NIMS fatigue data sheet on low- and high-cycle fatigue properties of A2017-T4 (Al-4.0Cu-0.6Mg) aluminium alloy

ABSTRACT The new fatigue data sheet, No. 133, discloses low- and high-cycle fatigue properties of A2017-T4 aluminium alloys. Two types of low-cycle fatigue tests were conducted under strain-controlled conditions. One comprised incremental step tests to clarify cyclic stress-strain curves. The other comprised constant- strain amplitude tests to clarify the fatigue lives. The aluminium alloys revealed strong cyclic hardening that resulted in very high cycle yield stresses. The cyclic yield stresses were higher than those of titanium alloys and steels. Constant-strain amplitude tests showed plastic strain amplitudes to be smaller in most cases than elastic strain amplitudes. These appeared to be attributable to the small elongation, low elastic modulus and strong cyclic hardening of the aluminium alloys. Three heats of A2017-T4 revealed equivalent fatigue lives when their fatigue strengths were evaluated by strain amplitude. Evaluation by stress amplitude, in contrast, resulted in differences between heats, likely caused by the effects of cyclic yield stress. The high-cycle fatigue tests up to 108 cycles were uniaxial and carried out under load-controlled conditions. Many specimens fractured at over 107 cycles, so fatigue limits were not confirmed. The high-cycle fatigue strengths were proportional to the tensile strengths. The 107-cycle fatigue strengths were close to those of austenitic stainless steels and normalized carbon steels, while the 108-cycle fatigue strengths were lower. In short, the fatigue strengths deceased at over 107 cycles, showing the need for gigacycle fatigue tests. These will be described in the next issue. GRAPHICAL ABSTRACT IMPACT STATEMENT This paper presents a new fatigue data sheet, titled No. 133, on low- and high-cycle fatigue properties of A2017-T4 aluminium alloy, followed by the gigacycle versions.


Introduction
NIMS fatigue data sheets comprise a huge database of fatigue properties of structural materials [1].The total number of fatigue data sheets is currently 133 (Nos.0-132) and is still increasing.This paper presents a new fatigue data sheet, titled No. 133.
This fatigue data sheet discloses the low-and highcycle fatigue properties of A2017-T4 aluminium alloy.A2017-T4 is a typical high-strength aluminium alloy, so-called Duralumin.This alloy is widely used because of its superior market availability.A series of former fatigue data sheets disclosed gigacycle fatigue properties of A5083P-O, A7075-T6 and A6061-T6 alloys [2][3][4][5][6][7][8][9][10].The former data sheets were comprised of three types.The first was low-and high-cycle fatigue properties tested under strain-and load-control conditions, respectively.The second was gigacycle fatigue properties tested by rotating-bending and ultrasonic fatigue testing.The third was gigacycle fatigue properties at high stress ratios.This fatigue data sheet is thus the first one in the A2017-T4 series.
The low-cycle fatigue tests conducted were of two types.One was a set of incremental step tests [11], which clarified cyclic stress-strain curves.The other comprised constantstrain amplitude tests to clarify the fatigue lives.The highcycle fatigue tests were of the uniaxial loading type at 100 Hz.The cut-off cycle numbers were 10 8 cycles.The details of the new fatigue data sheet are as follows.

Materials
Tables 1 and 2 show the processing details and chemical compositions of the tested alloys.The tested alloys were a hot-rolled plate and extruded round bars sampled in 2020.Table 3 shows their mechanical properties.The tensile strengths of the tested alloys ranged from 400 to 539 MPa.The tensile strengths of A2017-T4 were higher than those of A5083P-O and A6061-T6, but lower than those of A7075-T6.
Figure 1 shows the microstructures of the tested alloys.The microstructures of heats A and C had recrystallized, while those of heat B were fibrous structures without recrystallization.The fibrous structures were similar to those observed in A7075-T6.

