LUCHT- EN RUIMTEVAARTLABORATORIUM

Transcription

1 z r- :- -1 0co C. c NATIONAAL LUCHT- EN RUIMTEVAARTLABORATORIUM NATIONAL AEROSPACE LABORATORY NLR THE NETHERLANDS NLR U VARIABILITY OF FATIGUE CRACK GROWTH PROPERTIES FOR 2024-T3 ALUMINIUM ALLOY by A. Oldersma and R.J.H. Wanhill.DIC USERS Ot NLR

2 U IN CLA5511WILL) AD NUMBER AD- j.z LZ NEW LIMITATION CHANGE TO DISTRIBUTION STATEMENT A - Approved for public release; Distribution unlimited. Limitation Code: 1 -F k FROM DISTRIBUTION STATEMENT Limitation Code: AUTHORITY THIS PAGE IS UNCLASSIFIED

3 Reprints from this document may be made on condition that full credit is given to the Nationaal Lucht- en Ruimtevaartlaboratorium (National Aerospace Laboratory NLR) and the author(s). NATIONAAL LUCHT- EN RUIMTEVAARTLABORATORIUM NATIONAL AEROSPACE LABORATORY NLR y Anthony Fokkerweg 2, 1059 CM AMSTERDAM, The Netherlands P.O. Box 90502, 1006 BM AMSTERDAM, The Netherlands Telephone 31 -(0) Fax :31 -(0)

4 DOCUMENT CONTROL SHEET ORIGINATOR'S REF. U SECURITY CLASS. Unclassified ORIGINATOR National Aerospace Laboratory NLR, Amsterdam, The Netherlands TITLE Variability of fatigue crack growth properties for 2024-T3 aluminium alloy PRESENTED AT AUTHORS DATE pp ref A. Oldersma and R.J.H. Wanhill DESCRIPTORS Aging (materials) Aluminium alloys Corrosion resistance Crack propagation Cyclic loads Damage assessment Environment effects Fatigue tests Mechanical properties Speciment geometry ABSTRACT Discussions between D-BAA and the NLR have shown that significant variability in fatigue crack growth properties is possible for the industry standard damage tolerant aluminium alloy, 2024-T3 sheet. This is important for two reasons: (1) design assumptions and (2) comparisons of different candidate materials, such as 6013-T6, with 2024-T3. Another important topic is the possibility of changes in damage tolerance properties owing to long term natural ageing. In this report a comparison was made between 2024-T3 aluminium alloy sheet crack growth obtained from two NASA reports (from 1958 and 1959) and an NLR report (from 1966) to examine batch to batch variability of fatigue crack growth rates. Two other NASA reports (from 1969 and 1988), which refer to test specimens from the same batch of material, were used to determine if there is any long term natural ageing effect on fatigue crack growth properties. From the results it can be concluded that: 1. There is a significant effect of long term natural ageing on crack growth properties of 2024-T3, 2. Significant variability in fatigue crack growth properties is possible for different batches of 2024-T3. 3. The effect of long term ageing is less than or similar to the effect of batch-to-batch variation. 4. The effect of cycle frequency has to be considered even when tests are done in laboratory air

5 NLR TECHNICAL PUBLICATION U VARIABILITY OF FATIGUE CRACK GROWTH PROPERTIES FOR 2024-T3 ALUMINIUM ALLOY by A. Oldersma and R.J.H. Wanhill This investigation has been done as a contribution to topics of mutual interest to Daimler-Benz Aerospace Airbus and the NLR. Division Structures and Materials Completed Prepared AO/,q J /. Order number: Approved RJHW/ Typ. MM

6 -3- Summary Discussions between D-BAA and the NLR have shown that significant variability in fatigue crack growth properties is possible for the industry standard damage tolerant aluminium alloy, 2024-T3 sheet. This is important for two reasons: (1) design assumptions and (2) comparisons of different candidate materials, such as 6013-T6, with 2024-T3. Another important topic is the possibility of changes in damage tolerance properties owing to long term natural ageing. In this report a comparison was made between 2024-T3 aluminum alloy sheet crack growth data obtained from two NASA reports (from 1958 and 1959) and an NLR report (from 1966) to examine batch to batch variability of fatigue crack growth rates. Two other NASA reports (from 1969 and 1988), which refer to test specimens from the same batch of material, were used to determine if there is any long term natural ageing effect on fatigue crack growth properties. From the results it can be concluded that: 1. There is a significant effect of long term natural ageing on the crack growth properties of 2024-T3. 2. Significant variability in fatigue crack growth properties is possible for different batches of 2024-T3. 3. The effect of long term ageing is less than or similar to the effect of batch-to-batch variation. 4. The effect of cycle frequency has to be considered even when tests are done in laboratory air.

