You are on page 1of 17


Very High Cycle Fatigue Failure Analysis and Life
Prediction of Cr-Ni-W Gear Steel Based on Crack
Initiation and Growth Behaviors
Hailong Deng 1 , Wei Li 1, *, Tatsuo Sakai 2 and Zhenduo Sun 1
Received: 10 October 2015; Accepted: 18 November 2015; Published: 2 December 2015
Academic Editor: Robert Lancaster


School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081, China; (H.D.); (Z.S.)
Research Center of Advanced Materials Technology, Ritsumeikan University, Kusatsu 5258577, Shiga, Japan;
Correspondence:; Tel./Fax: +86-10-6891-8730

Abstract: The unexpected failures of structural materials in very high cycle fatigue (VHCF) regime
have been a critical issue in modern engineering design. In this study, the VHCF property of a
Cr-Ni-W gear steel was experimentally investigated under axial loading with the stress ratio of
R = ´1, and a life prediction model associated with crack initiation and growth behaviors was
proposed. Results show that the Cr-Ni-W gear steel exhibits the constantly decreasing S-N property
without traditional fatigue limit, and the fatigue strength corresponding to 109 cycles is around
485 MPa. The inclusion-fine granular area (FGA)-fisheye induced failure becomes the main failure
mechanism in the VHCF regime, and the local stress around the inclusion play a key role. By using
the finite element analysis of representative volume element, the local stress tends to increase with the
increase of elastic modulus difference between inclusion and matrix. The predicted crack initiation
life occupies the majority of total fatigue life, while the predicted crack growth life is only accounts
for a tiny fraction. In view of the good agreement between the predicted and experimental results,
the proposed VHCF life prediction model involving crack initiation and growth can be acceptable for
inclusion-FGA-fisheye induced failure.
Keywords: Cr-Ni-W steel; very high cycle fatigue; interior failure; local stress distribution; initiation
and growth; life prediction

1. Introduction
In consideration of the current economics and energy resources, a great number of machinery
parts or components, such as axle, wheel, gear, blade, etc., are expected to have a very high cycle
fatigue (VHCF) life beyond 107 cycles [1,2]. However, many unexpected failures of structural materials
in the VHCF regime are reported one after another [3–5], and some peculiar failure mechanisms are
not yet well understood. From the viewpoint of safety and reliability, the analysis and assessment
methods based on high cycle fatigue can no longer ensure the safety of parts or components in the
VHCF regime. To elucidate the VHCF failure mechanisms of structural materials, and then to establish
the VHCF damage assessment and life prediction methods, has been more of a concern in the current
fatigue field.
The VHCF failure is often found in the experiments of alloy steels as one of the most widely
used structural materials [4]. In addition, this failure is often accompanied with a change of failure
mechanism from surface to subsurface or interior. The surface failure is mainly induce from surface
machining flaws such as scratches or cavities [3], where the interior failure is mainly induced from

