Embed Size (px)
Citation preview
Evaluation of Low-and Medium-CarbonNb-Microalloyed Plate Steelsfor Wind. Tower ApplicationsInNorth America, steel plates for the primary struc-_tural members of wind turbine towers are commonlyordered to ASTM A572/ A709 Grade 50 or E1 10025-2Grade S355 standards. For wind tower sites that mayexperience low ambient temperatures, the plate specifi-cations often require stringent low-temperature CharpyV-notch (CVN) toughness requirements, e.g., 30 ft-Ibsat -4 of, 42 ft-Ibs at -40°F or 25 ft-Ibs at -58°F. If platesare ordered to ASTM A572/ A709 Grade 50 standards, alow-carbon, high-strength, low-alloy (HSLA) steel is typi-cally employed to meet the specified toughness require-ments. However, plates ordered to EN 10025-2 GradeS355 with these low-temperature CVN requirements
often specifY a requirement for "normalized rolling"(e.g., +N delivery condition). Structural steel plates sup-plied in the normalized rolling condition require con-trol of the hot rolling process such that the final defor-
This work presents the findings of an
investigation that compared low- and medium-
carbon steels as they relate to weldability,
toughness and fatigue resistance in wind
turbine towers.
Keith Taylor (top left], principal research engineer; Richard Bodnar (top right],director - research and development; and Todd Nelson (bottom left], director -applications engineering, SSAB Iowa, Muscatine, Iowa, USA ([email protected],[email protected], [email protected]); Steven Jansto (bottom center],market development manager, CBMM-Reference Metals Co., Bridgeville, Pa., [email protected]);Henrietta Tsosie, Zhiyong Hu, Sheldon Mostovoy,and Philip Nash (bottom right], director, Thermal Processing Technology Center,Illinois Institute of Technology, Chicago, III., USA ([email protected], [email protected],[email protected], [email protected])
Thickness & Condition
Authors
mation is accomplished over a temperature range whichis intended to produce a microstructure and propertiesequivalent to that obtained in a normalizing heat treat-ment. Furthermore, steel plates supplied in the normal-ized rolling condition are, by definition, required tomaintain the specified mechanical properties follow-ing a normalizing heat treatment. Based on internalstudies conducted by SSAB Americas to evaluate thecapability of the company's typical low-carbon A572/A709 Grade 50 to comply with the normalized rollingdefinition of EN 10025-2, it was determined that theselow-carbon HSLA steels, which achieve their excellentbalance of strength and toughness principally via ferritegrain refinement and a reduced carbon level, did nothave sufficient capability to retain the specified tensileproperties after normalizing, as exhibited in Figure l.Based on these results as well as similar studies, it wasdetermined that in order to comply with the EN 10025-2normalized rolling defi-nition, a medium-carbonsteel must be employedto ensure that the speci-fied minimum as-rolledmechanical propertiesare retained after nor-malizing. This paper
oYield (ksi)
mTensile (ksi)
-- - --= - ~ - - -
- - - ---i - -
- - - - - - -- - - - - - -- - - - - - -
L- L- L-
CD CD f..- f..- a a a aN N co co ~ ~ LO LO<0 ~ •.... •.... .•.... .•.... C'! C'!0 0 0 0 .•.... .•....
AR Norm AR Norm AR Norm AR Norm
807060
'iii~ 50.cC, 40c:~ 30-(j)
20100
~~~~l~~~ _
Comparison of tensile properties of as-rolled vs. normalizedlow C-Nb HSLA steels (A572-50) of various thicknesses.Normalizing generally results in yield strengths that arebelow the required minimum of 50 ksi.
October 2011 • 127
4
Compact specimen for fracture toughnesstesting.
Figure 3
0.90.80.7
Zone III(high under most
conditions)
0.60.5
Zone I(low under most conditions)
0.4
MedC-NbSteels _
0.2 0.3
lowC-NbSteel •
Zone II(depends on conditions)
0.1
0.05
0.35
0.00o
~ 0.30
10.25 .-=~= 0.20=U= 0.15 .=of~ 0.10
1IIIimIIIII _0.40
Carbon Equivalent
Graville weldability diagram (susceptibility to HAZ cracking).
reports on a collaborative investigation which comparesthese alternate approaches, namely low C-Nb controlledrolled steel, a medium C-Nb normalized rolled steeland furnace-normalized medium C-Nb steel. The evalu-ation includes microstructural characterization, tensileproperties, impact toughness, fracture toughness anduniaxial fatigue. A review of available data on fracturetoughness and fatigue properties of structural steelssimilar to those investigated here revealed a surprisinglack of such data. Therefore, the authors hope that theinformation reported here will be useful to the windturbine design community.