7 -4-0 Contents 1 Introduction 5 2 Batch variability 6 3 Frequency and thickness effects Effect of frequency Effect of sheet thickness 7 4 Long term variation 8 5 Discussion 9 6 Conclusions 10 7 References 11 6 Tables 9 Figures (33 pages in total)

8 -5-1 Introduction Discussions between D-BAA and the NLR have shown that significant variability in fatigue crack growth properties is possible for the industry standard damage tolerant aluminium alloy, 2024-T3 sheet. This is important for two reasons: (1) design assumptions and (2) comparisons of different candidate materials, such as 6013-T6, with 2024-T3. Another important topic is the possibility of changes in damage tolerance properties owing to long term natural ageing, which has already been demonstrated to be deleterious to corrosion resistance. With increasing age there is a trend towards more intergranular corrosion at the expense of general pitting.

9 -6- ( 2 Batch variability Schijve [1] compared fatigue crack growth data of 2024-T3 Alclad sheet material from several manufacturers to examine batch-to-batch variation. The investigation also included differentbatches from the same manufacturer. Constant amplitude fatigue tests were carried out at the NLR with the same test conditions: specimen dimensions, cycle frequency, environment and testing machine. It was found that the variability between batches from the same manufacturer was similar to the variability between batches from different manufacturers. In the investigation three stress amplitude levels were used with the same mean stress level, resulting in three different R-values: R = 0.037, R = 0.27, and R = The results are shown in figure 1 (scatterbands) and listed in table 1. Each test series (same R,AK) contained 21 specimens. The standard deviation of the log(da/dn) values of each test series was calculated in order to describe the scatter in crack growth rates. The results are listed in table 2 and shown in figure 2. There is a tendency towards a higher standard deviation with increasing AK. The Schijve results have also been compared with NASA (NACA) fatigue crack growth test data from the following reports [2, 3]: - NACA TN 4394, prepared by McEvily, A.J. and ll1g, W., September NASA TN D-52, prepared by McEvily, A.J. and Illg, W., October The test conditions (specimen dimensions, cycle frequency and testing machine) for these tests differed from the NLR test conditions. The crack growth rate data are listed in tables 3 and 4. Figure 3 compares these data with the NLR data. Although tested at different frequencies and at slightly different R-values ( ) the data from NACA TN 4394 fit into the scatterband for R = from reference 1. The NASA TN D-52 data illustrate the effect of negative R- value (R = -1). Taken as a whole, the crack growth results in figure 3 show that there is an effect of cycle frequency: lower frequencies result in slightly higher crack growth rates. This effect is discussed in the next section.

10 -7-3 Frequency and thickness effects 3.1 Effect of frequency The effect of cycle frequency is interrelated with the effect of atmospheric humidity. This is discussed in ref[4]. Figure 4 illustrates both effects and shows that reducing the frequency increased crack growth rates more in dry air than in wet air at low AK, but did the opposite at high AK. This figure also shows that the humidity effect decreases with decreasing frequency, i.e. at low frequencies there are only slight differences in crack growth rates in dry or wet air. The data used in the present report are all from tests carried out in laboratory air. Although one might think that the humidity effect can therefore be neglected, figure 3 shows that there is still a frequency effect, which has to be taken into account when comparing crack growth rate data. 3.2 Effect of sheet thickness Constant amplitude tests in normal and nominally dry air on plate and sheet materials demonstrate that crack growth rates tend to be higher in thicker section materials. (Note that specimen thickness, rather than sheet thickness, has a negligible influence for 2024-T3, but not for some materials, notably aluminium-lithium alloys.) The data used in the present report are from test specimens with sheet thicknesses ranging from 2.0 to 2.6 mm. It is assumed that in this thickness range the thickness effect is negligible.