Materials 2015, 8, 8338–8354; doi:10.3390/ma8125459

they to some extent reflect the failure behaviors of materials under a given condition. Experimental Procedure 2. Nakamura et al. the stress concentration factor of specimen.5 mm and 60 mm. and the relevant S-N property. Although the unanimous understanding or conclusion has not yet been drawn. The heat treatment processes were described as follows: (a) first quenching: 950 ˝ C ˆ 0. failure mechanism and crack size characteristics were analyzed. 2. and the relationship between the maximum principal stress and fatigue life was established.97 W. some methods or models combined with Paris law [14–16]. The tensile strength. studieson the VHCF life prediction associated with the crack initiation and the crack growth are relatively rare. the VHCF property of a gear steel was experimentally investigated under axial loading. and (3) Completed formation of the fine granular area (FGA). (b) secondary quenching: 850 ˝ C ˆ 0. σb . That is. Based on stress concentration handbook [21]. 0. the local stress around the inclusion during cyclic loading must be responsible for the formation of FGA. the researchers share a same belief or opinion.5 h + air cooling. Based on the finite element modeling. This region was called fine granular area (FGA) by coauthor Sakai [8]. 205 GPa and 0. 8. thought that the cyclic compression between the fine concavo-convex surfaces leads to the formation of FGA around the inclusion [13]. However. 8338–8354 the metallurgical defects such as non-metallic inclusions [6] or inhomogeneous microstructure [7]. Material and Specimen A Cr-Ni-W gear steel was investigated in this study. 0. From the standpoint of dislocation movement and irreversible slip. as shown in Figure 1. Based on the definition of crack initiation and grow processes.19] are proposed to predict fatigue life or strength in the VHCF regime. the fatigue life consumed in the crack initiation process should be considered. Furthermore. yield strength. respectively. Shiozawa et al. Besides that. proposed that the formation processes of FGA can be divided into three stages [9]: (1) Formation of a fine granular layer.Materials 2015. ? The area model associated with material hardness and defect size was developed by Murakami to evaluate the fatigue strength of material with defect-induced failure. Kt is evaluated to be around 1. pointed out that the dispersive decohesion of spheroidal carbides from matrix around the inclusion during the long-term cyclic loading is attributed to the formation of FGA [10]. the local stress distribution and the stress gradient effect around the inclusion was addressed. Based on this. and its main chemical composition (wt %) is 0.17. The minimum diameter and round notch radius of specimen are 4. proposed that the refinement of crystalline structure around the inclusion during cyclic loading leads to fracture surface of FGA [12].33 Mn. Sakai et al.22 Ni. a fatigue life model involving both crack initiation and growth was established to evaluate the VHCF life for inclusion-FGA-fisheye induced fatigue failure. 0. respectively. cumulative damage [19].01 V.55 Cr. etc. In the present study. 0. showed that the formation of FGA is caused by the synergistic effect between cyclic stress and hydrogen trapped around the inclusion [11]. From the annealed steel bar with a diameter of 16 mm. (2) Nucleation and coalescence of micro-debondings.5 h + air cooling. (c) tempering: 170 ˝ C ˆ 3 h + air cooling. FGA size [15. a special region can occur in the vicinity of the inclusion if the number of cycle is larger than around 106 cycles. 8339 .19 Si. Although these empirical methods or models are not necessarily appropriate for the experimental results of any steel. the hardness distribution along the cross section of specimen is almost uniform and the average value of Vickers hardness (HV) is about 503 kgf/mm2 . Murakami et al.1. are measured to be around 1520 MPa. 1. and the FGA governs the VHCF behavior of material.18].20] and tensile strength [7. Especially for VHCF occurring at relatively low stress levels.3. With respect to the interior inclusion-induced failure.02. elastic modulus and Poisson's ratio of Cr-Ni-W gear steel. and will be used throughout the paper. it is known that the fatigue process consists of crack initiation process and crack growth process. Grad et al. specimens were first machined into the shape of hourglass and then grinded in a direction parallel to the axis of specimen by the grades 600–2000 abrasive paper to final shapes. E and ν. σy . 4. 1190 MPa. stress ratio [16. Unfortunately.16 C.

