ASD-STAN/D 4/WG 5

Date: 2026-02-19

prEN 3988:2026

Secretariat: DIN

Aerospace series — Test methods for metallic materials — Constant amplitude strain-controlled low cycle fatigue testing

Luft- und Raumfahrt — Prüfverfahren für metallische Werkstoffe — Dehnungsgesteuerter Kurzzeit-Ermüdungsversuch (LCF) mit konstanter Amplitude

Série aérospatiale — Méthodes d'essais applicables aux matériaux métalliques — Essais de fatigue oligocyclique en déformation imposée

CCMC will prepare and attach the official title page.

Contents Page

European foreword 4

Introduction 5

1 Scope 6

2 Normative references 6

3 Terms and definitions 6

4 Principle 10

4.1 General 10

4.2 Definitions 10

4.2.1 General 10

4.2.2 Test section 10

4.2.3 Gauge length 10

4.2.4 Cross-section area 10

4.2.5 Cycle 10

4.2.6 Stress-strain loop 11

4.2.7 Creep-fatigue 12

4.2.8 Failure 13

4.2.9 Mid-life stress-strain loop 15

5 Test equipment 15

5.1 Test machine 15

5.1.1 General 15

5.1.2 Test machine calibration 15

5.2 Cycle counting 15

5.3 Extensometer 16

5.3.1 General 16

5.3.2 Extensometer calibration 16

5.3.3 Waveform generation and control 16

5.3.4 Test fixtures 16

5.4 Heating device 17

5.5 Temperature measurement 17

5.6 Data recorders 18

5.6.1 General 18

5.6.2 Calibration 18

6 Test piece 19

6.1 Design 19

6.2 Sampling, storage and handling 20

6.3 Test piece preparation 21

6.4 Test piece measurement 21

7 Test method 22

7.1 Test piece insertion 22

7.2 Test piece heating 22

7.3 Test commencement 22

7.3.1 General 22

7.3.2 Waveform optimization and control 22

7.3.3 Data recording 24

7.4 Test termination 25

8 Post-test checks 25

8.1 Accuracy of control parameters 25

8.2 Examination of the fracture surface 25

8.3 Determination of the fatigue life 25

8.4 Examination of the stress-strain loops 26

9 Test report 26

9.1 Essential information 26

9.2 Additional information 27

9.3 Presentation of results 28

Annex A (informative) Use of thermocouples 29

Annex B (informative) Test piece preparation 30

Annex C (normative) Guidelines on test piece handling and degreasing 32

Annex D (informative) Failure criteria 33

Bibliography 34

European foreword

This document (prEN 3988:2026) has been prepared by ASD-STAN.

After enquiries and votes carried out in accordance with the rules of this Association, this document has received the approval of the National Associations and the Official Services of the member countries of ASD-STAN, prior to its presentation to CEN.

This document is currently submitted to the CEN Enquiry.

Introduction

This document is part of the series of EN metallic material standards for aerospace applications. The general organization of this series is described in EN 4258.

This document applies to uniaxial strain-controlled low cycle fatigue testing of metallic materials governed by EN aerospace standards. It defines the properties that need to be determined and the terms used in describing the tests and test pieces.

It specifies the equipment, the test pieces, the method of testing and the presentation of results. It applies to testing at ambient and elevated temperatures.

The purpose of this document is to ensure the comparability and reproducibility of the test results. It does not cover the evaluation or interpretation of the results.

This document is restricted to the use of test pieces having a circular cross-section. In some particular cases the practice can be applied to flat test pieces. The major difficulties concern the preparation of the test pieces and their alignment in the grips.

The following documents are referred to in the text in such a way that some or all of their content constitutes requirements of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.

ASTM E1012:2019,^[1] Standard Practice for Verification of Testing Frame and Specimen Alignment Under Tensile and Compressive Axial Force Application

For the purposes of this document, the following terms and definitions apply.

ISO and IEC maintain terminology databases for use in standardization at the following addresses:

— ISO Online browsing platform: available at https://www.iso.org/obp/

— IEC Electropedia: available at https://www.electropedia.org/

3.1

force

F

instantaneous load applied to the test section, in kN

Note 1 to entry: Tensile forces are considered to be positive and compressive forces negative.

3.2

strain

ɛ

extension of the test piece gauge length, due to the force which is applied to the test piece, divided by its original gauge length

Note 1 to entry: It is taken to be positive when the gauge length increases in length and negative when it contracts as a percentage.

3.3

maximum strain

ɛ_max

highest algebraic value of strain applied during the strain cycle as a percentage

3.4

minimum strain

ɛ_min

lowest algebraic value of strain applied during the strain cycle as a percentage

3.5

mean strain

ɛ_m

half the algebraic sum of maximum and minimum strains as a percentage

3.6

strain range

Δɛ

algebraic difference between the maximum and minimum strains as a percentage

Note 1 to entry: The total strain range includes elastic and plastic strain ranges.

3.7

strain amplitude

ɛ_a

half the strain range as a percentage

3.8

strain ratio

Rɛ

algebraic ratio of the minimum strain to the maximum strain

Note 1 to entry: The A ratio, which is defined as the ratio of strain amplitude to the mean strain, is sometimes used.

3.9

stress

σ

force divided by the nominal cross-sectional area, in MPa

Note 1 to entry: It is the independent variable in a stress-controlled fatigue test.

Note 2 to entry: The nominal cross-sectional area (engineering stress) is that calculated from measurements taken at ambient temperature and no account is taken for the change in section as a result of expansion at elevated temperatures.