11 -8-4 Long term variation To investigate the effect of long term natural ageing, NASA fatigue crack growth test data for 2024-T3 sheet material were used from the following reports [5, 6]: - NASA TN D-5390, prepared by Hudson, C.M., August AGARD report no.732, annex G, E.P. Phillips, The long term natural ageing effect can be studied because the materials tested came from the same batch, which was manufactured before The crack growth rate data are listed in tables 5 and 6. Figures 5-7 compare data for the same R, but tested in different decades. The crack growth rates for the Phillips 1988 data (ref[6]) were slightly higher than those for the Hudson 1969 data (Ref. 5) at the same frequency. Note, however, that figure 7 shows a frequency effect for the Hudson 1969 data (0.5 Hz versus 30 Hz), and that the Phillips 1988 data still are above the Hudson 1969 data even though the latter were obtained at a lower frequency (14 Hz versus 30 Hz).

12 -9-5 Discussion The batch-to-batch variation in crack growth is characterized by the scatterbands obtained from the data in reference 1 and shown in figure 1 for R = 0.037, R = 0.27 and R = The scatterband representing R=0.47 is smaller, indicating a somewhat lower variability. This is also indicated by the calculated standard deviations, table 2 and figure 2. Figures 8 and 9 show that the NASA 1969 and 1988 data (Refs. 5, 6) fit mostly in the scatterbands for the NLR 1966 data (Ref. 1) although there are slight differences in R values and cycle frequencies. This means that the effect of long term ageing on the NASA material is less than or similar to the effect of batch-to-batch variation. Nevertheless, both effects should be considered as significant, taking into account the log-log method of data presentation. The frequency effect in laboratory air could not be studied over the total AK-range, since the test frequencies were generally adjusted to the applied stress amplitude. Therefore the low frequency data mostly relate to higher AK-values and the high frequency data to lower AKvalues.

13 Conclusions In this report a comparison was made between 2024-T3 aluminum alloy sheet crack growth data obtained from two NASA reports (from 1958 and 1959) and an NLR report (from 1966) to examine batch-to-batch variability of fatigue crack growth rates. Two other NASA reports (from 1969 and 1988), which refer to test specimens from the same batch of material, were used to determine if there is any long term natural ageing effect on fatigue crack growth properties. From the results it can be concluded that: 1. There is a significant effect of long term natural ageing on the crack growth properties of 2024-T3. 2. Significant variability in fatigue crack growth properties is possible for different batches of 2024-T3. 3. The effect of long term ageing is less than or similar to the effect of batch to batch variation. 4. The effect of cycle frequency has to be considered even when tests are done in laboratory air.

14 -11- l 7 References 1. Schijve, J.; de Rijk, P., The fatigue crack propagation in 2024-T3 Alclad sheet materials from seven different manufacturers, National Aerospace Laboratory Report NLR M2162, Amsterdam, May McEvily, A.J.; Illg, W., The rate of fatigue-crack propagation in two aluminium alloys, National Advisory Committee for Aeronautics Report NACA TN 4394, September McEvily, A.J.; Illg, W., The rate of fatigue-crack propagation for two aluminium alloys under completely reversed loading, National Aeronautics and Space Administration Report NASA TN D-52, October Wanhill, R.J.H., Environmental effects on fatigue of aluminium and titanium alloys, National Aerospace Laboratory NLR, Amsterdam, paper in "Corrosion fatigue of aircraft materials", AGARD Report NO.659,"Corrosion Fatigue of Aircraft Materials", April Hudson, C.M., Effect of stress ratio on fatigue-crack growth in 7075-T6 and 2024-T3 aluminium-alloy specimens, National Aeronautics and Space Administration Report NASA TN D-5390, August Phillips, E.P., Long-crack growth rate data -constant amplitude and FALSTAFF loading-, AGARD Report NO.732: "Short-Crack Growth Behaviour in an Aluminum Alloy- An AGARD Cooperative Test Programme", December 1988.

