in the weld metal and 54 Jnext to the weld. The impactvalues are on the same level asthe base material as shown inFigure 3 with the ductile to brittletransition curve for the platematerial before manufacturing ofthe pipes. Compared with otherduplex stainless steels with nickelcontents in the range of 6 to 8%,this low nickel grade appears tohave a lower upper shelf value.However, the transition beha-viour and lower shelf values aresimilar to that of standard duplexalloys.

acom

1–2/2002 20(24)

Figure 2: Isocorrosion curves for LDX 2101 and othersteels with a rate of 0.1 mm/y in H2SO4-solution.

acom

1–2/2002 21(24)

Manufacturing trials of rectangularhollow sections were made at StalaTube Oy in Finland. The hollowsection was produced out of coilin a continuous line for formingand welding to a section of 140x80x3 mm followed by cuttingto length. The continuous weldingin the line was performed withPAW keyhole technique, followedby a GTAW-dressing of the weldcap. The welding was made without filler and with a duplex2209-filler. No annealing of thehollow sections was made.

The weld ductility in the hollow sections was evaluated byU-bend tests. Samples weldedwith filler and without filler weretaken across the weld. 180°bending was performed withboth the weld cap and the weldroot out over a mandrel with adiameter of 16 mm. All samplespassed the test without cracking.

Microstructure examination ofthe welds in the hollow sections

showed that the reforming ofaustenite in the weld metal andin the heat affected zone is veryfast. The fast forming of austeniteis controlled by the nitrogenalloying in LDX 2101. Figure 4shows the microstructure of anautogenous PAW joint withoutany post weld heat treatment. Thetotal arc energy was 1.3 kJ/mm.The ferrite content was determinedby image analysis to 61% in theweld metal, 64% in HAZ and 49%ferrite in the parent material.

Practical implications

The property profile of LDX 2101shown in the previous sections isto great extent typical for asecond-generation duplex stain-less steel. With comparatively highnitrogen content the mechanicalstrength is on level with that oftype 1.4462 and also on level withthose of several quenched andtempered steels. The weldmentproperties, also due to the nitrogenlevel, seem to be excellent evenwhen welding without filler. Thisindicates that the material can bewelded without the close restric-tions necessary for many otherduplex grades. Ongoing work inthis area will produce more datain this respect.

Other specific features of thislean duplex grade are the veryhigh structure stability. Due to thelow nickel content this steel hassomewhat reduced impact tough-ness of the base material compredto standard duplex steels.

Altogether, the property profile of LDX 2101 shown in thispaper indicates that it has apotential for use in a large varietyof applications. However, for use

Figure 3: Impact toughness of LDX 2101. Ductile to brittle transition curve for 16 mmthick hot rolled plate and of seam weld in pipe produced of the plate.

Figure 4: Microstructure in an autogenous PAW joint in a rectangular hollow section of LDX 2101. Beraha etchant.

Figure 5: The Apaté-bridge at Sickla channel in the centre of Stockholm. Bridge made ofDSS (1.4462) under construction. (Created by Magnus Ståhl & Erik Andersson)

acom

1–2/2002 22(24)

as a stainless structural steel, further documentation is needed.Fatigue testing of parent materialas well as welded joints is onearea being evaluated at present. It is also important to get suchsteels included in constructionstandards.

Below, some potential appli-cation areas for LDX 2101 are discussed.

A fairly new area whereduplex stainless steels are used is the construction of bridges. The materials choice is based onthe high strength and low main-tenance and life cycle cost butstrong reasons are also environ-mental concern and aestheticvalues. An example where thishas been utilised in a new design,

with duplex steel as load-carryingstructural material, is shown inFigure 5.

Several other applications inthe civil engineering area are com-ponents needing high mechanicalstrength and corrosion protection,equipment that today is made ofhigh strength, galvanised or painted carbon steels.

Stainless steels, includingduplex, are already being used asreinforcement bar material inconcrete structures to improve theservice life in chloride-contamina-ted environments. Having a highstructural stability, LDX 2101 ispossible to produce in as hot rol-led condition with adequate pro-perty profile; high strength, goodcorrosion resistance.

Storage tanks and towers withmodest requirements on corrosionresistance is another examplewhere LDX 2101 has an advantagecompared to 1.4301 and 1.4401due to the high strength givingweight savings.

Duplex stainless steel is apotential material and used con-struction material for transportapplications. For train carriagescertain components where highstrength and low maintenancecost is required, LDX 2101 is agood alternative.

Other applications were LDX2101 has an advantage is in waterheaters and hot water tanks dueto the fact that duplex stainlesssteels have a high resistance tostress corrosion cracking.

acom

1–2/2002 23(24)

Conclusions

A new lean duplex stainless steel,AvestaPolarit LDX 2101, designed for construction purposesis presented.

• Full-scale manufacturing experience has been made with a great variety of products, including hot rolled plate, coils in different surface conditions, seam welded pipe, seam welded hollow section, bar, rod and reinforcement bar.

• Low nickel content combined with manganese and nitrogen additions gives the steel a high structural stability.

• Low nickel content also gives less sensitivity to price fluctuations.

• Mechanical strength is high, on a level with other duplex steels and also with age-hardening steels.

• The corrosion resistance of LDX 2101 is in general at least as good as 1.4301 and in most cases as good as 1.4401.

• The weldability, even without filler addition is very good giving excellent weldment properties in as-welded condition.

References

1. J Olsson, M Liljas, Corrosion 94,paper no 395

2. S Hertzman et al, Duplex Stainless Steels ’86, The Hague,26-28 October 1986, p. 257

3. M Liljas, R Qvarfort, ibid p.244

4. HL Groth, ML Erbing, J Olsson,Duplex Stainless Steels ’91, Beaune, 1991, p. 1255

5. US Patent No 3,736,131, May, 1973

6. US patent No 4,828,630, August, 1989

7. 5,38 mmJ B Emmons, J Douthett, Advanced Materials& Processes, 8/96, 1996, p. 23

8. J W McEnerny, Stainless Steel World 2001, The Hague, 13–15 Nov., 2001, p. 245

9. S Nana, M B Cortie, J. S. Afr. Inst. Min. Metall., vol.93, no 11/12, Nov/Dec, 1993, p.307

10. A van Bennekom, Stainless Steel World, December 1998, p. 54

11. C D Van Lelyveld, A Van Benekom, Materials Science and Engineering, A 205 (1996), p. 229

12. JM Hauser et al, Stainless Steel ’99, Science and Market, Sardinia, Italy, June, 1999, p.85

13. J Wang, Dissertation ETH No. 12649, Swiss Federal Institute of Technology, Zurich,1998

AvestaPolarit ABResearch and DevelopmentSE-774 80 Avesta, SwedenTel: +46 (0)226 810 00Fax: +46 (0)226 813 05

www.avestapolarit.com

ISSN

110

1–06

81

Tekn

isk

Info

rmat

ion

/ G

rap

hiu

m V

ästr

a A

ros

2001

acom is distributed to persons actively involved in process industry development and other areas where stainless steels are important.

All rights reserved.

If you have any queries regarding acom,

please contact me at jan.olsson@avestapolarit.com or by

telephone on +46 (0) 226 812 48, fax +46 (0)226 813 05.

Jan Olsson

Technical Editor, acomAvestaPolarit

acom