Materials and ProceduresSteel Selection and Processing - Two basic steel com-positions, both suitable for wind tower applications,were selected for this investigation. One is a low-carbongrade (0.06% C) with an addition of about 0.03% Nb.The other is a medium-carbon grade (0.15% C) withan addition of about 0.02% Nb. Henceforth, thesesteels are referred to as "low C-Nb" and "mediumC-Nb," respectively. Steel samples were obtained fromproduction plates produced at SSAB Americas' facilityin Mobile, Ala. Mobile employs a fine-grain practice,
with a low sulfur content and calcium treat-ment for sulfide inclusion shape control. Castslabs 6 inches in thickness were reheated to atemperature high enough to dissolve all of theNb. Plates were rolled to a final thickness ofabout 0.75 inch. The rolling temperatures were
controlled to produce final microstructures consistingofa mixture offine polygonal ferrite grains and pearlite.In addition to the samples from as-rolled plates,
a medium C-Nb plate that had been normalized at1,650°F was also sampled for evaluation. This steel willbe referred to as "medium C-Nb normalized" or "medi-um C-Nb norm."The results of a product chemical analysis on each
production plate are provided in Table 1. This tablealso includes the calculated carbon equivalent (CE).The carbon content and carbon equivalent values placethe low C-Nb steel in zone 1 of the Graville weldabilitydiagram 1 (Figure 2), while the medium C-Nb steels fallwithin zone 2. Although weldability was not investigatedin this work, the low C-Nb steel can be expected to havebetter weldability than the medium C-Nb steels.
MetaUography- Specimens for light optical microscopyand scanning electron microscopy were prepared usingconventional methods. Longitudinal cross-sections wereexamined in the as-polished condition for inclusionevaluation as well as after etching with 2% Nital. Theinclusion content of each steel was determined in accor-dance with ASTM E2142, Method 1.2 Specimens were
mDII _Steel Compositions (wt. %)
Steel C Mn P S Si Cu Ni+Cr+Mo Nb CEI
Low C-Nb 0.06 1.27 0.011 0.001 0.34 0.27 0.41 0.031 0.35
Med C-Nb 0.15 1.39 0.012 0.003 0.22 0.20 0.32 0.019 0.43
Med C-Nb norm 0.15 1.32 0.011 0.003 0.21 0.22 0.23 0.014 0.42
ASTM A572-50 and A709-50 [max) 0.23 1.60 0.04 0.05 0.40
EN 10025-2 S355K2 (max) 0.23 1.70 0.035 0.035 0.60 0.60 0.45
ICE = C+Mn/6+[Cr+Mo+V]/5+[Ni+Cu)115
128 • Iron & Steel Technology
examined in a light optical microscope as well as inZeiss Ultra 55 and Zeiss EVO/MAI 0 scanning electronmicroscopes (SEMs).
(Eq.2)
v is the Poisson's ratio (0.3) andE is the Young's modulus (29,000 ksi).
October 2011 • 129
where
where
cry is the effective yield strength (i.e., the average of theUTS and 0.2% offset yield strength as determinedfrom a tensile test).
(Eq. 1)
The relationship between lIe and KIe for plane strainconditions is:
Typicalfatigue test specimen.
2. Allow the machine to ramp for 10 minutes or toa destination of 1 inch, without exceeding theupper limit incremental of 0.005 inch (beginning at0.01 inch).
3. Once the upper limit is tripped, hold at displace-ment (crosshead extension) for 5 minutes to allowfor load relaxation.
4. Unload to a negative displacement of 0.1 inch with-out exceeding a lower limit of 0.005. inch (upperlimit minus 0.005 inch).
5. Once the lower limit is tripped, the process beginsagain with an increased displacement which is deter-mined by the user while maintaining ASTM E1820requirements.
This procedure is repeated several times to obtaina J-~a curve. Once the program has completed, thespecimen is heat-tinted by placing on a hot plate for atleast 5 minutes or until the specimen turns blue. Thespecimen is then fractured by fatiguing in a closed-loopmachine.For each specimen, several incremental measure-
ments of initial crack length and final crack lengthare made using a comparator. The total thickness (B)and side-grooved thickness (BN) are measured usingthe comparator as well. The results of each test pro-vide information to determine unloading compliances,1 elastic and plastic areas. The ASTM analysis allowsthe vallie lQ to be determined using a set of offsets andcurve fits. For the value of lQ to be considered lIe' thewidth of the specimen and the initial ligament (bo) mustmeet the following requirement:
0.020
__ Unloading Compliance:y=668226x-371R2;0.99996
-- Unloading Compliance:r-673076x+232R2;().99991
o0.000
2,000
6,000
~.f 4,000
0.005 0.010 0.015
Displacement pn.]