15 -12- ( 00 6l \66 N' 10 In m 0 10 e 16 c -- : C'0 C cn C' 'S mc' \0- ' \0 0 v~ a\ 'n 'C 0o 'C m mc.4 N N ms Cs N q 'C Nl CSC4 C-4 N C' c N C4 0 N (ccqn 'C - mc ccn 10~V N6 \0 ' m'~ S -0 s C) 0q 'is m rs 'C 'C m 00 ~ CM IM ;0"N S 0 C4 _' N 00 v 14sVS lc m 00 -(' 00 eq C r0- N0cq 07\ C4 (. 'C o 0 0 a, a, - ) n r \0 00 W In %0 \,R e0s (' \ V' 'C I: 'Cq In CS -l 'C \R - C \q C C. C\ 00 0 N C r, S (S - (S 00 ' s 'S 05 CD N CD 00 CD C5 N5 N N Q \0, m C, CYS1 V') C C-4 i O 00 0 N 10 No 'i' 6' mn ' 00 t- a ýn 00 el ms NY ' 0 5 N N v~ N ICN~ 6D 0n & N m 00 " ;'S ~,5 C ' i ' o N -- In I OS 0 N C N 'C('S 'C - 'S N SCr Nt-V 0 0 C S N ~ (S N (i 0 - N N N ' N- cq 0M No 00 'C - 'S N CS ' 5 N 0 C- N 'S N O~~~~~ 0 00 N CFC C - N C 5 N (S(S0 ~~ Ii 0(S ms (S 'SC4 is N N ' N (S (S N 0 ('S 05 N 0 N (' 2 ' ' m' - ('S ('S ('S ' 0~ \ G \0 Cm mi N. 5 C I ' g. A N V C ' S S 0 C 0 0 ' i '

16 mn Mn Mn e O n n n w 2 10 n C1 e m3 V: c0 C> rn n en. en n Cý 0 - '3 ' 0 - m3 000 ( (N enlqc (N (N - q Nn -q- m V' -t e ' m 3 ' Vn m'- en M3 en en C4 MN M0 C4* ~'4 cq t~*e en 0 '3 w An m 0 r- e e n 0 on en en m0 e e 0-4n e C e ' - - n (N (N -n - (N - - ~ N ( 0 \- '3- w3 3 ( e N ' n n C4 a, ;Z' e ' 00 en("r n e '3 3 3 e 3 ( w '3 eo m3 w3 en en3 Vn ' 'n ( 0 en 0 0q '3 ' 3e C, w3' '30 m~ 0-0 t- w0 w0i'd0[-. %D '3~~~~~ (N '3 e n e N ( n e v ' e 0 a, 3 \30 a, en n 3'0 '3 0 0 n e ' n 043 (N e n e ci 0 en en en ( 3 ( C3 0 '3 0 Q C. ( 0N en M en en ":r (N en 00 0 as n ' n 3 ' 0 n en e0 en '3-2 m 4. " 00m0-00 ' -V en IV, en en en C3 '3 enm en (N en en ' w30 en en3 ' ent en e e -It n '3 en In In l: I: In -0 \6m (N ~ ~ '3 a ( n '3 3 n- 3' '3 e n r- 0 0 mn n 0 C m en (N en '3 en (N 4 C5 en C5 en 1 0* 10 1C 0 01C 0 en '3. u~0 C> 5~ ' 0 en en 0 - en en en 00 0 ae en en 00 en en e ( 3a(, e 00 e 0 e m3 n m m5 ws en en w 3 en (uo e n - 0n en 30 en (N en I-4

17 -14- IV" (N U, N N 'n o 00 '0 t 0, 0 0 '0 ( r- N U N N ( 0 In UQ (N O,~ N, I, (N - N ( 0 0 ;n N 0 ". % ( cn - ' r- 0 (N N 'o O( C4 N - ( r ON m wn C' ' 0 o c ( N U, 0 N ouq%0 0 N U N 0 m0 ~ m eq (N (N (N (N LN ci (N (N N - (N m (N cq r- r- F4 I" F (N q ' -U -N - - ' N - % U, (N (N (N 0 0 't:,l (N U, (N 0 0 c-00 0 (N! 0 r, tno 405 cq '0 0 0%mc (N -I I 0mm (N ~~ ~ 0 ' NU,U 0 '0 0 0 ' 0 %( U 0o ( S 0 % ( 0 ~ 00 In- ~ - ' 0 U, U, ( (N 0o cq 00~o (Nr U, In,, ~ w? In (.. N ~ 0-0, 0 N ~ 0 ( N 0, - ~ N ~ ( Ni~ ~ ' N cn c o o w6a,,o In oo -~ -q ~ N N 0 0 ( In 0 0 N % ( ~, 0 ' 0 O 0 0% % 0 e, 0 ( '0 m% 00m 0 N 0 0 N ' (S( ( 50* (N(N( (N(N( 0(N 0 ~ N0 ( (N- (N - N ( N - o%0 U N ( r- -o c- (N(N*(N*o(N*(a, (N~~~~ (o 0" ~~ U, 0 0 -' 0 00~ I m 0 I% (N ~ ~ ~ ~ ~ ~ ~ ~ o-0 % 0 (, 0 ' 00 N N U % - 0 U % ' 0 ~ m ~ ( ~ ~ 0 ~ 0 ~ -q 0 ~ ( ~ - ( ~. l S - ( m~( l o6 t-- c '0 t- lo N en 0 In~- 00 \o, ( 0 N ( 0 ~ - (N 0 0 U, 0 '0 a, 0q '0D 0 0 ~ % ( 0 O~~o 0" (N - N'mU % ( 0 U, 0 % 0 % U o~~q g o In U N N ( ~ (, 0 ', 0 U N U 0-0 r( go0 0 0 N U U ( ( I.0 o ooao o o o o o o o 00 '0UA 0 ( N ~ - ( N (