3.10

maximum stress

σ_max

highest algebraic value of stress applied, in MPa

3.11

minimum stress

σ_min

lowest algebraic value of stress applied, in MPa

3.12

stress range

Δσ

arithmetic difference between maximum stress and minimum stress, in MPa

Note 1 to entry: Δσ = σ_max - σ_min

3.13

primary stress range

Δσ₁

algebraic difference between the maximum and minimum stresses, for creep-fatigue tests, in MPa (see Figure 3)

3.14

secondary stress range

Δσ₂

difference between the stresses at the points of reversal of strain, for creep-fatigue tests, in MPa (see Figure 3)

3.15

stress amplitude

σ_a

half the stress range, in MPa

3.16

stress ratio

R_s

ratio of minimum stress to maximum stress during a fatigue cycle

Note 1 to entry: R_s = σ_min/σ_max

3.17

initial modulus of elasticity

E_O

modulus of elasticity determined on the loading portion of the first cycle or alternatively prior to start of test, at test temperature, in MPa

3.18

mid-life modulus of elasticity

E_m

average of the modulus of elasticity determined on the loading and unloading portions of the mid-life stress-strain loop, in MPa

3.19

inelastic strain range

Δɛ_e

stress range (secondary stress range for the creep-fatigue tests) divided by the modulus of elasticity, as a percentage

3.20

inelastic strain range

Δɛ_p

difference between the total strain range and the elastic strain range, as a percentage

3.21

number of cycles

N

number of strain sequences applied

3.22

number of cycles during failure (general)

N_f(X)

number of cycles corresponding to a decrease of x % in the stress value extrapolated over the maximum stress versus number of cycles curve when the maximum stress falls sharply to failure (see Figure 4)

3.23

fatigue life

N_f

number of cycles corresponding to a decrease of 10 % in the stress value extrapolated over the maximum stress versus number of cycles curve when the maximum stress falls sharply to failure (see Figure 4)

3.24

number of cycles to initiation of an apparent macro-crack

N_i(0)

number of cycles corresponding to the first discernible decrease of the maximum stress versus number of cycles curve when the maximum stress falls sharply to failure (see Figure 4)

3.25

number of cycles to complete separation

N_f(100)

number of cycles corresponding to the complete separation of the test piece into two distinct parts

3.26

total number of cycles

N_t

total number of sequences applied

3.27

frequency

f

expressed in Hertz

3.28

parallel length

L_p

length in the gauge test section of a specimen or test piece that has equal test diameter or test width and is parallel, in mm

3.29

specimen length

L_z

overall length of test specimen, in mm

The uniaxially-loaded strain-controlled low cycle fatigue test consists in maintaining a test piece at a uniform temperature and subjecting it to a constant strain-amplitude waveform. The test piece has a gauge length of constant circular cross-section on which an axial extensometer is mounted and the applied force is varied such that the strain is controlled between set limits in accordance with a cycle of chosen waveform. The magnitude of the applied cyclic force affects the development of microscopic plastic strain within the test section, thus determining the fatigue life. A series of such tests, on nominally identical test pieces allows the relationship between the applied strain and the number of cycles to failure to be established.

The test is continued until a crack develops in the test piece so that the desired failure criterium is reached, or until a specified number of cycles is reached.

The fatigue lives generated are typically less than 100 000 cycles to failure and the test regime is said to be that of low cycle fatigue (LCF).

For the purposes of this document, the following definitions apply:

The test section is defined as the region between the blending fillets at the gripping section of the test piece.

The gauge length of the test piece is the portion of the test section where the extensometer is attached, to measure the strain. The original gauge length is the gauge length measured at test temperature, prior to the application of any force to the test piece. This “hot” gauge length may be determined by monitoring the extensometer output during heating (see 7.2).

The area of the gauge section of the test piece. The cross-section area shall be measured at ambient temperature.

NOTE When designing a component which is intended to operate at elevated temperature, the endurance curves (stress versus number of cycles to failure) need to be corrected to take into account the thermal expansion of the test piece. This correction can be included in the component design calculation code.

A cycle is defined as the smallest section of the strain-time function which is repeated periodically.

This is shown in Figure 1, together with appropriate nomenclature which further defines the strain cycle.

Key

X time

Y strain

a strain amplitude ɛ_a

b mean strain ɛ_m

c minimum strain ɛ_min

d strain range Δɛ

e maximum strain ɛ_max

f one cycle

Figure 1 — Fatigue strain cycle

The stress-strain path during one cycle is called stress-strain loop (see Figure 2).

Key

X stress σ

Y strain ɛ

σ_max maximum stress

σ_min minimum stress

Δɛ_p inelastic strain range

Δσ stress range

Δɛ_t strain range

Figure 2 — Typical stress-strain hysteresis diagram

When the strain cycle includes a hold period at the maximum strain, the test is defined as a creep fatigue test (see Figure 3).

Key

X strain

Y stress

a plastic strain range

b secondary stress range

c primary stress range

Figure 3 — Typical hysterisis diagram for a creep-fatigue test

Failure is defined as the moment when a macro-crack develops in the test piece that is large enough to affect the compliance of the test piece. The number of cycles to failure is conventionally defined as the number of cycles corresponding to a decrease of 10 % in the stress value extrapolated over the maximum stress versus number of cycles curve when the maximum stress falls sharply, see Figure 4a) and Figure 4b). Depending on the material which is tested and the test conditions, the number of cycles to failure may be significantly lower than the number of cycles to complete separation of the test piece into two distinct parts.

a) For materials with steady-state behaviour after hardening

b) For materials with continuous softening

Key

X cycles

Y stress

σ_max maximum stress

N_f fatigue life

N_i number of cycles

X % tbd

Figure 4 — Determination of N_i and N_f

The stress-strain loop which is the closest recorded to half the number of cycles to failure is named mid-life stress-strain loop (or half-life stress-strain loop).