Example of J-integral test data.
Tensile and Impact Toughness Testing - Full-thicknessrectangular tensile and CVN impact toughness speci-mens were prepared in both the longitudinal and trans-verse orientations and tested in accordance with ASTME83 and E23.4 The Charpy specimens were machinedfrom the plate quarter-thickness. Charpy specimens forthe medium C-Nb steels were full-size (10 x 10 mm)while those for the low C-Nb steel were 3/4-size (10 x7.5 mm), in anticipation of higher absorbed energies.The test results for the low C-Nb steel were convertedto a "full-size equivalent" (FSE) value by multiplying by4/3. CVN testing was conducted at various temperaturesfrom room temperature down to -100°F.
8,000
Fracture Toughness Testing - Compact specimensfor room-temperature single-specimen fracture tough-ness testing (I-integral) were prepared and tested inaccordance with ASTM E1820.5 These specimens weremachined in a T-L orientation and were 0.75 inchthick (essentially full plate thickness) with integral slotsfor the displacement gauge attachment (Figure 3).Specimens were fatigue pre-eracked in a closed-loopmachine under load control with R = 0.1. The maxi-mum load was about 6 kip at the onset of pre-eracking,and was reduced to about 4 kip as the target pre-cracklength of 0.9 inch was reached. Side grooves 0.0375 inchin depth were machined on the pre-cracked specimens.Testing was conducted with a CGS/Lawrence InstronScientific Corp. model 8500R closed-loop machinecontrolled by MTS Multipurpose TestWare software. Anexample of the data that is collected by this program isshown in Figure 4. The data collected are: displacement(crosshead extension), strain (the displacement record-ed by clip gauge) and load. The program runs througha simple set of repetitions as outlined as follows:
1. Preload and unload five times to determine anaverage unloading compliance of the pre-crackedspecimen.
Figure 7
.•
MnS inclusions in the medium (-Nb norm steel.
Figure 8
Fatigue testing setup.
IIIBIIIIIL- _
A typical (a-AI oxysulfide inclusion (low (-Nb steel).
level. Testing was carried out until failure or when 107cycles were reached.
Electron Image 1
0.50
0.00
0.50
Heavy
Globular 0
1.00
1.50
0.50
Thin
ResultsMetallography - The results of inclusion analysis aresummarized in Table 2. Overall, the internal cleanlinessof each steel is very good. Because of the low sulfurcontents of these steels, essentially no type A-elongatedMnS inclusions were detected in the SEM-based analysis.However, close inspection revealed the presence of verythin MnS inclusions in both of the medium C-Nb steels,
especially near the mid-thick-ness (Figure 7). In contrast, theextremely low sulfur contentof the low C-Nb steel (0.001%)resulted in essentially no elon-gated MnS inclusions. Eachsteel contains a dispersion ofspheroidal type D oxysulfideinclusions containing calciumand aluminum, an exampleof which is shown in Figure 8.The low C-Nb steel exhibits a0.00
0.00
0.00
Heavy
Silicate C
0.00
0.00
0.00
Thin
Inclusion Analysis ResultsSulfide A Alumina B
Steel Thin Heavy Thin Heavy
Low C-Nb 0.00 0.00 0.50 0.00
Medium C-Nb 0.00 0.00 0.00 0.00
Medium C-Nb norm 0.00 0.00 0.00 0.00
Uniaxial Fatigue Testing - Specimens for fatigue test-ing were prepared and tested in accordance with ASTME466.6 Hourglass specimens with a diameter of 0.3 inch(Figure 5) were machined from the plate mid-thicknessin a transverse orientation. All machining marks wereremoved by finishing the reduced section starting withan appropriate grit size and finishing with 600 gritsandpaper. Care was taken to ensure that all transversemachining marks were removed from the reducedsection of the test specimens. A laser micrometer wasused to measure the final specimen diameter (averageof three measurements). Fully reversed (R = -1) load-controlled fatigue testing was carried out with an MTS880 testing machine (Figure 6). A fan was suspendedfrom the upper specimen clamp to minimize specimenheating during testing. Specimens were tested at alter-nating stresses (Sa) up to about 54 ksi at a frequency of4 Hz, except at stress levels near the endurance limit(Se) when a frequency of 8 Hz or 10 Hz was used. Threedifferent tests were conducted at each alternating stress
130 + Iron &. Steel Technology
Figure 9'", -
EHT- 300W
WO- $Smm
(a2)
0.'. 23Feb201'o.aG-tOMount' qtrl SSAB
(b1) (b2)
(c1)
,...H
EHT. '.DOWWO- 8.8nm
StgMlA- SE2Mag. 10.00 I( X
(c2)
0.. :24F.b201.032.09 ••••••••5"". SSAB
Microstructure of the low (-Nb steel (light-optical (a1)) and SEM secondary electron images (a2). The microstructure ispredominantly a mixture of polygonal ferrite and pearlite. The mean ferrite grain diameter is about 7.2 Jlm and the pearlitecontent is about 7%. Microstructure of the medium (-Nb steel (light-optical (b 1)) and SEM secondary electron images (b2).The microstructure is predominantly a mixture of polygonal ferrite and pearlite. The mean ferrite grain diameter is about6.7 Jlm and the pearlite content is about 28%. Microstructure of the medium (-Nb normalized steel (light-optical (c1)) and
. SEM secondary electron images (c2). The microstructure is predominantly a mixture of polygonal ferrite and pearlite. Themean ferrite grain diameter is about 7.4 Jlm and the pearlite content is about 27%.
