Available online at www.sciencedirect.

com

Procedia Engineering 10 (2011) 11841189

ICM11

Manufacturing influences on the fatigue properties of

quenched and tempered SAE 4140 specimens
M. Klein*,P. Starke, D. Eifler
Institute of Materials Science and Engineering, University of Kaiserslautern, D-67663 Kaiserslautern, Germany

Abstract

To analyse interactions between the manufacturing process chains, its design, variations in material properties,
especially the microstructure and the resulting component properties, specimens with tension screw geometry were
machined with specific varying process chains. According to the specific process chain the surface integrity and
consequently the fatigue behaviour of the specimens is influenced in a characteristic manner.
The cyclic deformation behaviour of these components can be benchmarked with conventional strain measurements
as well as with high-precision temperature and electrical resistance measurements. The development of temperature-
values provides substantial information. Furthermore, electrical resistance measurements enable the monitoring of
proceeding fatigue damage during scheduled maintenances without any mechanical load. The physically based
fatigue life calculation method PHYBAL was modified to calculate the lifetime.
2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Selection and peer-review under responsibility of ICM11
Keywords: Fatigue behaviour; Mechanical stress-strain hysteresis measurements; Temperature and electrical resistance
measurements;Steel SAE 4140; Microstructure; Automotive components; Process chains

Nomenclature
PC process chain
EA extension areas
CAT constant amplitude test
LIT load increase test

* Corresponding author. Tel.: +49-631-205-5057; fax: +49-631-205-2137.

E-mail address: [email protected].

1. Introduction

The demands on highly loaded automotive components are increasing steadily. For this reason,

1877-7058 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Selection and peer-review under responsibility of ICM11 doi:10.1016/j.proeng.2011.04.197
 M. Klein et al. / Procedia Engineering 10 (2011) 11841189 1185

components with higher efficiency are required for the competitive ability of metallic construction units.
Therefore a comprehensive understanding of the fatigue behaviour of metallic materials and components
as well as the development of new methods for reliable fatigue life calculations are of great importance.
The fatigue properties of components are a multi parameter property and strongly affected by the surface
integrity, since the surface and the near surface zones are the most stressed component areas. The surface
integrity is influenced by each manufacturing process, its parameters, the process order and the process
combination. Because of their high complexity, the influence-cause-effect-relations of process chains are
not understood satisfactorily until now [1]. Investigations concerning the design of process chains as well
as manufacturing alternatives are mentioned in [2 - 4].
The main purpose of the present investigation is the evaluation of the influence of the individual
process chain design on the surface integrity and consequently on the fatigue properties of highly loaded
metallic specimens with tension screw geometry of the quenched and tempered steel SAE 4140. In
particular, microscopic changes at the surface and the near surface areas were regarded as a function of
the entire process chain from the semi-finished parts to the finished specimens.

2. Material and process chains

To estimate the influencing factors of the individual process, its parameters and of the design of the
process chain on the fatigue behaviour, specimens with tension screw geometry (Fig.1a) were machined.
Five typical manufacturing process chains (PC), consisting of the processes grinding, turning and heat
treatment, were applied (Fig. 1b). The studies are focused on the fatigue relevant extension areas (EA).
The investigated SAE 4140 was supplied in a quenched and tempered condition as round steel bars with a
diameter of 25 mm. The accordance of the chemical composition to ASTM A 829 M standard was
verified. The microstructure is characterised by a tempered martensite with ferrite and fine dispersed Fe3C
carbides.

a) b)

Fig. 1. Specimen with tension screw geometry (a), process chains PC 1-5 (b)

After normalisation (Ta = 860C, t = 120 min.) the components of PC 1 and 2 were soft-turned and in
step three quenched and tempered (Ta = 850C, t = 120 min.; T = 550C, t = 90 min.). The extension
areas of PC 1, 3, 4 and 5 were machined with only one chipping process: soft turning, hard turning or
grinding. The extension areas of PC 2 were machined in two steps (soft turning with an allowance of
300 m, grinding) and consequently lower chipping volume.
1186 M. Klein et al. / Procedia Engineering 10 (2011) 11841189

3. Experimental setup

Light and scanning electron microscopy, surface roughness, micro hardness and X-ray-measurements
as well as fatigue tests were carried out. The experimental setup for the fatigue tests is shown detailed in
[5]. In addition to stress-strain hysteresis measurements Ha,p, deformation-induced changes in temperature
'T [6-7] and changes in electrical resistance 'R [8-10] were measured with high precision.
The mechanical stress-strain hysteresis measurements were carried out with a commercial
extensometer. The change in temperature was measured with four thermocouples (T1-4). The
thermocouples T1 and T2 were fixed in the middle of the extension areas, T3 and T4 at the elastically
loaded specimen shafts. The local change in temperature for EA 1 and EA 2 can be evaluated with the
subtraction of the mean of T3 and T4 from T1 and T2 respectively. For the change in electrical resistance
measurements, a DC-power supply was connected with the shafts and 'R was measured integrally across
the extension areas with two wires spot-welded at the transition between extension areas and shafts. The
experiments yield cyclic deformation Ha,p-N, temperature 'T-N and electrical resistance 'R-N curves
representing the actual fatigue state.
CATs interrupted by load-free sequences were performed to separate the influence of 'T on 'R.
During these load-free sequences a fully cooling down of the specimen is guaranteed. Increasing 'Rload-
free-values indicate proceeding fatigue damage in specimens and components in a suitable and non-
destructive manner.