 Microstructure and inclusion. In view of the linear distribution characteristics of  test data. as shown in Figure 2. page–page  Materials 2015. the  By means of the 4% alcohol nitric acid solution and scanning electronic microscope (SEM). the observed microstructure of heat‐treated material is the tempered martensite. respectively.  fracture  surfaces  of  all  the  failure mechanisms. open an  environment  at  room  temperature.1. Based on the reference [22]. the gear  steel  exhibits  the  used constantly  decreasing  S‐N  property  and gear the  steel traditional  fatigue  limit  Basquin model [23] was to establish the S-N3  curve of Cr-Ni-W under axial loading. Shape and dimensions of specimen (units: mm). 8.1. Shape and dimensions of specimen (units: mm). andat  theroom  testing frequency the  is around 100stress  Hz.  were carefully observed by SEM.are  Based on the preliminary SEM observation ofThis  crack initiation site. the Basquin model [23] was used to establish the S‐N curve of Cr‐Ni‐W gear steel under  becomes the unique failure mode of Cr‐Ni‐W gear steel in the VHCF regime.2. respectively. E i and υi are around 390 GPa and 0.  This  means  interior  failure  corresponding to 10 exhibits the constantly decreasing and traditional fatiguethat  limitthe  corresponding to 7 cycles test data. i andi are around 390 GPa and 0. S‐N Property  3. paying attention to the crack initiation sites and  temperature measuring instrument.  By means of the 4% alcohol nitric acid solution and scanning electronic microscope (SEM). Furthermore.  Figure 1. The fatigue test was performed in an  experiment. Based on the reference [22]. the relevant stress ratio was given ´1. Overall. Based on the preliminary SEM observation of crack initiation site. In view of the linear distribution characteristics of test data. Based on the preliminary SEM observation of crack initiation site. the Cr‐Ni‐W  shown in Figure 3.   Figure 1. fracture surfaces of all the failed specimens an  infrared  temperature  measuring  instrument. the tested specimen has no “self‐heat” phenomenon by means of the measurement with  failed specimens were carefully observed by SEM.  some   non‐metallic inclusions of Al 2O3 are found in the steel matrix in Figure 2.  Figure 2. the Basquin model [23] was used to establish the S‐N curve of Cr‐Ni‐W gear steel under  8340 3  . corresponding to 107 cycles cannot be observable.  Furthermore. Results and Discussion  3. This means that the interior failure becomes becomes the unique failure mode of Cr‐Ni‐W gear steel in the VHCF regime.  the elastic modulus and Poissonʹs ratio of inclusion. fatigue failures  gear  steel  exhibits  constantly gear decreasing  S‐N VHCF property  and  the  traditional  fatigue  gear limit steel the unique failure modethe  of Cr-Ni-W steel in the regime.  non-metallic inclusions of Al2 O23Oare found in the steel matrix in Figure 2. the Cr-Ni-W 7 cycles cannot be observable.2. Microstructure and inclusion. Overall. The fatigue test was performed in an  An axial‐type high frequency fatigue testing machine was used to carry out the fatigue test of  open gear environment  temperature. 8338–8354 Materials 2015.  3. 8. Overall. failed specimens were carefully observed by SEM. as shown in Figure 2. Ei and υ υi are around 390 GPa and 0.25.  means  that  the  6−109  cycles  is  The  S‐N  data  of  Cr‐Ni‐W gear  steel  under axial loading  in  the  life regime  of 10 specimens are all induced from the subsurface or interior. paying attention to the crack initiation sites and  3. Figure 2.  combined  with  the  analysis  of  energy  dispersive  X‐ray  spectrometer  (EDS). The fatigue test was Cr‐Ni‐W gear steel.  performed During  the in an Cr-Ni-W steel. E   Figure 2. page–page    Figure 1.  After  the  experiment.Materials 2015. the tested specimen has no “self‐heat” phenomenon by means of the measurement with  open environment at room temperature. paying attention to the crack initiation sites and failure mechanisms. of During  infrared  temperature  measuring  instrument. fatigue failures  3. Experimental Method  An axial‐type high frequency fatigue testing machine was used to carry out the fatigue test of  An axial-type high frequency fatigue testing machine was used to carry out the fatigue test of Cr‐Ni‐W gear steel.25. Based on the reference [22].  After  experiment. and the testing frequency is around 100 Hz. Microstructure and inclusion. S‐N Property  shown Figure 3. In view of the linear distribution characteristics of  of  specimens  are  all  induced  from S-N the property subsurface  or the interior.1. Results and Discussion 3. During the experiment. Results and Discussion  failure mechanisms. 8. After the experiment.  some   observed microstructure of heat‐treated material is the tempered martensite. Shape and dimensions of specimen (units: mm). respectively.    2.interior  fatiguefailure  failures of of in specimens  all  induced  from  the  subsurface  or  interior.25. the  observed microstructure of heat-treated material is the tempered martensite.  relevant  ratio  was  given  −1. the Cr‐Ni‐W  10 cannot be observable. some Furthermore. combined with the analysis of energy dispersive X-ray spectrometer (EDS).  By means of the 4% alcohol nitric acid solution and scanning electronic microscope (SEM).  surfaces  all infrared the  the  the tested specimen has no “self-heat” phenomenon bythe  means of the measurement with an experiment.  the  relevant  stress  ratio fracture  was  given  −1.2. non‐metallic inclusions of Al 3 are found in the steel matrix in Figure 2. and the testing frequency is around 100 Hz. as shown in Figure 2. Experimental Method 2. Experimental Method  2. S-N Property The  S‐N  data  of  Cr‐Ni‐W gear  steel  under axial loading  in  the  life regime  of 106−109  cycles  is  The S-N data of Cr-Ni-W gear steel under axial loading in the life regime of 106 ´109 cycles is shown in Figure 3.  combined  with  the  analysis  of  energy  dispersive  X‐ray  spectrometer  (EDS).  the elastic modulus and Poissonʹs ratio of inclusion.  the elastic modulus and Poisson's ratio of inclusion.