October 2011 + 131
1IImmIII _Tensile and -600 F CVN Results
Steel
Low C-Nb
Orientation
L
T
YS UTS EI. in 8 inches CVN @ -GO°F(ksi) (ksi) (%) (ft-Ibs)
.- '-"'-'~'. .H_---..,..,... .._ -63.3 74.5 29.7 2831.-.._,..,-_ ...~~-.. -65.2 75.6 28.1 2741
Med C-Nb normalized ------- ....--.--.
Med C-Nb
ASTM A572-50 and A709-50
EN 10025-2 S355K23
L
T
L
T
63.6
64.1
55.7
56.7
50 min.
50 min.
81.3
82.5
76.5
76.8
65 min.
68-91
21.9
23.3
28.3
27.6
18 min.
20 min.
76
31
179
97
25 min. LCVN at 10°F2
30 min. LCVN at -4 ° F
1Full-size equivalent.2ASTM A 709 requirement for zone 3 fracture critical tension components.316 < L; 40 mm.
somewhat greater number of these inclusions than themedium C-Nb steels.
Figures 9a-9c show light-optical and SEM images ofetched metallographic cross-sections. All microstruc-tures consist predominantly of a mixture of polygonalferrite and pearlite. In each case, the ferrite grain size(mean grain diameter) is about 7 11m.The low C-Nbsteel contains about 7% pearlite, while both of themedium C-Nb steels contain about 28% pearlite in abanded distribution.
Tensile Properties and Impact Toughness - Tensileproperties and Charpy absorbed energy at a test tem-perature of -60°F are provided in Table 3. A tempera-ture of -60°F is about the lowest test temperature forwhich designers of wind turbine towers have a tough-ness requirement. Overall, the tensile properties of eachsteel are similar for both test orientations. The as-rolledsteels have similar yield strength of about 63-65 ksi. Thetensile strength of the low C-Nb steel is about 75 ksi,while that for the medium C-Nb steel is about 82 ksi. The
most notable difference between these as-rolled steels isthe impact toughness. The -60°F absorbed energiesfor the low C-Nb steel are over 250 ft-Ibs, while valuesfor the medium C-Nb steel are less than 100 ft-Ibs. Thelow C-Nb steel also exhibits less toughness anisotropy(i.e., a smaller energy difference between longitudinalvs. transverse test orientations), presumably due to itslower sulfide-inclusion content.
Normalizing decreases the yield strength of themedium C-Nb grade from about 64 to about 56 ksi, anddecreases the tensile strength from about 82 to about77 ksi. A substantial increase in Charpy-absorbed energyis gained by the normalizing treatment, presumably dueto a strength-toughness tradeoff.
Figure 10 plots the CVN energy for each steel vs. testtemperature for transverse and longitudinal orienta-tions. The superior toughness of the low C-Nb steel isreadily apparent, with upper-shelf behavior extendingto temperatures as low as about -60°F. It should benoted that each steel meets the tensile and Charpy
1IIimIIlII _Longitudinal Tests Transverse Tests
o300
III 250.c~~ 200E'Gll:W 150
"CD€ 1000III.c
<CZ 50>()
0-150 -100
8
-50
b 8
o
n 300
III 250.c..•.-=~ 200E'Gll:W 150
"Gl€ 1000III.c<C
OMedC-Nb Z 50>
<> Med C-Nb Norm ()
050 100 -150
o
-100
o
-50
o 0
o 50 100
TestTemperature, OF
(harpy-absorbed energy VS. test temperature.
132 • Iron & Steel Technology
TestTemperature, OF
Measured Calculatedao af ao af
Specimen (in.) (in.) (in.) (in.)