4. Results

After each machining process, light and scanning electron micrographs as well as X-ray-measurements
were performed to analyse the influences of each manufacturing process and its parameters on the surface
and the near surface area. Due to the grinding process, the near surface area of the specimens machined
according to PC 2 is characterized by significant plastic deformation reaching to a depth of 3 m (Fig.
2a). Compressive residual stresses Vrs = -330 MPa were measured. The grinding process in PC 5,
inducing compressive residual stresses of -350 MPa, leads to changes in the microstructure up to a depth
of about 7 m. The specimens in PC 3 show microstructural changes with two sections. The primary (P)
zone near the surface up to a depth of 2 m where the carbide lines are compressed and oriented
according to the machining direction. In the secondary (S) zone for another 3 m the carbide lines are
strongly deformed (Fig. 2b).
a) b)

Fig. 2. Changes of the near surface area in the extension areas due to the process chains PC 2 (a) and PC 3 (b)

The feed rate v = 50 m in combination with an infeed ap = 0.6 mm leads to a residual-stress-free-state.

The double feed rate and half infeed used in PC 4 lead to tensile residual stresses of about +430 MPa.
 M. Klein et al. / Procedia Engineering 10 (2011) 11841189 1187

The surface roughness for the PC 2-5 was determined in average to 5 m. The micro-hardness-values of
the grinded surfaces are slightly higher in comparison to the turned ones. Fig. 3a shows the development
of the plastic strain amplitude measured in EA 2, 'T measured individually in EA 1 and EA 2 and 'R
measured integrally across both extension areas for a CAT with a stress amplitude of 440 MPa of a
specimen machined according to PC 4.
a) b)

Fig. 3. Development of plastic strain amplitude Ha,p, change in temperature 'T and change in electrical resistance 'R in a constant
amplitude test for a specimen with tension screw geometry of PC 4 (a), cyclic temperature curves and residual stresses for
specimens with tension screw geometry of PC 3, PC 4 and PC 5 (b)

Due to the proceeding fatigue damage in EA 2 a change in the slope of 'T2 and 'R can be observed at
4.6104 cycles. The course of EA 1 ('T1) increases at significant higher number of cycles, right before
failure caused by thermal conductions. Temperature and electrical resistance measurements indicate
fatigue failure slightly earlier than plastic strain amplitude measurements. This is caused by the very
small and strongly localized plastic deformation in high-strength steels leading to very small plastic strain
amplitudes which are extremely difficult to measure with conventional extensometers. In contrast the
local application of change in temperature and change in electrical resistance allow to identify precisely
fatigue critical spots of the specimen. Due to the detection of fatigue damage on the basis of temperature
and electrical resistance measurements in a very early fatigue state components can be replaced in a
sufficient and save distance to final failure at a defined change in the slope of 'T- and 'R-curves.
Specimens with higher (lower) 'T-values reach lower (higher) number of cycles to failure. With
increasing residual stresses from compression to tension, 'T increases and the lifetime consequently
decreases (Fig. 3b).
The change in temperature and electrical resistance for a CAT (Va = 500 MPa) interrupted by load-free
sequences (Va = 0 MPa for t = 30 min.) is plotted for a standard fatigue specimen versus time in Fig. 4a.
While 'T and 'R are measured during cyclic loading, 'Rload-free is measured at the end of the load-free
state. As can be seen, during the load-free sequences the change in temperature decreases to 'T = 0 K. In
contrast, 'Rload-free is characterized by a consistent increase caused by the increasing deformation-induced
changes of the microstructure in the bulk material as consequence of proceeding fatigue damage.
To verify the application potential of electrical resistance measurements for online monitoring in
service and even for microstructure-induced damage detection during scheduled maintenances, 'Rload-free
measurements were applied on specimens with tension screw geometry. The 'Rload-free-N-curve shown in
Fig. 4b provides explicit information about the actual fatigue state. The loading of the PC 5 specimen
with the stress amplitude 480 MPa leads to a significant change in the slope of 'Rload-free after
N = 500 cycles.
1188 M. Klein et al. / Procedia Engineering 10 (2011) 11841189

a) b)

Fig. 4. Development of the change in temperature and change in electrical resistance, fatigue specimen (a), development of the
change in load-free electrical resistance in a constant amplitude test for a specimen with tension screw geometry