 the interior failure is all induced by the  Based on the further SEM observation and EDS analysis. the interior failure is all induced by the  Based on the further SEM observation and EDS analysis.      8341 4  4  .  the inclusion and the ambient matrix due to the inconsistency of deformation. causing the FGA formation.2. the Moreover. cycles Number of cycles to failure Nff .  Nf = 2. Figure 4. the value of σw is evaluated to be 485 MPa.93 ˆ 107 cycles). it should be noted that the crack morphology outside the  Moreover.aMPa .  S‐N curve. 3. Observation of inclusion. S‐N curve of Cr‐Ni‐W alloy steel with interior failure. page–page  Materials 2015. R= -1 Experimental Fitting curve data. it should be noted that the crack morphology outside the  causing the FGA formation. furthermore. Furthermore. as indicated by a solid circle in Figure 4. The inclusion nearly nearly  located  center  of  the  as fisheye. An approximately circular crack shaped like a fisheye  nonmetallic inclusion contained in material. it is undeniable that the local stress distribution around the inclusion play a key role in  inclusion affects the stress distribution on a micro-level and contributes to the interior crack initiation.  It  should  be  noted  that  the  FGA  with  rougher  surface  located at the center of the fisheye. In the  morphology occurs in the vicinity of the inclusion. as indicated by a dashed circle in Figure 4. as indicated by a solid circle in Figure 4. origin of the crack. N f = 2. the  Under cyclic loading. Fatigue Failure Mechanism Based on the further SEM observation and EDS analysis. the  the inclusion and the ambient matrix due to the inconsistency of deformation. Furthermore. and a solid line was plotted in Figure 3. An approximately circular crack shaped like a fisheye can can be observed on the fracture surface.  to the different crack growth rate. This is mainly attributed  causing the FGA formation. To a great extent. The inclusion is  be observed on at  thethe  fracture surface. the stress amplitude corresponding to 109 cycles is to  1099  cycles  is  approximately  defined  as  the  fatigue  limit  σw  of  material. Stress Stressamplitude amplitudea. 8. indicated by a dashed circle in Figure 4. it should be noted that the crack morphology outside the to the different crack growth rate.Materials 2015. 8338–8354 Materials 2015. the stress amplitude corresponding  and a solid line was plotted in Figure 3. fine granular area (FGA) and fisheye (σa = 600 MPa. According to the fitted S-N curve. Nf = 2. the severe stress concentration inevitably occurs on the interface between inclusion affects the stress distribution on a micro‐level and contributes to the interior crack initiation. It should be noted that the FGA with rougher surface morphology morphology occurs in the vicinity of the inclusion. some radial or “river” patterns can be seen and point to the  region outsider the FGA. the value of σ w is evaluated to be 485 MPa. To a great extent. the stress amplitude corresponding  axial loading.  Figure 3. MPa 750 750 Experimental data. fine granular area (FGA) aand fisheye (σ a = 600 MPa. some radial or “river” patterns can be seen and point to the  the FGA.2.  inclusion affects the stress distribution on a micro‐level and contributes to the interior crack initiation. Fatigue Failure Mechanism 3.      Figure 4. cycles     Figure 3.  According  to  the  fitted   approximately  defined  as σthe  fatigue  limit  σw  of  material. The inclusion is  can be observed on the fracture surface. S‐N curve of Cr‐Ni‐W alloy steel with interior failure. Observation of inclusion. page–page  axial loading.93 × 107 cycles).2. Moreover. 8.  origin of the crack. To a great extent. fine granular area (FGA) and fisheye (σ  = 600 MPa. Fatigue Failure Mechanism 3. In the  occurs in the vicinity of the inclusion.  of σw is evaluated to be 485 MPa.  Figure 3. S-N curve of Cr-Ni-W alloy steel with interior failure. An approximately circular crack shaped like a fisheye  nonmetallic inclusion contained in material. Moreover. the severe stress concentration inevitably occurs on the interface between  the inclusion and the ambient matrix due to the inconsistency of deformation. the severe stress concentration inevitably occurs on the interface between  Under cyclic loading. furthermore. 8.  According  to  the  to  10   cycles  is defined approximately as the fatigue limit material.93 × 107 cycles). Furthermore. Observation of inclusion. Moreover. the interior failure is all induced by the nonmetallic inclusion contained in material.  It  should  be  noted  that  the  FGA  with  rougher issurface  nearly  located  at  the  center  of  the  fisheye. In the region outsider region outsider the FGA. as indicated by a solid circle in Figure 4.  Figure 4. Under cyclic loading. it is undeniable that the local stress distribution around the inclusion play a key role in fisheye is obviously different from that within the fisheye outside the FGA. it is undeniable that the local stress distribution around the inclusion play a key role in  Moreover. and a solid line was plotted in Figure 3. furthermore. R= -1 Fitting curve 700 700 650 650 600 600 550 550  = 485 MPa ww= 485 MPa 500 500 450 4501066 10 7 107 10 8 108 10 9 109 10 10 1010 10 Number of cycles to failure N . thefitted   value w of S‐N curve. some radial or “river” patterns can be seen and point to the origin of the crack. This is mainly attributed  fisheye is obviously different from that within the fisheye outside the FGA. as indicated by a dashed circle in Figure 4.