0.95 1.04 0.93 1.03
2 0.93 1.02 0.90 1.03
3 0.95 1.05 0.92 1.02
• Fitted Datao Excluded Data
_._.- Power Fit-- Blunting Line- - - 0.008 in. Offset Line..... Exclusion Line
0.10 0.15 0.20 0.25
M[in.]
Ja K~%Diff %Error (in-lb/in2) 25 [Ja/cry) Valid? (ksil in.)
8.1 3.0 1.8 0.8 2,466 0.94 Yes 399
11.1 3.6 2.6 1.3 1,912 0.73 Yes 351
9.3 0.7 3.0 3.0 3,124 1.19 No
0.05
".i:'/ .'e
,1 I., .•.i I
./ ...i .'1 •
:i /:;~/., I
ti/:,/.,''I./~JQ= 2466 [in-lb/in2
].• :.i.,
10,000
9,000
8,000
7,000
6,000"'c~ 5,000
.~ 4,000.,3,000
2,000
1,000
00.00
Fracture toughness results for the low C-Nb steel: an example of a J-~a plot and a tabular summary of the testdata.
0.350.300.25
oo
o
• Fitted Datao Excluded Data
_._.- Power Fit-- Blunting Line- - - 0.008 in. Offset Line. . . .. Exclusion Line
0.20
oo
0.15
o
0.100.05
.,.j
.11j
,I.,j
.11j
,I.,.j
:'1j .:
:' / ." /':'
.j ";' /
":l1'- Ja=976 [in-lb/in2
]
o0.00
7,000
8,000
6,000
2,000
5,000
1,000
N~
C1i 4.000
~':::; 3.000
t.a [in.]Measured Calculatedao af ao af Ja K~Specimen (in.) (in.) (in.) (in.) %Diff %Error (in-lb/in2) 25[Ja/cry) Valid? (ksi/ in.)
0.95 1.29 0.92 1.24 8.8 22.1 2.5 3.6 976 0.37 Yes 251
2 0.94 1.28 0.92 1.23 9.2 21.5 2.3 3.7 741 0.28 Yes 219
3 0.96 1.27 0.94 1.23 6.5 21.2 1.7 3.4 875 0.33 Yes 238
Fracture toughness results for the medium C-Nb steel: an example of a J-~a plot and a tabular summary of thetest data.
October 2011 •. 133
0.250.200.15
o
• Fitted Datao Excluded Data
_._.- Power Fit-- Blunting Line- - - 0.008 in. Offset Line. . . .. Exclusion Line
o
0.100.05
./
.i:'.,
.i:'.,
.i,1.,
.i /e.:.;" ../.
.i /./ ./
.i /:' e/., /./-,'
.; ,I.i,! •
:'1:Ie:--Ja= 713 [in-lb/in2)
10,000
9,000
8,000
7,000
6,000"'~c~ 5,000
~ 4,000...,3,000
2,000
1,000
00.00
6.a [in.)
Specimen
2
3
Measuredao af(in.) (in.)
0.90 1.12
0.92 1.19
0.94 1. 16
Calculatedao af Ja(in.) (in.) %Diff %Error (in-lb/in2) 25lJa/ay)-"._--- ..~.~ - -0.92 1.07 7.1 5.0 2.3 4.3 713 0.27
..._--_ ..
0.92 1.07 7.9 1.8 0.3 10.1 1,294 0.49
0.92 1.07 8.6 2.7 1.7 7.8 924 0.35.... .- __ .... ~1IIl~ •••• ._a __ .~
.ft ••
Valid?
Yes
Yes
Yes
KI~(ksiNin.)
215
289
244
Fracture toughness results for the medium (-Nb normalized steel: an example of a J-~a plot and a tabular sum-mary of the test data.
requirements of the ASTM and EN specifications bycomfortable margins (Table 3).
Fracture Toughness - Three individual tests wererun for each steel, and the results are summarized inFigures 11-13. The results for the tests are valid for Krccalculation with the exception of one of the tests on thelow C-Nb steel. The low C-Nb steel exhibits Krc levels ofabout 350 and 400 ksi/"in, while values for the mediumC-Nb steels are in the range of about 215-290 ksi/"in.
Uniaxial Fatigue - Figures 14-16 plot alternating stressvs. cycles to failure for each of the steels. These figuresalso show linear fits to the individual data points. A com-plete SoNcurve was developed for each steel by suitableselection of an endurance limit based on the runouttest results. These SoN curves are plotted together inFigure 17 to facilitate comparison of the fatigue lifebehavior. Based on the results, endurance limits for thethree steels are as follows: low C-Nb - 44 ksi; mediumC-Nb - 39 ksi; medium C-Nb normalized - 35 ksi.