18 -15- "(2 Table 2 Calculated standard deviations of log (da/dn) as a function AK Report: NLR M2162 [1] Material: 2024-T3 Alclad Frequency: 53 Hz. Specimen width: 160 mm sheet thickness: 2.0 mm R AK std(log R AK std(log R AK std(log (MPa V'm) (daldn)) (MPa V'm) (da/dn)) (MPa Vm) (da/dn))

19 -16-0 Table 3 Crack growth results from McEvily and IlIg (Ref. 2) Report: NACA TN 4394 [2] Material: 2024-T3 bare frequency: see table specimen width: 305 mm sheet thickness: 2.6 mm Sdaldn R i AK da/dn R MPa Vm 10-6 m/cycle MPaVm 10- m/cycle frequency 30 Hz frequency 20 Hz frequency 30 Hz frequency 20 Hz frequency 0.33 Hz frequency 20 Hz frequency 0.33 Hz

20 -17- " Table 4 Crack growth results from McEvily and Illg (Ref 3) Report: NASA TN D-52 [3] Material: 2024-T3 bare frequency: see table specimen width: 305 mm sheet thickness: 2.1 mm AK da/dn R AK da/dn R (MPaV,/m) 10-6 m/cycle MPa -/m) 10-6 m/cycle frequency 30 Hz frequency 0.22 Hz frequency 30 Hz frequency 0.15 Hz frequency 20 Hz I

21 -18- V- Table 5a Crack growth results from Hudson (Ref 5) R=-1 Report: NASA TN D-5390 [5] Material: 2024-T3 bare frequency: see table specimen width: 305 mm thickness: 2.3 mm AK da/dn R AK da/dn R MPa-Vm 10-6 m/cycle MPa Vm 10- m/cycle frequency 0.5 Hz frequency 13.7 Hz frequency 0.5 Hz frequency 13.7 Hz frequency 13.7 Hz

22 -19- Table 5b Crack growth results from Hudson (Ref 5) R=0 Report: NASA TN D-5390 [5] Material: 2024-T3 bare frequency: see table specimen width: 305 mm thickness: 2.3 mm AK daldn R AK dadn R MPaV/m 10- m/cycle MPav/m 10-6 m/cycle frequency 0.3 Hz frequency 20 Hz frequency 20 Hz frequency 30 Hz frequency 20 Hz

23 -20- Table 5c Crack growth results from Hudson (Ref. 5) R = 0.33 Report: NASA TN D-5390 [5] Material: 2024-T3 bare frequency: see table specimen width: 305 mm thickness: 2.3 min AK da/dn R AK da/dn R MPa qm 10-6 m/cycle MPa "m 10-6 m/cycle frequency 20 Hz frequency 20 Hz frequency 20 Hz frequency 30 Hz frequency 20 Hz

24 -21- Table 5d Crack growth results from Hudson (Ref. 5) R = 0.5 Report: NASA TN D-5390 [5] Material: 2024-T3 bare frequency: see table specimen width: 305 mm thickness: 2.3 mm AK da/dn R AK da/dn R MPa 4 mm 10-6 m/cycle MPa 'IM 10-6 m/cycle frequency 20 Hz frequency 20 Hz frequency 20 Hz frequency 30 Hz frequency 20 Hz