Sizes of inclusion. the parameters Rinc .4.05 11.13 6. The effect of microstructure such as grain size is not considered in the FE modeling.Materials 2015. Based on the SEM observation. FGA and fisheye.. the increasing of fatigue life.86 229. the inclusion and the ambient matrix are defined as the different linear elastic materials. in the direction of the Y axis. It can be seen that the difference for the values of Rinc is so slight. 8342 .45 30. the RVE is subjected to a tensile stress σ8 which is equivalent to the applied normal stress during the experiment.12 7.4.69 421. the inclusion causing fatigue failure can be viewed as the maximum inclusion on the minimum cross section of specimen under the axial loading. the representative volume element (RVE) proposed by Hill [24] is used to evaluate the local stress distribution characteristics around the inclusion. the measured values of Rinc .14 514.62 47. The average value of Rinc is evaluated to be 8. i.83 366. RFGA and RFisheye are presented in Table 1.49 36. and there is no special contact between them.89 8. as shown in Figure 5a.68 27. σa (MPa) Rinc (µm) RFGA (µm) RFisheye (µm) 700 675 650 625 625 600 600 575 550 525 10. 8.48 334. During the finite element (FE) modeling of RVE.86 10.7 µm.e. the inclusion is defined as the area within the red line and the rest area is the matrix. the local stress distribution characteristics around the inclusion have a great effect on the VHCF failure of material. The RVE can be considered to be the smallest volume over which a measurement can be made that will yield a value representative of the whole. By using the graphical software.7 µm which is the average value of Rinc . In this study. a cube that contains the maximum inclusion and the adjacent matrix is defined as a RVE. Local Stress Distribution around Inclusion 3. Crack Size Characteristics In view of the approximately circular shape of inclusion.85 8. RVE Model As mentioned earlier. as shown in Figure 5b. Table 1. FGA and fisheye.3. the sizes of FGA and fisheye are larger.12 7. considering material mechanical property.04 132. 3. FGA and fisheye. and can be viewed as a sphere. With the decreasing of applied stress amplitude.32 40. Furthermore. In this figure.76 3. The side length of RVE is 100 µm.42 33. It can be concluded that the sizes of inclusion are regardless of the applied stress amplitude or fatigue life and dependent on the melting techniques of steel.75 260.14 23. The RVE which contains the inclusion can be approximately considered as a heterogeneous material.23 65. That means that the lower stress level is. Moreover.89 31. the values of FGA and fisheye all tend to increase.61 32.56 8. and the radius of inclusion is 8.16 8.57 271. the type of 4-node-shell elements (CPS4R) as a two-dimension area model was used to simplify the calculation.34 24. To some extent. respectively.1.43 299. This is mainly attributed to the different crack growth rate. RFGA and RFisheye are adapted to indicate the radiuses of inclusion. 8338–8354 fisheye is obviously different from that within the fisheye outside the FGA. the local fatigue property around the inclusion governs the overall fatigue property of material in the VHCF regime. Under this circumstance.

4.2. 3.  the  maximum  difference between inclusion and matrix. Herein the inclusion property is mainly associated with the change of elastic module. page–page    (a)    (b)  Figure 5. In In  other  words. This means that the harder the inclusion is.  means that the harder the inclusion is.  stress distribution diagram is shown in Figure 6b. the maximum principle stress around the inclusion is larger.Materials 2015. and (b) Model and boundary conditions. Herein the inclusion property is mainly associated  maximum principal stress around the inclusion. It is known that the elastic module reflects the resistance capacity to elastic deformation.4. 8. the relationship between σmax /σ8 and Ei/E is shown in Figure 6a. the maximum principle stress around the inclusion is larger. 8343 6  . the effect of different inclusion properties on the local stress  The value of σ8 is set as 525 MPa. In fact. 8. This  Ei /E.  the  local  stress  around  the  inclusion  tends  to  increase  with  the  increase  of  elastic  other words.  The  parameters  σmax  is  used  to  denote  the  distribution was investigate under monotonic tension. and (b) Model and  Figure 5. the maximum principal principal stress distribution diagram is shown in Figure 6b. Finite element modeling based on RVE method: (a) Definition of RVE. the effect of different inclusion properties on the local stress distribution  was  investigate  under  monotonic  tension. the value of σmax /σ8 corresponds to the stress centration factor of inclusion. Effect of Inclusion Properties The value of σ∞ is set as 525 MPa. Obviously. the value of σmax/σ∞ corresponds to the stress centration factor of  Ei /E is shown in Figure 6a. The parameters σmax is used to denote the maximum principal stress around the inclusion. the relationship between σmax/σ∞ and  with the change of elastic module. it can be found that the values of σmax/σ∞ tend to increase with the increase of  inclusion. Based on the FE calculation.  boundary conditions. Obviously. the local stress around the inclusion tends to increase with the increase of elastic modulus modulus  difference  between  inclusion  and  matrix. Corresponding to Ei = 420 GPa.  Corresponding  to  Ei  =  420  GPa. Based on the FE calculation. Effect of Inclusion Properties  3. Finite element modeling based on RVE method: (a) Definition of RVE. it can be found that the values of σmax /σ8 tend to increase with the increase of Ei/E. It is known that the elastic module reflects the resistance capacity to elastic deformation. 8338–8354 Materials 2015. In fact.2.