Low C-Nb Steel
S =88(N ),().0518• t
R2=O.953
Medium C-Nb Steel•'\
• Med C-Nb Steel-e> Did Not Fail
-- Linear Fit- - - 95% Confidence Limit
38
~0056
54
52
'<if 50~~'" 48(/)
g 46~i/544Clc16 42E.$ 40«
S =75(N )'()0345• f
R2=O.969
• Low C-Nb Steel-e> Did Not Fail
-- Linear Fit_._.- 95 % Confidence Limit
52
~ 50
(/)'"<Iien 48~i/5Cl.5 46(5E.$« 44
Cycles to Failure, Nt
SoNcurve for the low (-Nb steel. SONcurve for the medium (-Nb steel.
421,000 10,000 100,000 1.000.000 1E7 1E8
361,000 10,000 100,000 1,000,000
Cycles to Failure, Nt1E7 1E8
134 • Iron & Steel Technology
1III!maII _Mechanical Property Summary
YS UTS -60°F TCVN Upper-shelf Fracture tou~hness, Fatigue enduranceSteel (ksi) (ksi) (ft-Ibs) TCVN (ft-Ibs) Klc (ksi/ in) limit, Se (ksi)
Low C-Nb 65 76 270 280 375 44
Medium C-Nb 64 82 30 120 235 39
Medium C-Nb normalized 57 77 80 160 250 35
In general, fatigue crack initiation occurred at the speci-men surface (Figure 18), as expected.
DiscussionTable 4 provides a summary of the mechanical proper-ties (transverse) of the steels evaluated in this investiga-tion. Values for KIc represent the average of the two orthree valid test results reported earlier. As mentionedearlier, available fracture toughness data is limited for
-------
structural steels similar to those examined in this study.However, Wilson 7 has investigated the fracture proper-ties of a low-sulfur (0.003%) ASTM A588 Grade A steel.The A588 steels have mechanical properties and chem-istry similar to the current medium C-Nb steels, exceptthat Grade A is microalloyed with vanadium instead ofniobium. Wilson reported a fracture toughness of 420ksi/-Vin for the A588 steel, but the j-integral test wasnot valid for a KIc determination. A valid test would
-------
Cycles to Failure, Nt
SoN curve for the medium (-Nb normalized steel.
Medium C-Nb Norm Steel
lE8lE7
Q•• 'iN' ••.
.....- -::»- ...•.
""
""
"
.... .•."
10,000
,- - . Low C-Nb Linear Fit '_. _. _. -lEJ- •••
o Low C-Nb DNF•..• Med C-Nb Linear Fit'I Med C-Nb DNF
_. - - Med C.Nb Normalized Linear Fit 0
o Med C-Nb Normalized DNF
1,000
55
30
50
'iii 45=-C/)"vi<Il 40~enClc:'E 35E.Sl<i:
100,000 1,000,000
Cycles to Failure, Nt
Fitted SoN curve for all of the steels.
lE8lE7
S =96(N )-Q07ZZa t
RZ=O.823
1,000,000100,000
," "" ", ",.'.",,,
10,000
• Med C-Nb Normalized-e- Did Not Fail
-- Linear Fit- - • 95% Confidence Limit
54
52
50
48
'en 46
=-44~C/) 42vi<Il 40~en 38Cl.5 36CiiE.Sl 34<i:
32
30
1.000
Figure 18
Fatigue crack initiation consistently took place on the specimen surface (arrow) withductile crack propagation.
October 2011 .• 135
Contributions to yield strength for the various strengthening mechanisms.
I ~01 _
Figure 19 shows the individual con-tribution of each mechanism to theyield strength of each steel. It is evidentthat Nb(C,N) precipitation contrib-utes substantially (by about 10 ksi)to the yield strength of the low C-Nbsteel, modestly (by 3.6 ksi) to that ofthe medium C-Nb steel, and makesessentially no contribution to theyield strength of the normalized steel.Precipitation in the low C-Nb steeloccurs during plate rolling and subse-quent cooling, but at temperatures lowenough that precipitate coarsening islimited. The medium C-Nb steels havelower Nb contents, and this is part of
L1YSNb(C,N) = YSmeas' - YScalc
(Eq. 4)
to that reported by Chen, et al.,s for an ASTM A709Grade 70 steel.