The tests shall be carried out on a tension-compression machine designed for a smooth start-up with no backlash when passing through zero. In order to minimize the risk of buckling of the test piece, the machine shall have great lateral rigidity and accurate alignment between the components used to grip the test piece ends.

The machine loading system shall be a controlled system in which the straining of the test piece is servo-controlled. It may be hydraulic or electromechanical.

During elevated temperature tests the machine load cell shall be suitably shielded and/or cooled such that it remains within its temperature operating range.

The force measurement system shall be verified at intervals not exceeding one year. The method to be used is that of ASTM E1012 with the following amendment, related to the application of test forces, to cover calibration in tension and compression going through zero:

Three series of measurements shall be carried out. Each series shall comprise at least 20 force steps as follows:

a) Five increasing force steps in tension at regular intervals from 20 % to 100 % of the full scale,

b) Ten decreasing force steps at regular intervals from 100 % of the full scale in tension down to the full scale in compression,

c) Five increasing force steps at regular intervals from 100 % of the full scale in compression up to zero.

The relative errors of accuracy, repeatability, reversibility, and zero shall be within the limits stated for class 1.

During the calibration process, an initial calibration shall be performed prior to adjustment of the test machine, such that the effect of any errors outside of the class 1,0 requirement can be understood. If initial errors are present, then the calibration period is to be reviewed accordingly.

The number of cycles applied to the test piece shall be recorded such that for tests lasting less than 10 000 cycles, individual cycles can be resolved, while for longer tests the resolution shall be better than 0,1 % of the indicated life.

NOTE A calibrated timer is a desirable adjunct to the cycle counter. When used to indicate total elapsed time to failure, it provides an excellent check against the cycle counter frequency for a fixed waveform frequency.

The strain applied to the test piece shall be measured and controlled using an axial extensometer attached directly to the gauge section of the test piece.

The geometry of the contact zones and pressure of the extensometer on the test piece shall be such that they prevent slippage of the extensometer without damaging the test piece.

The transducer section of the extensometer shall be protected from all heat fluctuations which are likely to give rise to a drift or fluctuation in the strain signal originating from the heating or cooling system and ambient air. Temperature compensation may offer an additional advantage.

It is suggested that tests be conducted in an enclosure within which the ambient temperature is controlled. The region immediately surrounding the testing machine shall be protected from draughts.

The extensometer, associated with its measuring system, shall be verified at intervals not exceeding one year as class 0,5 of ASTM E1012:2019, for tension as well as compression. In addition, hysteresis shall be equal to or less than 0,1 % of the measuring range.

During the calibration process, an initial calibration shall be performed prior to adjustment of the extensometer conditioner, such that the effect of any errors outside of the class 0,5 requirement can be understood. If initial errors are present then the calibration period is to be reduced accordingly.

The strain cycle waveform shall be constant and is to be applied at a fixed frequency throughout the duration of a test. The waveform generator in use shall have a repeatability such that the variation in specified strain levels between successive cycles is within the calibration tolerance of the extensometer as stated in 5.3.1, for the duration of the test.

A triangular or trapezoidal waveform is generally used for high temperatures. A sinusoidal waveform may be used at lower temperatures if the effect on viscoplastic strains is negligible. The strain rate (preferred) or the frequency shall be held constant within a test programme.

NOTE The range of frequencies for low-cycle fatigue tests is most often between 0,1 Hz and 1 Hz. In terms of the total strain rate, the majority of tests are carried out within the interval ranging from 0,1 %/s to 2 %/s unless the influence of rate on the behaviour of the material is being studied.

General

The gripping device shall ensure that the arrangement of the test piece is reproducible. It shall have surfaces ensuring the alignment of the test piece in order to meet the requirements specified in 5.5.2, and surfaces allowing transmission of tensile and compressive forces without backlash throughout the duration of the test.

To improve the parallelism, concentricity and perpendicularity of the reference surfaces of the grips, the distance between the crosshead and actuator shall be as small as possible.

Materials shall be selected so as to ensure correct functioning throughout the test temperature range.

It is recommended that the tolerances for parallelism, concentricity, and perpendicularity of the reference surface of the grips be less than 0,02 mm in order to achieve the alignment requirements described in 5.5.2. A further benefit can be realized by minimizing the number of mechanical interfaces in the load train.

Alignment verification

Alignment of the load train assembly shall be checked at intervals not exceeding one year or 100 tests, whichever occurs sooner. In addition, it shall be checked following disassembly of the text fixtures, movement of the machine crosshead or following a compressive failure that has caused the two test piece halves to overlap.

It is recommended that the alignment is checked by means of a strain-gauged test piece of geometry identical to that to be tested and that has been manufactured to the same tolerances.

The maximum bending strain determined in accordance with ASTM E1012:2019, Method 1 shall not exceed 5 % of the mean axial strain induced at the lowest maximum tensile force and the maximum compressive force to be encountered in the test program. This criterion shall be meet at each of four positions as the test piece is rotated through 90°.

The use of two sets of strain gauges in groups of four, fixed at 90° intervals around the test section is recommended. The gauges shall be equally distant from the test piece centre line, 3/4 of the parallel gauge length apart. Any strain induced into the gauge length due to the gripping mechanism shall be minimized to less than 100 micro strain.