Fatigue testing
Table 4 shows the low-cycle fatigue test conditions.The tests were conducted using servo-hydraulic fatigue testing machines (Shimadzu, Japan) under strain-controlled conditions.The strain-controlled tests used triangular waveforms with a strain rate of 5 × 10 −3 s −1 .Both incremental-step and constantstrain amplitude tests were carried out.The incremental step tests employed waveforms in which the strain amplitude gradually changed, repeating 25cycle blocks.
Table 5 shows the high-cycle fatigue test conditions.The uniaxially loaded high-cycle fatigue tests were conducted using an electromagnetic resonance fatigue testing machine (ZwickRoell, Germany) under load-controlled conditions.The stress ratio was R = −1.
The fatigue tests were conducted at room temperature in air.The specimens were round bar types with minimum diameters of 8 or 6 mm.Finishing of the specimens' surfaces consisted of longitudinal polishing using 600-grade silicon carbide papers.    min 215 min 375 min 12 --1) JIS Z 2241 (2015), No.14A type specimen with 8 mm diameter and 40 mm gage length.The nominal strain rate of the specimen was controlled to 0.00025 S −1 . 2)JIS H 4040 (2015).'Aluminium and aluminium alloy bars and wires'. 3)JIS H 4000 (2014).'Aluminium and aluminium alloy sheets, strips and plates'.

Low-cycle fatigue test results
Figure 2 shows the cyclic stress-strain curves obtained in the incremental step tests.The stress amplitudes were at half of the fatigue lives.The results of all heats revealed clear cyclic hardening.Figure 3 shows cyclic yield stresses.The results of A2017-T4 were located about midway between A7075-T6 and other aluminium alloys.The relationships between cyclic yield stress σ yc and tensile strength σ B were very close to σ yc = 0.90σ B .These cyclic yield stresses were higher than those of titanium alloys (σ yc = 0.80σ B ) and steels (σ yc = 0.61σ B ).
Figure 4 shows the results of constant-strain amplitude tests.The total strain amplitudes were divided into plastic and elastic strain amplitudes.The plastic and elastic strain amplitudes were at half of the fatigue lives.The plastic strain amplitudes were smaller than the elastic strain amplitude in most data except under very high total strain amplitude conditions.The results of linear regression analysis are summarized in Table 6. Figure 5 shows cyclic hardening curves.Cyclic hardening occurred very early and saturated after 10% of the fatigue lives.Figure 6 shows an S-N diagram that compares the results between the heats.The differences between heats are very small in this diagram.

High-cycle fatigue test results
Figure 7 shows a S-N diagram in which low-and highcycle fatigue test results are presented together.The stress amplitudes of the low-cycle fatigue tests were at half of the fatigue lives.Many specimens fractured at over 10 7 cycles, meaning that conventional fatigue limits were not confirmed.The results showed differences between heats: heat B was the highest, heat A was in between and heat C was the lowest.This order corresponds to that of tensile strength.The high-cycle fatigue test results were continuously related to the low-cycle fatigue test results.The lowcycle fatigue test results in this figure showed clear differences between heats, in contrast to those in Figure 6.
Table 7 shows numerical values of 10 7 -and 10 8cycle fatigue strengths.The fatigue strengths are average values between the maximum stress amplitude at which no specimen is fractured, and that just above it.Figure 8 shows a comparison of the fatigue strength between A2017-T4 and other materials.The fatigue strength of the aluminium alloys revealed linear relationships with tensile strength.The results of A2017-T4 also showed this trend.The relationships between fatigue strength σ W and tensile strength σ B are σ W = 0.53σ B for the quenched and tempered (QT) steels and σ W = 0.39σ B for austenitic stainless steels and normalized (N) carbon steels.The fatigue strength at 10 7 cycles of the aluminium alloys was very close to σ W = 0.39σ B , while that at 10 8 cycles was lower.This was also the case for A2017-T4.
Figure 9 shows typical fracture surfaces after the high-cycle fatigue tests.All the specimens failed from surface-nucleated fatigue cracks, and the fracture surfaces revealed typical transgranular fatigue fracture morphologies.