 first. The obtained maximum  3 cycles are shown in Figure 7.1 1.  in Figure 7a.4.9 0.5 Simulation data 1. Then.0 3.0 1.5 3. the number of loading cycles has no significant effect on the local stress distribution around the inclusion. it can be concluded that cyclic loading has a little effect on the local stress distribution around the inclusion. and (b) Stress distribution characteristics for Ei = 420 GPa. the number of loading cycles has no significant effect on the local stress distribution  the local stress distribution around the inclusion.5 1. based on the comparison between results around the inclusion. first. Effect of Cyclic Loading Under the condition of σ∞ = 525 MPa and Ei = 420 GPa. 8.2 1. It should be noted that considering likelihood  of  cyclic  hardening  of  inclusion material  under  cyclic  loading.4. Conversely.  only the linear-elastic constitutive relation was defined under static loading. page–page  1.3. Then. it can be concluded that cyclic loading has a little effect  Compared with the result in Figure 6b.0 0.  between σmax /σ8 and Ei /E. Compared with the result in Figure 6b.0 Ei / E (a)  (b)  2. Effect of Cyclic Loading 3. The obtained maximum principal stress distribution diagrams under cyclic loading of 10 and 10 principal stress distribution diagrams under cyclic loading of 10 and 103 cycles are shown in Figure 7. 3.b. only the linear‐elastic constitutive relation was defined under static loading.5     Figure 6. the effect of cyclic loading on the local  Under the condition ofthe  σ8inclusion  = 525 MPa Ei = 420 It  GPa.  likelihood of cyclic hardening of material under cyclic loading. Stress distributions around different inclusions under monotonic tension: (a) Relationship between σmax/σ∞ and Ei/E. 8338–8354 Materials 2015. a combined isotropic-kinematic hardening constitutive relation was accounted for the FE-modeling under cyclic loading. the be  effect of that  cyclic loading on the stress  distribution  around  was and investigated. and (b) Stress distribution characteristics for Ei = 420 GPa. 8.Materials 2015.b.  should  noted  considering  the  local stress distribution around the was investigated. 7  8344 .  a  combined  isotropic‐kinematic  thehardening constitutive relation was accounted for the FE‐modeling under cyclic loading. Conversely.3. based on the comparison between results  on in Figure 7a.5 2.3 1.4 max 1. Stress distributions around different inclusions under monotonic tension: (a) Relationship  Figure 6.

 It can be found that with the increase of distance from the  max 8 theedge of inclusion. the maximum principal stress has a tendency to increase with the decrease of fatigue life. 7.14 μm and the stress amplitude of 525 MPa. ∞ begins to approach 1. Maximum Principal Stress Life Curve 3. On the other hand. 8  8345 . On the other hand. the value of and stress gradient effect around the inclusion are greatly related to the formation of FGA. Similarly. the maximum principal stress has a tendency  RVE model. Approximately.  10 cycles. to the inclusion size of 8.  N = 10 3 N = 10 cycles. About at the distance of 20 μm. page–page      (a)    ( (b)  Figure  Stress  distribution around  inclusion  under under  cyclic  and  (b) (b) at  at Figure 7. 8338–8354 Materials 2015. as shown in Figure 8.4. Approximately. Similarly.14 µm and the stress amplitude of 525 MPa. it can be concluded that the higher stress concentration and stress gradient effect around the inclusion are greatly related to the formation of FGA. the ratio of σ /σ as a function of the distance from the edge of inclusion is shown in Figure 9. as shown in Figure 8. it can be concluded that the higher stress concentration  edge the value of σmax /σ8 tends to decrease. The relationship between the obtained maximum principal stresses and the fatigue lifeto increase with the decrease of fatigue life. and 3 cycles. corresponding to the inclusion size of  can be established.4.Materials 2015.  Stress distribution around inclusion cyclic loading:  loading:(a)  (a)at atN N= =10 cycles. Maximum Principal Stress Life Curve Based  on  the  applied  stress  amplitudes  and  the  sizes  of  inclusions  in  Table  1.4. 8.  the  relevant  Based on the applied stress amplitudes and the sizes of inclusions in Table 1. the ratio of σ maxcorresponding /σ∞ as a function of the distance from  8. 3.  σmax /σ8 begins to approach 1.4. the value of σ edge of inclusion is shown in max Figure 9. It can be found that with the increase of distance from the /σ∞ tends to decrease. About at the distance of 20 µm. the value of  σmaxof/σinclusion. the relevant maximum principle stresses around different inclusions at different stresses can be obtained by using  maximum principle stresses around different inclusions at different stresses can be obtained by using the RVE model. The relationship between the obtained maximum principal stresses and the fatigue  thelife can be established. 8.