The “Measurement Good Practice Guide No 1” from The National Physical Laboratory (NPL) is recommended as a good detailed best practice document.

The use of dial gauge indicators in checking alignment shall be avoided. When they are used, the tolerances adopted shall ensure an equivalent alignment error to that obtained using strain gauges. However, bending induced by an aligned, but off-centred load train is not detected by this technique.

Testing is generally conducted in air at ambient or elevated temperatures, although there may be a requirement to test in vacuum or in a controlled atmosphere.

Where additional apparatus is used such as furnaces, chambers, etc., it is essential that the full force indicated by the force indicator is being applied to the test piece and is not being diverted through the auxiliary apparatus (e.g. by friction).

For elevated temperature tests the heating device employed shall be such that the test piece can be uniformly heated to the specified temperature, and an indicated temperature variation along the test section of less than or equal to 4 °C can be maintained for the duration of the test.

A resistance furnace with three control zones is recommended. If a direct induction heating system is used, it is advisable to select a generator of medium frequency (f ≤ 100 kHz) to achieve minimal radial thermal gradient in the test piece.

The temperature measuring system comprising sensors and readout equipment, shall be capable of operating continuously for the duration of the test and have a resolution of at least 1 °C and an accuracy of ± 2 °C. It shall be verified at intervals not exceeding one year over the working temperature range, traceable to national standards by a documented method.

The use of thermocouples is recommended. Annex A describes their method of use.

For short test sections (<25 mm), two thermocouples should be positioned at the extremities of the test section. For longer test sections at least three thermocouples equi-spaced along the test section shall be used.

The variation in indicated temperature anywhere on the test section shall not exceed 4 °C.

For a given set of heating device, thermocouple position, grips, and test piece geometry, this number of thermocouples may be reduced if experience shows that the temperature gradient in the test piece is reproducible within the above tolerances.

The permitted deviations between the specified test temperature and the indicated temperature measured at the surface of the test section are as indicated in Table 1.

Table 1 — Permitted deviations between indicated temperature and specified test temperature

For tests at ambient temperature (10 °C to 35 °C) it is not necessary to measure the test piece temperature. In case of dispute the test shall be performed at a temperature of 23 °C ± 5 °C and the test piece temperature shall be measured.

NOTE The effect of compounding errors could result in the real tolerance in temperature from the specified level to be 3 °C greater.

The temperature rise due to plastic deformation shall be minimized (see 7.3.1) and shall be compensated for within the Table 1 tolerances.

An automatic microcomputer system capable of carrying out the task of collecting data and processing it simultaneously shall be used to monitor stress, strain and temperature as a function of time. This information is required particularly to determine failure and to monitor changes in stress that occur during hold periods.

This system will be used to plot stress-strain loops of force versus deformation or stress versus strain. The sampling frequency of stress-strain data points shall be sufficient to ensure adequate definition of the stress-strain loop specially in the regions of strain reversal (at least 200 data points per loop).

The system described above may be replaced by a strip chart recorder and a potentiometric X-Y recorder. The recorders shall be used only when the test conditions result in a maximum pen velocity that will not cause inaccurate records e.g. less than half of the recorder's slewing speed. For higher frequencies, a digital storage oscilloscope or an oscilloscope equipped with a camera are acceptable alternatives to the X-Y recorder.

The indicated test temperature shall be monitored throughout the test, within the accuracy stated in 5.7. A temperature observation shall be made at least every 5 min.

The whole system of transducers, associated electronic conditioning or amplification and data recording shall be calibrated as an entity over their working range at intervals not exceeding one year and the deviations recorded on the calibration certificates.

The accuracy of all recording systems shall be kept within 1 % of full scale.

The type of test piece used depends on the objectives of the test program, the equipment capacity and the form in which the material is available. The design, however, shall meet certain general criteria as outlined below, in order to ensure a uniform distribution of stresses and strains in the gauge section and the localization of failures inside the gauge length.

The geometry of the test piece shall fulfil the following conditions:

a) it shall have a parallel length longer than the extensometer base length, to allow strain measurements to be taken using an axial extensometer. However, it shall not be excessively long, to avoid the occurrence of failure outside the extensometer base length;

b) it shall be sufficiently compact to avoid the risk of compressive buckling and sufficiently slender to avoid failure at the fillet;

c) it shall ensure uniform distribution of stresses and strains over the whole gauge length;

d) it shall have reference surfaces to ensure correct alignment.

Taking into account these requirements, the geometric dimensions mentioned in Figure 5 are recommended.

The diameter of the gauge section of the test piece shall not be less than 4,5 mm and shall be held constant in a test program to reduce test result variability.

When testing coarse grain alloys (for example cast alloys) the test piece diameter shall be at least five times the average grain size.

For alloys that contain a small number of randomly-dispersed metallurgical defects, e.g. powder metallurgy alloys, the test piece size shall be chosen such that the probability of finding a defect in the gauge section is sufficiently high. The ratio of gauge length to diameter is also an important parameter since near-surface defects lead to failure sooner than internal defects.

When characterizing thin wall cast alloys like turbine blade alloys, it may be necessary to use test pieces having the same thickness as the real part. It this case a tubular test piece is recommended.

It is important for the general tolerances of the reference surfaces and test section of the test pieces to show the three following properties:

— parallelism: 0,02 mm;

— concentricity: 0,02 mm;

— perpendicularity: 0,02 mm;

(these values are expressed in relation to the axis or reference plane).