The physically based fatigue life calculation method PHYBAL affords the calculation of Woehler
(S-N) curves of metallic specimens on the basis of generalized Morrow and Basquin equations with data
taken from only one load increase test (LIT) and two constant amplitude tests (CATs). This procedure is
explained in detail in [11].
a) b)

Fig. 5. Cyclic stress-temperature (CST) curve of a load increase test, of two constant amplitude tests and calculated Va, 'T-relation
of constant amplitude tests (a), calculated and experimentally determined Woehler (SN) curves (b)

The cyclic stress-temperature (CST) curve of the load increase test (LIT) is shown in Fig. 5a. The load
levels in the range of 400 MPa < Va < 480 MPa are presented as a function of the mean values of 'T of
the EAs (). Furthermore, the data points of the Va, 'T-relation of the two constant amplitude tests
(CATs) at 104 cycles are plotted ({). At first the ratio between the measured values of the CAT at 104
cycles and the appropriate load levels of the LIT was determined dividing the CAT-value by LIT-value.
The CST-curve for constant amplitude loading () was calculated by multiplying the values of the load
increase test with the respective ratio. The Morrow-parameters as well as the Basquin-exponent can be
determined out of the calculated CST-curve. The Basquin-factor can be estimated using the number of
cycles to failure of one of the constant amplitude tests in the Basquin-equation. Finally, the Woehler-
curve can be calculated on the basis of these data. As shown in Fig. 5b, the conventionally determined
Woehler (SN) curve and the Woehler (SN) curve calculated on the basis of the change in temperature
of only three fatigue tests agree excellently.
 M. Klein et al. / Procedia Engineering 10 (2011) 11841189 1189

5. Conclusions

The individual manufacturing process, its parameters, the process order and the process combination
influence the surface integrity and consequently the fatigue behaviour in a characteristic manner. The
plastic strain amplitude, temperature and electrical resistance measurements can be used for the fatigue
assessment of specimens with tension screw geometry. Temperature and electric measurements are not
limited to a defined gauge length. Therefore these measurement techniques can be applied even on
complex component geometries. The measured values can be used as input data for the physically based
fatigue life calculation method PHYBAL. Even though the geometry is more complex than a standard
fatigue specimen, the calculated lifetimes are in a very good accordance with the conventionally
determined S-N curves.
Significant changes in the arrangements of the microstructures were observed as a result of different
soft- and hard-turning as well as grinding processes. The process type and the process parameters,
respectively, lead to different developments of microstructure and the resulting amount and sign of
residual stresses. With increasing residual stresses from compression to tension the lifetime decreases.
Due to the detection of fatigue damage on the basis of temperature and electrical resistance
measurements in an early fatigue state, components can be replaced in sufficient distance to collapse as
specified changes in the slope of 'T- and 'R-curves occur.
'Rload-free measurements provide very useful information about the proceeding fatigue damage during
scheduled maintenances without any mechanical load.

References

[1] Klein D, Thoben K-D, Nowag L. Using indicators to describe distortion along a process chain. Mat.-wiss. u. Werkstofftech.
2006, 37, 19.
[2] Van Luttervelt C A, Childs T H C, Jawahir I S, Klocke F. Present situation and future trends in modelling of machining
operations. CIRP Annals Manuf. Tech. 1998, 47, 587.
[3] Schmidt B-C, von Gleic C F. Technical supply chain planning. CIRP J. Manuf. Syst., 2003, 32, 5.
[4] Warnecke G, Eifler D, Aurich J C, Mauren F, Klein M. Investigation of interactions within manufacturing process chains.
Res. and Develop. in Germany, 2005, 12, 85.
[5] Klein M, Walther F, Eifler D. Einfluss der Fertigung auf das Ermdungsverhalten hochbeanspruchter Dehnschrauben.
Proceedings Werkstoffprfung, 2007, 243.
[6] Luk P, Klesnil M. Cyclic stress-strain response and fatigue life in metals in low amplitude region. Mat. Sci. Eng. A,
1973, 11, 345.
[7] Strk K F. Temperaturmessung an schwingend beanspruchten Werkstoffen. Z. Werkstofftech., 1982, 13, 333.
[8] Polk J. Electrical resistivity of cyclically deformed copper. Czech. J. Phys., 1969, 19, 315.
[9] Charrier J, Roux R. Evolution of damage fatigue by electrical measure on smooth cylindrical specimens. Nondestr. Test.
Eval., 1991, 113.
[10] Piotrowski A, Eifler D. Bewertung zyklischer Verformungsvorgnge metallischer Werkstoffe mit Hilfe mechanischer,
thermometrischer und elektrischer Messverfahren. Mat.-wiss. u. Werkstofftech., 1995, 26, 121.
[11] Starke P, Eifler D. Fatigue Assessment and FatigueLife Calculation of Metals on the Basis of Mechanical
Hysteresis,Temperature, and Resistance Data. Mat. Test, 2009, 261.