To understand the differences in fatigue endurancelimits of the current steels, it is useful to consider thedifferent strengthening mechanisms at play in thesesteels. Strengthening mechanisms that pertain to thesesteels are the lattice strength of iron (base strength),solid-solution strengthening, pearlite strengthening,ferrite grain size (FGS) strengthening and precipitationstrengthening. It is possible to estimate the contribu-tions of these strengthening mechanisms by referring tothe literature and utilizing quantitative metallographyresults.9 Consider the empirical regression equation foryield strength of C-Mn steels developed by Grozier andBucher, shown in Equation 3.10
Irvine and Pickeringll considered additional solid-solution elements and proposed the following coef-
ficients for the elements Mn, Si, Cu,Mo, Ni and Cr: +4.7, +12.2, +12.2,+2.0, 0 and -4.5 ksi/wt. %, respec-tively. The Mn and Si coefficients forsolid-solution strengthening of Irvineand Pickering are similar to those ofGrozier and Bucher.
The yield strength contributionsfrom solid-solution strengthening varybetween about 11 and 13 ksi for thepresent steels. The contribution ofpearlite to strengthening is minimal forthe low C-Nb steels, but is about 6 ksifor the medium C-Nb steels. Grain sizestrengthening is similar in all steels,at28-29 ksi. Of the various strengthen-ing mechanisms, that which is due toprecipitation of Nb(C,N) is expectedto show the largest variation amongthe present steels. This strengtheningincrement can be inferred from thedifference between YScalc and the mea-sured yield strength, i.e.,
Med C-Nb NormMedC-NbLowC-Nb
----------3.6 ....
Nb(C/N) ••....
10.1........ -----.:*:,......•. u•.•...•...
FGS 29.4 2828.4
.". ... "" ...••- ---------.....-... 6.2 Pearlite 5.9u --------- ---------
13.4 Mn/Si11.6 11.4
CU, etc
--------- ---------13.3 Base 13.3 13.3
I
70
60
50-'iii=.oS: 40.•..lI.Oc::QI~.•..
VI 30"CQj>=
20
10
0
YScalc (ksi) = 13.3+ 5.9(%Mn) + 10.2(%Si) + 0.22(%Pearlite) + 2.4(1/-Jd)(Eq. 3)
d is the ferrite grain size in mm,the constant, 13.3 ksi, represents the base or lattice strength of iron and
Mn and Si contribute to solid-solution strengthening.
where
presumably have produced a lower result, perhaps closeto the 235 and 250 ksi/-Jin of the current medium C-Nbsteels. In any case, the low C-Nb steel with a Krc of about375 ksi/-Jin clearly has superior fracture toughnessas compared to the medium C-Nb steels. This highertoughness can be attributed to its lower pearlite andsulfide-inclusion con ten ts.
A somewhat surprising result of this investigationconcerns the higher endurance limit of the low C-Nbsteel as compared to the medium C-Nb steels. Typically,the fatigue endurance limit is expected to correlate withtensile strength, and the low C-Nb steel has a somewhatlower tensile strength than the medium C-Nb steels. Thelower sulfide-inclusion content may contribute to theimproved fatigue resistance of the low C-Nb steel. Theendurance limit for the low C-Nb steel, 44 ksi, is similar
136 • Iron &. Steel Technology
the reason that the precipitation strengthening incre-ments are smaller. The higher carbon content of themedium C-Nb steels also results in precipitation at high-er temperatures with attendant precipitate coarsening.In the normalized steel, further coarsening takes placeduring heat treatment. It is speculated that Nb(C,N)precipitates contribute to cyclic strain hardening duringfatigue testing, and that this cyclic hardening contrib-utes to the higher fatigue life and endurance limit of thelow C-Nb steel as compared to the medium C-Nb steels.
It is generally known that fatigue cracks in polishedspecimens initiate due to dama~e to persistent slipbands at the specimen surface. 2 In ferrite-pearlitesteels, the damage generally occurs in the ferrite asopposed to the pearlite, which can arrest cracks frompropagating.I3 Ferrite which has been more effectivelystrengthened by Nb(C,N) precipitates (e.p" low C-Nbsteel) will exhibit greater cyclic hardening. 4 This great-er cyclic hardening provides more resistance to slip,thereby leading to a higher fatigue endurance limit.Others. have also 0bs.erve? cyclic hardenin~ and excel-lent fatIgue strength III mlCroalloyed steels. 5-17
ConclusionThis investigation has shown that controlled rolledlow-carbon Nb-microalloyed steels offer considerableadvantages to wind turbine tower producers in the formof improved impact toughness and fracture toughnessas compared to medium-carbon steels. The toughnessbenefit is attributed primarily due to the lower, moreuniformly distributed pearlite content of low-carbonsteels. Precipitation of Nb(C,N) provides a substantialcontribution to yield strength in low-carbon steels,which also appears to improve fatigue resistance, per-haps due to a cyclic strain hardening mechanism.Based on these results, the implications of specifyingthe EN 10025 normalized rolling delivery conditionare clear: the steel chemical composition constraintsimposed by the EN 10025 normalized rolling require-ment result in wind turbine tower plates with reducedweldability, toughness and fatigue resistance.