The recommended gripping heads are as follows:

— shouldered heads (preferred);

— threaded heads;

— smooth cylindrical heads (with hydraulic jaws).

Recommended dimension of cylindrical test pieces are shown in Table 2.

a) Plain test piece

b) Test section profile for a cylindrical test piece

Figure 5 — Test section profile for cylindrical test pieces

Table 2 — Recommended dimension of cylindrical test pieces

The position and orientation of test piece blanks cut out of components or billets can have a significant effect on the fatigue properties of a material. It is therefore important that their identity is maintained throughout the test piece manufacture process, and that this is traceable to their position and orientation in the original material stock. Reference to EN ISO 3785 is recommended.

Each test piece blank and ultimately each test piece shall therefore be suitably marked in a reliable manner. The test piece shall be marked at each end away from the test section, such that the two halves can be identified post-fracture.

Machined test pieces shall be stored in a manner that protects them from mechanical damage such a scratching, and environmental effects such as extreme humidity, etc.

Throughout the testing process, any special handling requirements for the material under investigation shall be adhered to. The use of clean cotton gloves is recommended.

The condition of the test piece and method of preparation are of the utmost importance. Inappropriate methods of preparation, which may be material specific, can greatly bias the data generated. While it may be the purpose of some tests to establish the effect of a particular representative surface finish, for standard test pieces the following guidelines shall be adhered to.

The technique established and approved for a specific material and test piece configuration shall not be changed without first demonstrating that no bias is introduced by the alternative technique.

The final machining of the test pieces shall be performed in a manner consistently producing a smooth surface with low residual compressive stresses. The recommended procedure, for test pieces with circular cross section, comprises a fine turning or low stress grinding sequence followed by longitudinal polishing (see Annex B for example of machining sequence which may be used in the production of test pieces). The final polishing methods employed shall eliminate all circumferential machining marks or scratches on the test piece gauge length or end transitions. A low-magnification examination (x 20) is recommended as a final inspection check.

The magnitude of residual compressive stress at the surface of the test piece shall be less than 500 MPa. Moreover, after removing 10 µm from the surface of the test piece, the magnitude of residual compressive stress shall be less than 200 MPa, and at 50 µm from the surface less than 50 MPa.

The effect of contaminants such as cutting fluids and degreasing agents is also to be understood.

NOTE Assurance that compressive residual stresses are maintained at a low level throughout the manufacturing route can be achieved by the use of X-ray residual stress measurement techniques.

The dimensions used for calculating the cross-sectional area of the test piece shall be measured prior to the test on individual test pieces, to an accuracy of 0,2 % or 0,005 mm, whichever is the greater value. The integrity of the surface finish shall not be jeopardised during this activity; the use of an optical measurement method is highly recommended.

The dimensions (outer diameter and possibly inner diameter) of the test section shall be measured at three positions along the gauge length. The average of these values are used to calculate cross sectional area.

Applied stresses shall be calculated based on ambient temperature measurements and no compensation is to be made for the change in section and effective stress due to heating for elevated temperature tests, or due to deformation of the test section during the test.

When designing a component which is intended to operate at elevated temperature, the endurance curves (stress versus number of cycles to failure) need to be corrected to take into account the thermal expansion of the test piece. This correction shall be included in the component design calculation code.

The method employed to insert the test piece into the test fixture shall not jeopardize the alignment mechanisms, surface finish integrity or material properties. Excessive torsion shall be avoided and compressive stresses within the test section limited to a maximum of 50 % of yield strength.

Prior to testing the test piece shall be degreased and kept clean; the requirements given in Annex C shall be used.

Where a contacting extensometer is used, the contact pressure shall be high enough to prevent extensometer slippage, but shall not be so high that test piece failure initiates from the contact position. The mounting system of the extensometer shall be rigid enough so that the contact pressure is kept constant throughout the test.

In order to identify possible problems within the force or strain measuring systems before starting a test, a good practice is to measure the modulus of elasticity of the material at ambient temperature, limiting loading to 50 % of yield stress. This value shall not deviate by more than 10 % from the expected value.

The test piece shall be heated to the specified temperature at a rate not exceeding 50 °C/min and shall be maintained at that temperature for a sufficient period to ensure that the temperature has fully stabilized. In general, a period of 30 s per mm² of cross-sectional area shall be allowed, with a minimum period of 15 min.

If the total time to reach and stabilize at the test temperature exceeds 12 h, then the actual soak time shall be reported.

During the heating process, the temperature of the test piece shall not exceed the specified temperature within the tolerances outlined in 5.7.

Expansion during the heating process shall not result in compressive forces being applied to the test piece.

The force applied to the test piece shall therefore be controlled throughout the heating process and shall not exceed 10 % of yield stress.

The extensometer gauge length at test temperature shall be determined by monitoring the thermal strain recorded by the extensometer under zero force after the temperature stabilizes at test temperature. Then the strain output shall be set to zero.

In order to identify possible problems within the temperature or strain measuring systems before starting a test, it is advisable to check during heating the mean coefficient of thermal expansion of the material. This coefficient should not deviate by more than 10 % from the expected value.

In order to check the test set up, it is recommended that the initial modulus of elasticity E_o be measured at the test temperature prior to start the test, limiting loading to 50 % of yield stress.

The same waveform shall be retained throughout the whole test programme unless the aim of this programme is to study the effect of the waveform on the behaviour of the material.