. Crack Initiation Life 3.  σb σ=  3. Ni [25]. Local Stress‐Life Model  3. Evaluation of maximum principal stress at the interface between inclusion and matrix. Local Stress‐Life Model  The local stress‐life curve can be used to predict the crack initiation life.  somewhat  larger the than  the  experimental  of  1520  MPa.between  somewhat larger than the experimental value of 1520 The difference between them  is  slight.1.5.1 1. 8.2 1. definition  of  fatigue  limit  on σbdifference  = 3.  The  σh = 1609  700 MPa. For the VHCF failure. For the tested gear steel in this study. Local Stress-Life Model and expressed as [26] 4 4 log10 b  2.5.0 1.Materials 2015.85  700  log ˆ10  ˙ 700  104b  2. σ h = 700 MPa. a crack initiation life model limited in the life regime of 10 –107 cycles was proposed and expressed as [26]   3. i. the evaluated value of σb is about  1609  MPa  based  on  σb  =  3. where σb also approximately indicates the fatigue strength  diagram is shown in Figure 10.  Figure 9.e. 8338–8354 Materials 2015. where σ also approximately indicates fatigue at fatigue 2 b at  fatigue  life  of  102  cycles  and  the  parameter  σh  denotes  the  fatigue the strength  at strength fatigue  life  of   at  fatigue  life  of  10   cycles  and  the  parameter  σh  denotes  the  fatigue  strength  at  fatigue 7 life  of   2 7 life of1010 cycles and the parameter σh denotes the fatigue strength at fatigue life of 10 cycles.2  HV. a crack initiation life model limited in the life regime of 10 The local stress-life curve can be used to predict the crack initiation life. the fatigue strength  however.  and  700  MPa  is  (1) where  the  tensile  strength  b  is  related to to the the local local  hardness.  the MPa. Establishment of maximum principal stress gradient curve.2 1. the evaluated value of σ b is about  7 10  cycles.2  HV.  Figure 9. Crack Initiation Life  3.  The  relevant  7 assumed  to  be  the  fatigue  limit  value  corresponding  to  fatigue  life  of  10   cycles.. Based on the Basquin equation  strength σw corresponding to the w corresponding to the fatigue life of 10 σa = A(N)BB.5.. where σ schematic diagram is shown in Figure 10. Based on Basquin equation [23]..1 1.1.  the  definition  of  fatigue  limit  7 cycles is apparently not suitable. 8.2 HV. 8.  The  relevant  assumed to be the fatigue limit value correspondingb also approximately indicates the fatigue strength  to fatigue life of 10 cycles.e. Figure 8.  however. Based on Basquin  2–107 cycles was proposed  equation [23]. evaluated value of value  σb is about 1609 MPa based MPa  based  σb  =  3.  m m Distances 45 45 50 50    Figure 9.85   a 10 Ni “ 10i 7 σa a  7 7 (1) (1) where  the  tensile  strength  σb  is  related  to  the  local  hardness.  about  σb  =  3.5.2 700 MPa assumed  to  be  the  fatigue  limit  value  corresponding  to  fatigue  life  of b 1077  cycles. the definition of fatigue limit corresponding to fatigue 9 cycles is used to replace σh.  Figure 8.0 0 0 5 5 10 10 15 15 20 20 25 25 30 30 35 35 40 40 Distances from from inclusion inclusion .5.e.. Therefore.1. the value of 9  σw can be given by 9  ´ ¯B σw “ A 109 (2) 8346 .about  about = 3. Evaluation of maximum principal stress at the interface between inclusion and matrix.7 cycles is apparently not suitable. HV. R= -1 Experimental data.  For  the  VHCF  failure.  cycles.  however. Based on the Basquin equation σa = A(N)B . page–page  900 900 Experimental data. Establishment of maximum principal stress gradient curve. σh = 700 MPa. Based on the Basquin equation  σw10  corresponding to the fatigue life of 10 7 cycles is apparently not suitable. Therefore. R= -1 max .MPa MPa max 850 850 800 800 750 750 700 700 650 650 600 6001066 10 7 8 107 10 9 108 10 Number cycles cycles to to failure failure N Nf . For the tested gear steel in this study. N 2–10 7 cycles was proposed  i [25].5.  Simulation data Simulation data max // . i. page–page  Materials 2015. and and 700  MPa  is  is where the tensile strength σbσis related hardness.2  HV. cycles cycles Number f 109 10    Figure 8. For theon  tested gear steel in this study. Based on Basquin  The local stress‐life curve can be used to predict the crack initiation life. Establishment of maximum principal stress gradient curve. Evaluation of maximum principal stress at the interface between inclusion and matrix.85   N  i N  10   700 log σb ´ 2. the fatigue strength  corresponding to fatigue life of 10 themcorresponding to fatigue life of 10 is slight. a crack initiation life model limited in the life regime of 10 2 and expressed as [26]   equation [23].2  HV. i. the fatigueh. Crack Initiation Life  3. the value of σ w can be given by  9 cycles is used σa = A(N) . Ni [25]. the value of σ w can be given by  fatigue life of 10 to replace σh . 9 cycles is used to replace σ life ofσ Therefore. 3. .  The  difference  between  them  is  slight. The relevant schematic schematic diagram is shown in Figure 10..MPa MPa max  1.  somewhat  larger  than  the  experimental  value  of  1520  MPa.  For  the  VHCF  failure.