The authors plan to conduct additional fracturetoughness testing at -40°F, as well as room-temperaturefatigue crack growth testing. The results of this addi-tional testing will be reported at a later date.
AcknowledgmentsThe authors gratefully acknowledge the assistance fromScott Smith with the tensile and Charpy testing, EricGilbertson with the scanning electron microscopy andJustin Raines with the inclusion analysis.
References1. BA Graville, "Cold Cracking in Welds in HSLASteels,"Welding
of HSLA (Mieroalloyed) Structural Steels, ASM International,Materials Park, Ohio, 1976.
2. ASTM Standard E2142, "Test Methods for Rating andClassifying Inclusions in Steel Using the Scanning ElectronMicroscope," Annual Book of ASTM Standards, Vol. 03.01, ASTMInternational, Conshohocken, Pa., 2010.
3. ASTM Standard E8, "Test Methods for Tension Testing ofMetallic Materials," Annual Book of ASTM Standards, Vol. 03.01,ASTM International, Conshohocken, Pa., 2010.
4. ASTM Standard E23, "Test Methods for Notched Bar ImpactTesting of Metallic Materials," Annual Book of ASTM Standards,Vol. 03.01, ASTM International, Conshohocken, Pa., 2010.
5. ASTM Standard E1820, "Test Method for Measurement ofFracture Toughness,"Annual Book of ASTM Standards, Vol. 03.01,ASTM International, Conshohocken, Pa., 2010.
6. ASTM Standard E466, "Practice for Conducting ForceControlled Constant Amplitude Axial Fatigue Tests of MetallicMaterials," Annual Book of ASTM Standards, Vol. 03.01, ASTMInternational, Conshohocken, Pa., 2010.
7. A.D. Wilson, "Influence of Inclusions on the FractureProperties of A588A Steel," Fracture Mechanics: FifteenthSymposium, ASTM STP 833, ASTM International, Philadelphia,1984, pp. 412-435.8. H. Chen, G. Grondin and R.G. Driver, "Characterization
of Fatigue Properties of ASTM A709 High-Performance Steel,"Journal of Constructional Steel Research, Vol. 63, 2007, pp.838-848.
9. R.L. Bodnar, EB. Fletcher and M. Manohar, "The PhysicalMetallurgy of Normalized Plate Steels," MS&T 2004 ConferenceProceedings, AIST,Vol. 1,2004, pp. 89-109.
10. J.D. Grozier and J.H. Bucher, "Correlation of Fatigue LimitWith Microstructure and Composition of Ferrite-Pearlite Steels,"Journal of Materials, Vol. 2, 1967, pp. 393-407.
11. K.J. Irvine and EB. Pickering, "Low-Carbon Steels WithFerrite-Pearlite Structures," JISI, Vol. 201, November 1963, pp.944-959.
12. R.W. Hertzberg, Deformation and Fracture Mechanics ofEngineering Materials, John Wiley & Sons, New York, 1976, p.460.
13. K. Hussain and R.R. De Los Rios, "Microstructural Effect onTensile and Fatigue Behavior of C-Mn Steel," Journal of MaterialScience,Vol. 32, 1997, pp. 3565-3569.
14. R.W. Landgraf, "Control of Fatigue Resistance ThroughMicrostructure - Ferrous Alloys," Fatigue and Microstructure,ASM International, Materials Park, Ohio, 1979, pp. 453-455.
15. P. Watson and T.H. Topper, "An Evaluation of the FatiguePerformance of Automotive Steels," SAE Mid-Year Meeting,Montreal, Que., Canada, 7-11 June 1971, Paper No. 710597.
16. T. Abe, T. Sampei, H. Osuzu and I. Kozasu, "StaticStrengthening Mechanisms of Low- and Medium-Carbon Steelsand Their Fatigue Strength," Tetsu-to-Hagane, Vol. 70, No. 10,1984,pp.1459-1466.17. R.L. Bodnar and 5.5. Hansen, "Metallurgical Development
of Rolled Steel Channels for Truck Frames," 32nd MechanicalWorking & Steel Processing Conf. Proc., Vol. XXVIII, 1991, pp.61-72. +
This paper was presented at the 2011 International Symposium on the Recent Developments in Plate Steels,Winter Park, Colo., and published in the Conference Proceedings.
I-J......•.•.c.w''..c~!~
October 2011 + 137