The strain rate (or frequency) shall be low enough to allow the heating device to compensate for the temperature rise in the test piece due to straining such that the test piece temperature stay within the limits stated in Table 1. The same strain rate (or the same frequency) shall be retained throughout the whole test programme.

Unless the purpose of the test is to assess the effect of this parameter, the initial loading shall be in the tensile direction.

The transfer to the strain control mode from the force control mode shall be made carefully, without backlash prejudicial to the rest of the test.

Prior to commencing cycling the waveform generator shall be set such that the achieved maximum and minimum strains are between 95 % and 100 % of the intended values.

On commencing cycling, the specified strain shall not be adjusted in order to achieve the precise intended strain cycle. Actual strains and strain ratios shall be recalculated based on measured values. lf the test piece is overstrained during the start of the test; the strain range or maximum strain shall not be reduced. However, in the cases where the specified maximum and minimum strains are mandatory, the specified strain shall be adjusted so that the correct maxima and minima are achieved within the first 10 cycles. This shall be stated in the test report.

The achieved strain rate (or frequency) shall be within ± 10 % of the specified strain rate (or frequency).

In the case of trapezoidal or triangular waveforms, the discontinuities at the nodes shall be well defined and angular. Any adjustment to the waveform shape shall be complete within 100 cycles or

1 % of the expected life, whichever is smaller, and shall not introduce deviations from the specified waveform that exceed the limits defined below.

Thereafter it shall remain unadjusted throughout the duration of the test. Oscillations and rounding features at the nodes shall not exceed 1 % and 5 % of the intended strain respectively, and in the case of trapezoidal waveforms, shall not constitute more than 20 % of the hold time or a maximum of 0,2 s, whichever is smaller (see Figure 6).

Key

x ≤ 20 % of the specified hold time (max of 0,2 s)

y ≤ 1 %, of the intended force amplitude

y' ≤ 5 %, of the intended force amplitude

a specified hold time

Figure 6 — Force waveform optimization

In the case of sinusoidal waveforms, the waveform shall be smooth and free from discontinuities.

In the event of an inadvertent or accidental stop of the test, no attempt shall be made to restart is unless:

a) the extensometer has not slipped, and

b) the test piece has not been damaged by the stop (a return to zero strain may strain the test piece plastically, causing irreversible damage).

These two points may be verified by analysing the recordings. If the test is restarted it shall be reported.

The indicated test temperature and the achieved minimum and maximum strains shall be monitored throughout the test in order to ensure that the controlled test parameters have remained within the limits specified in 5.4 and 5.7.

The achieved minimum and maximum forces shall be monitored continuously or frequently enough to allow a precise determinate of the number of cycles to failure.

At the start of the test, a continuous recording shall be made of the initial stress-strain loops. During the course of the test a periodic recording is sufficient. The frequency of these recordings shall be appropriately chosen for the expected overall duration of the test. The option generally used consists in recording approximately the first ten cycles and then recording individual cycles based on a logarithmic increase (16, 25, 40, 63, 100, 160, etc.).

NOTE If there is a system of automatic acquisition, the acquisition of loops can be programmed either with a predefined interval or as a function of the progression of each of the two parameters (stress, strain).

In the absence of any particular specification, the failure criterion shall be a decrease of 10 % in the stress value extrapolated over the maximum stress versus number of cycles curve when the maximum stress falls sharply, as defined in 4.1.7. By agreement, complete separation of the test piece into two distinct parts may be used in place of this criterion (particularly for acceptance/rejection tests). The use of complete separation for failure criterion shall be reported.

Tests shall be continued without interruption until the failure criterion (as defined above) is reached, or until a predetermined number of cycles has been exceeded, and shall be stopped shortly after, in order to avoid any damage of the fracture surface detrimental to the identification of the initiation site. Heating of the test piece shall be stopped at the same time.

Annex D defines several failure criteria that have been used for determining the number of cycles to failure. The criterion that is defined in 4.1.7 shall be used for the purpose of this document.

In order to validate the test, the test temperature record and the strain record shall be consulted to ensure that there were no deviations outside the limits specified in 5.4 and 5. 7.

The location of the crack initiation site shall be identified. The following distinctions are to be made:

a) initiation within the extensometer gauge length;

b) initiation under an extensometer;

c) initiation in the gauge section, outside the extensometer gauge length;

d) initiation in the transition radius or in the test piece heads.

The test shall be considered valid when the crack initiates within the extensometer gauge length for the determination of N_t, Initiation in the gauge section, outside the extensometer gauge length gives a lower estimation of N_f100.

NOTE A more thorough examination, using a scanning electron microscope, can be necessary to see if the crack initiates from surface scratches generated by poor machining or handling procedures, or thermocouple interference and to identify the initiation site (inclusion, porosity, ...).

The number of cycles to failure N_f shall be determined by means of examination of the record of stress versus time (or maximum force versus number of cycles), as the number of cycles corresponding to a decrease of 10 % in the stress value extrapolated over the maximum stress versus number of cycles curve when the maximum stress falls sharply (see Figure 4).

NOTE The number of cycles to initiation of a macro-crack N_f0 can also be determined by examination of the records, provided the fatigue crack initiates inside the extensometer gauge length.

On the first cycle, measure the maximum stress, minimum stress and initial modulus of elasticity (if it has not been measured prior to starting the test).

Select a cycle where the stress-strain loop has been recorded, as close as possible to mid-life, N_f/2.

The mid-life modulus of elasticity Ern is determined by measuring the slope of the stress-strain diagram in the portions showing elastic behaviour after the maximum and minimum strain and averageing the two measurements.