25 -22-0 Table 5e Crack growth results from Hudson (Ref. 5) R= 0.7 Report: NASA TN D-5390 [5] Material: 2024-T3 bare frequency: see table specimen width: 305 mm thickness: 2.3 mm AK da/dn R AK da/dn R MPaV'm 10-6 m/cycle MPa-'m/ J 10-6 mr/cycle frequency 13.7 Hz frequency 13.7 Hz frequency 13.7 Hz frequency 13.7 Hz frequency 13.7 Hz

26 -23- Table 6a Crack growth results from Philips (Ref. 6) R-=-1 Report: AGARD-R-732 [6] Material: 2024-T3 bare frequency: 15 Hz specimen width: 50.8 mm sheet thickness: 2.3 mm AK da/dn R MPaVm 10-6 m/cycle I

27 -24- Table 6b Crack growth results from Philips (Ref. 6) R = 0.5 and R = 0 Report: AGARD-R-732 [6] Material: 2024-T3 bare frequency: 15 Hz specimen width: 50.8 mm sheet thickness: 2.3 mm AK da/dn R AK da/dn R MPaVm 10-6 m/cycle MPaVxm 10-6 m/cycle

28 r--r-I i R = R=0.27 ý53hz R 0R=.47I daldni (rn/cycle)jp 10-8 I I ~go 1 ~ ~ ~ ~ ~ 2II10 i Fi 0.6 AK (M P Contan apiueftgecakgotraesatrndrpeetingmtra batch~~ [Re11vaiailt

29 I 1 SR = R=0.27 x R = o std log 0 (da/dn) AK (MPa V) Fig. 2 Batch variability as function of AK [Ref 1]

30 -27- " o R (m/cycle) "- Ilot '' R 07 / Ito I, b I ' /A,,=00 _- R = I 0.33 Hz [2] R= Hz [3],I da/dn -adnr=-i Hz [3] I!i- I, it I 53Hz[1] 0 = R 10- ',,,"o i5hz [] 10-8 i AK (MPa ýfi) Fig. 3 Constant ampfitude fatigue crack growth results from McEvily and lig [Refs. 2, 3] compared with two of the scatterbands from figure 1

31 daldn (mm/cycle) Constant amplitude; R =0.1 * AU 0 Dry air (10-25 ppm H 2 I0/ 0) o A 0 Saturated wet air ~ Fi. 1O~ Hz, 5 H~ K(HZaA ffct Feqeny o 1mm204-3 lcadshet[rf.4

32 I F FO 0TN D-5390 da'dn20 Hz [5] (rn/cycle) 0 TN D Hz [5] * Qo AGARD-R Hz [6] 10, 0 iso _ AK (MPa,/iii) Fig. 5 Comparison of constant amplitude fatigue crack growth rates for R = 0

33 TN D ~ 0ad 20 Hz[5] CD (r/yl)30 TN D-5390 Hz [5]! AGARD-R OHz [6] ILj AK (MPa. -%i) Fig. 6 Comparison of constant amplitude fatigue crack growth rates for R = 0.5

34 -31-10Q-4 I I 0 0o- 5 0 TN D d/n0.5 Hz [5]0 da/dn TN D-5390 (mlcycle)14h[5 AGARD-R Hz [6] a i io ~ ~ ~ ~ ~ ~ ~ A -Uf Fig. 7 Comparison of constant amplitude fatigue crack growth rates for R = 1

35 R~ = H z [5] I l~ I ) _ad D-39 R= daldn 7TN D-5390 (rn/cycle) R=,03~(] I@ R = 3 _O, [ Hz [1]l 4.:- - l L L _-- L L AK (MPaFfi) Fig. 8 Comparison of crack growth rates for R =0 and a scatterband representing crack growth data for R =

36 da/d TN D-39 E SI I II I I I_ R _-Hz_[5 T I L AGARD-R-732 R =0.5, 30 Hz [6] daldn TN D-5390 da/dn ~ =0.5, ~R Hz [5] / I I 0-6_...,... i z : j - / R = i53 z [1] I I7 Ile l/i Sid lo-8 -- j './ AK (MPa %r) Fig. 9 Comparison of crack growth rates for R = 0.5 and a scatterband representing crack growth data for R = 0.47