Discussion
The first point to be noted in these fatigue test results is the very high cyclic yield stress of the aluminium alloys (Figure 3).It exceeded 90% of the tensile strength, much higher than seen in titanium alloys and steels.SUS630, a 17-4 PH stainless steel, also revealed very high cyclic yield stress.It is therefore possible that precipitation hardening caused the high    cyclic yield stress.However, A5083P-O, an annealed alloy without precipitation hardening, also revealed high cyclic yield stress.Our understanding, therefore, is that the very high cyclic yield stress characteristic of the aluminium alloy is caused by strong cyclic hardening.
The second point is the effect of strain amplitude on low-cycle fatigue properties.In most of the data from the constant-strain amplitude tests (Figure 4), the plastic strain amplitudes were smaller than the elastic strain amplitudes.Normally, the regions in which plastic strain amplitudes exceed the elastic strain amplitudes are 'low-cycle fatigue regions' to which the Manson-Coffin law applies.The low-cycle fatigue regions were very narrow in A2017-T4.This was also the case in other aluminium alloys.The width of the low-cycle fatigue regions varies with the ductility of the materials, i.e. this region is wide in ductile materials [1].The elongations of the aluminium alloys were around 10-20%, smaller than for ductile materials.Moreover, the low elastic modulus increases the elastic strain, and the strong cyclic hardening decreases the plastic strain.These characteristics of the aluminium alloys appear to reduce the width of the low-cycle fatigue regions.
On the other hand, the low-cycle fatigue test results show clear differences between heats in Figure 7, but not in Figure 6. Figure 7 shows the use of stress amplitudes to evaluate low-cycle fatigue strength, while Figure 6 uses strain amplitudes.They illustrate that strain amplitudes evaluate low-cycle fatigue strengths more distinctively than do stress amplitudes.: exponent and coefficient respectively in equation ε ea N ｆ ke = C e .ε pa : plastic strain amplitude, defined as half the width of hysteresis loop at zero load level.ε ea : elastic strain amplitude, defined as (ε ta -ε pa ) where ε ta is total strain amplitude.N ｆ : number of cycles to failure. 1) Determined by linear regression analysis for total strain amplitudes from 5 × 10 -3 to 1.4 × 10 -2 .
The stress amplitudes in the strain-controlled tests are affected by the cyclic yield stress.The effects of the cyclic yield stress appear to cause the differences on the low-cycle fatigue test results between heats in Figure 7.
The third point is that of high-cycle fatigue strength showing differences between heats.With high-cycle fatigue, the plastic strain amplitudes were negligible, so the total strain amplitudes were very close to the elastic strain amplitudes, meaning that the differences between heats would appear even when the total strain amplitudes were applied.Figure 7 shows that the high-cycle fatigue strengths are in the order of the tensile strength, i.e. tensile strengths have a major effect on highcycle fatigue strengths.Figure 8 more clearly shows that high-cycle fatigue strengths are proportional to tensile strengths.The 10 7 -cycle fatigue strengths of the aluminium alloys are close to those of the austenitic stainless steels and the normalized carbon steels.On the other hand, the 10 8 -cycle fatigue strengths are lower in A2017-T4, since the fatigue limits are not confirmed at 10 7 cycles.Previous issues [12] clarified that the fatigue limits were present at 10 7 cycles for A5083P-O, but not for A6061-T6 or A7075-T6.The 10 8 -cycle fatigue strength of A6061-T6 and A7075-T6 are thus also lower.The presence or absence of the fatigue limits will be discussed in the next issue, which will disclose the gigacycle fatigue test results up to 10 10 cycles.

Summary
NIMS fatigue data sheet No. 133 discloses low-and high-cycle fatigue test results for A2017-T4 aluminium alloys.The strain-controlled tests were applied to the low-cycle fatigue tests, conducting incremental step and constant-strain amplitude tests.The highcycle fatigue tests were a uniaxial loading type under load-controlled conditions.
The incremental step tests revealed that the cyclic yield stresses were much higher than for  titanium alloys and steels.In the constant-strain amplitude tests, the plastic strain amplitudes were, in most cases, smaller than the elastic strain amplitudes.These appeared to be a characteristic of aluminium alloys, attributable to their small elongation, low elastic modulus and strong cyclic hardening.
The high-cycle fatigue strengths were proportional to the tensile strengths.The 10 7 -cycle fatigue strengths were close to those of the austenitic stainless steels and the normalized carbon steels.On the other hand, the 10 8 -cycle fatigue strengths were lower since the fatigue strengths decreased at over 10 7 cycles, i.e. the fatigue limits were not confirmed at 10 7 cycles.

Figure 3 .
Figure 3.Comparison of cyclic yield stress between A2017-T4 aluminium alloy and other materials.

Figure 5 .
Figure 5. Change in stress amplitude during constant-strain amplitude fatigue testing of A2017-T4 aluminium alloy.

Figure 6 .
Figure 6.S-N diagram showing total strain amplitude fatigue properties of A2017-T4 aluminium alloy.

Figure 7 .
Figure 7. S-N diagram showing constant-strain or constant-stress amplitude fatigue properties of A2017-T4 aluminium alloy.

Figure 8 .
Figure 8.Comparison of fatigue strength between aluminium alloys and steels under constant-stress amplitude fatigue testing.

Table 4 .
Low-cycle fatigue test conditions.

Table 5 .
High-cycle fatigue test conditions.