Total, elastic and inelastic strain ranges are determined in the following way:

total strain range is measured:

∆ Ɛ_t = (Ɛ_max − Ɛ_min) (1)

— elastic strain range is calculated by dividing the stress range by the mid-life modulus of elasticity:

∆ Ɛ_e = ∆ σ /E_m (2)

— inelastic strain range is obtained by the difference:

∆ Ɛp = ∆ Ɛt – ∆ Ɛe (3)

NOTE When the linear elastic portion of the loading and unloading parts of the stress-strain loop are greater than the stress amplitude, the following method can be used as an alternative:

— total strain range is measured:

∆ Ɛ_t = (Ɛ_max. – Ɛ_min.) (4)

— inelastic strain range (∆ Ɛ_p) is measured as the width of the stress-strain loop at mean stress.

— elastic strain range is obtained by the difference:

∆ Ɛ_e = ∆ Ɛ_t - ∆ Ɛ_p (5)

The test report on each test piece shall state:

a) reference to this document;

b) material specification;

c) test piece identity, drawing number, and reference to a documented method of preparation;

d) temperature of the test plus any deviation from the specified limits;

e) waveform and strain ratio;

f) frequency or total strain rate;

g) characteristics of the first cycle: — Total strain range;

— Maximum stress;

— Minimum stress.

h) any adjustment of the strain range at the beginning of the test;

i) characteristics of the mid-life cycle:

— number of the cycle that was examined;

— total strain range;

— elastic strain range;

— strain range;

— maximum stress;

— minimum stress;

— mid-life modulus of elasticity.

j) number of cycles to failure N_f, number of cycles to complete separation N_f100, or total number of cycles N_t;

k) fracture location relative to the extensometer probes and transition radii;

l) any deviation from this document that might have an influence upon the test result.

The following information shall be included in the test report;

a) material composition, heat treatment, microstructure;

b) complete identification of the part or semi-finished product from which the test pieces are taken;

c) precise position and orientation of each test piece;

d) predominant material orientations due to the manufacturing process, such as rolling direction or casting direction;

e) testing machine, heating device, extensometer, gauge length of the extensometer;

f) initial modulus of elasticity, mean coefficient of thermal expansion;

g) number of cycles to initiation;

h) characteristics of additional cycles:

— cycle number;

— total, elastic and inelastic strain ranges;

— minimum and maximum stresses;

— modulus of elasticity.

i) fractographic examination of the two fracture surfaces to identify the initiation site and to determine any unusual causes of failure that might invalidate the test result.

The most popular presentation of results for a series of related tests is via a graphical Ɛ-N_f diagram. Its construction involves plotting the number of cycles to failure as the abscissa and the strain amplitude as the ordinate, with logarithmic scales. Such a diagram may be represented by the following relationship:

Ɛ_ta = Ɛ_ea + Ɛ_pa = (σ'_f/E_m) (2N_f)^b + Ɛ'_f (2N_f)^c (6)

where the parameters are defined in accordance with Table 3.

Table 3 — Stress-fatigue life and strain-fatigue life relationships

1. 
   (informative)
   
   Use of thermocouples
   1. Calibration

Thermocouples should be made from batches of wire that have been calibrated over the whole working range against the recognized fixed points for thermocouples calibration or by comparison with a similarly calibrated and carefully maintained secondary standard reference thermocouple.

• 1. Application

For short test sections (<25 mm), two thermocouples equally spaced along the test section are generally sufficient to guarantee uniformity of the temperature of the test piece unless prior verification has been performed in accordance with 5.6.1. For longer test sections at least three thermocouples should be used.

The thermocouple junctions should be maintained in close thermal contact with the surface of the test piece and be suitably screened from direct radiant heating caused by the heating system. However, they should not be welded to the test section or affect the test section in any way.

The accuracy of thermocouples can be affected by radio interference from induction coils. It is therefore recommended that they are not used as the only measurement system when induction heaters are used.

NOTE 1 Heat-resistant string can be used to tie the thermocouple to the test section.

Due to the time-temperature dependent nature of degradation of thermocouple performance, they should be periodically checked in order to ensure that measurement accuracy is not impaired, in line with good calibration practice.

1. 
   (informative)
   
   Test piece preparation
   1. General

In order to determine basic material properties, it is essential that the roughness of the surface contributes as little as possible to the fatigue failure and that the surface residual stresses and microstructural modifications are minimized.

• 1. Machining the test piece blank

Machine test pieces in the fully heat-treated condition. For test pieces that cannot easily be machined in this condition, give the final heat treatment prior to finish machining.

NOTE The fatigue properties of metals are often sensitive to microstructure, which can depend on the cross-section at the time of heat treatment. The user can ensure himself that the microstructure of the test piece is similar to the component in question.

In all other cases, take steps to ensure that any cutting or rough turning operation does not alter the metallurgical structure of the test piece, i.e. remove metal in cuts of decreasing depth to minimize work hardening of the surface.

• 1. Machining of the test piece
     1. General

The gauge length of the test piece and its transition radii should be machined following a procedure that minimizes the residual stresses at the surface of the test piece.

During the finish machining stage, at least 0,5 mm from the test piece diameter should be removed. The test piece diameter should be decreased to 0,025 mm over the final diameter, the final 0,025 mm in diameter being removed by polishing.

Low stress grinding followed by longitudinal polishing is the procedure that is used in most of the cases. However, particularly for titanium and aluminium alloys, turning may be an alternative to grinding. The chosen method should adhere to one of the following procedures.

• • 1. Turning

Remove metal by cuts of progressively decreasing depth. No single cut should exceed 0,4 mm and the final pass should not exceed 0,05 mm. The following sequence is suitable:

0,4 mm, 0,25 mm, 0,125 mm, 0,075 mm, 0,05 mm.

NOTE Cuts of less than 0,08 mm can result in squeezing rather than cutting in some cases, especially if the cutting edge is not renewed; this can be avoided.

• • 1. Grinding

A suitable lubricant should be used and its flow should be sufficient to prevent heating of the surface. The abrasive particles should be continuously removed from the lubricant.

The diameter of the test piece should be reduced at a rate of no more than 0,5 µm per turn (plunge grinding) or 0,005 mm per pass (traverse grinding) for the last 0,025 mm.

The grinding wheel should be frequently dressed.

The characteristics of the wheel and maximum grinding speed should be selected as a function of the alloy to be machined: a silicon carbide abrasive is suitable for titanium and aluminium alloys whereas and aluminium oxide abrasive is suitable for steels, nickel or cobalt base alloys.

• 1. Polishing of the test section

The final 0,025 mm in diameter should be removed by longitudinally polishing the test piece.

The polishing paper should be renewed periodically. The force applied to the polishing paper should be constant and low (2 N to 5 N). An automatic polishing device is recommended.

Recommended metal removals and paper grit are as follows:

— 0,012 mm with P800 paper (particle size: 20 µm to 30 µm);

— 0,008 mm with P1 000 paper (particle size: 15 µm to 25 µm);

— 0,005 mm with P1 200 paper (particle size: 10 µm to 2 µm).

All grinding or turning marks should be removed with the P800 paper before using a finer grit. Polish with each finer grit paper until all marks left by the previous paper are removed, and to impart a maximum surface roughness Ra of 0,2 µm in the test piece axis direction.

NOTE Extreme caution can be exercised in polishing to ensure that material is being properly removed rather than merely smeared to produce a smooth surface.

1. 
   (normative)
   
   Guidelines on test piece handling and degreasing
   1. General

Unless otherwise specified by the customer or where special surface treatments have been applied, the following guidelines shall be adhered to:

• 1. Steels

Degrease by full submersion in an acetone bath immediately prior to testing. The use of an ultrasonic bath is recommended. Test pieces that are degreased and then not tested for some reason shall be retreated with an anti-corrosion fluid.

No special handling requirements are necessary subsequently.

Thread lubricants are not necessary at the test temperatures encountered.

• 1. Nickel and cobalt base alloys

Degrease the test section with acetone and wipe dry with a clean soft cloth. No special handling requirements are necessary subsequently.

For tests at temperatures greater than 650 °C, a thread lubricant shall be used sparingly. Any excess evident once the test piece has been inserted shall be removed.

• 1. Titanium base alloys

Degrease the test section with acetone and wipe dry with a clean soft cloth.

Once degreased, the test section shall not be touched other than with clean cotton gloves. In addition, everything which touches the test section (extensometer probes, string used to tie on thermocouples, etc.) shall be clean and shall be handled with gloves (to prevent any transfer of salt from the skin to the test piece).

Thread lubricants shall not be used.

• 1. Aluminium and magnesium base alloys

Degrease the test section with acetone and wipe dry with a clean soft cloth. No special handling requirements are necessary subsequently.

Thread lubricants are not necessary at the test temperatures encountered.

1. 
   (informative)
   
   Failure criteria

There are various ways of determining a failure criterion, more often associated with an end-of-test criterion than with the total failure of the test piece. It may depend on the interpretation of the fatigue test result and on the nature of the material being tested. In fact, any development in the material behaviour should be taken into account during the test {hardening or softening) and consequently the appropriate end-of-test criterion selected. The failure criteria under consideration are generally based on the appearance, presence or intensification of a phenomenon which has been observed or recorded, and which indicates severe damage or imminent failure of the test piece.

Several criteria may be considered. The conventional number of cycles to failure is defined as the number of cycles corresponding to:

1) Total failure of the test piece: separation into two distinct parts.

2) A certain percentage decrease in the maximum tensile stress in relation to the level determined during the test.

3) A certain percentage decrease in the modulus of elasticity ratios in the tensile and compressive part of the hysteresis loops.

4) A certain percentage decrease in the maximum tensile stress in relation to the maximum compressive stress.

The use of criterion 2) based on a certain percentage decrease in the maximum tensile stress is recommended. This criterion allows a reduction in the scatter observed in the number of cycles-to complete-failure of the test piece and may be applied to different stress-strain behaviours (hardening, softening, stable behaviour, see Figure 4). In this case, the usual number of cycles to failure N_f is defined as the number of cycles corresponding to a decrease of x % in the stress value extrapolated over the tensile stress-versus number of cycles curve when the stress falls sharply. The recommended value of x is 10.

Criterion 2) relates to the presence of one (or more) macroscopic crack(s) in the test piece. In general, the fraction of cracked surface in the cross-section of the test piece is of the same magnitude as the fraction of stress decrease.

Bibliography

[1] EN 4258, Aerospace series — Metallic materials — General organization of standardization —Link between types of European Standards and their use

[2] EN ISO 3785, Metallic materials — Designation of test specimen axes in relation to product texture(ISO 3785)

[3] NPL MMS 001:1995,^[2] Code of Practice for the Measurement of Bending in Uniaxial Low Cycle fatigue Testing

1. www.astm.org. ↑

2. www.npl.co.uk. ↑