  μinc  is  the  shear  modulus  of  inclusion. respectively. Tanaka‐Mura Model  3. µinc is the shear modulus of inclusion. τoct and the [20]:  octyield  2strength  y / 3   of material. the values of ∆τ and τf in Equation (6) can be expressed as: ∆τ “ ? 2∆σ{3 and τ f “ 8347 ? 2σw {3 (9) . The relationship between E and shear modulus µ is given by:   E /  2  2    µ “ E{ p2 ` 2υq (7) Furthermore. the relationship between the critical octahedral shear stress. page–page  Materials 2015. Ws is the specific fracture energy for a unit area along Generally.5. Tanaka-Mura Model Based on dislocation‐energy method. 8.  The  relationship  between  E  and  shear  the slip band. Combined with the evaluated local maximum stress. W s is the specific fracture energy for a  where µm is the shear modulus of matrix. Combined with Equation (2). the relationship between  the critical octahedral shear stress. a crack initiation life model associated with the inclusion de‐bonding or cracking was proposed by Tanaka‐Mura and expressed as [27]  de-bonding or cracking was proposed by Tanaka-Mura and expressed as [27] 2      inc    h  4Ws 1 N„i   m m   ˆ ˙2   2 µm pµm ` µinc q   hh l  (4W inc  2sf ) R1inc Ni “ 2 µinc h`l (6) (6) p∆τ ´ 2τ q Rinc f where  μm  is  the  shear  modulus  of  matrix. the values of A and B can be respectively expressed as the values of A and B can be respectively expressed as   σ w AA “  9Bw9B   10 10 (3) (3) log σw b 1010 log w´log log1010σ b B“ B   77 (4) (4)   Figure 10.  and semi-major of the slip band zone respectively. Finally. Generally. based on the maximum octahedral shear stress criterion. 8338–8354  w  A 109    B (2) wherewhere parameters A and B are coefficient and exponent. σy can be defined as(8) ? τoct “ 10 2σy {3 (8) Accordingly.  Figure 10. Combined with the evaluated local maximum stress.  Δτ  is  the  local  shear  stress  range  and  τf  is  the  critical  shear  stress.5.Materials 2015. ∆τ is the local shear stress range and τf is the critical shear stress. h and l are semi-minor unit  area  along  the  slip  band. the crack initiation life model based  9 cycles can be established as  Finally. based on the maximum octahedral shear stress criterion. Basquin S‐N relationship in lg‐lg scale.2. respectively. the value of l modulus μ is given by:  is defined as the half of grain size.  h  and  l  are   semi‐minor and semi‐major of the slip band zone respectively.2.  the  value  of  l  is  defined  as  the  half  of  grain  size. a crack initiation life model associated with the inclusion  Based on dislocation-energy method. Basquin S-N relationship in lg-lg scale. σy can be defined as [20]:   (7) Furthermore. 8. Combined with Equation (2). the crack initiation life model based on the local stress‐life at 10 9 on the local stress-life at 10 cycles can be established as 7    log10 b log w  109  ˙ w  7 Ni ˆ    σ max   w log10 σb ´ logσw 9 Ni “ 10 (5) (5) σmax 3.  parameters A and B are coefficient and exponent. τoct and the yield strength of material.

the effective number of dislocations causing crack initiation. the value of c is given by c “ nc b (16) Combined with Equation (14). studies show that not all dislocations contribute to the crack initiation. nc . the crack initiation model with the FGA formation is obtained as: „ µm pµm ` µinc q Ni “ µinc ˆ h h`l ˙2 1 2pσmax ´ σw q Rinc 9Ws 2 (10) When h is much less than l. In a slip band zone. leads to „ Ni “  h4 µm pµm ` µinc q 9µm c 2 2 R 2 0. Combined with Equations (6) and (9). and herein defined as the fatigue strength corresponding to fatigue life of 109 cycles. the Equation (6) can become „ µm pµm ` µinc q Ni “ µinc  ˆ ˙2 h 1 9Ws 2 R l 2pσmax ´ σw q inc (11) If the slip band zone is circular. yield 1{2 Ni „ µm pµm ` µinc q “ 0.005 2l h ˙2 γs µm (17) Combined with Equation (10) and imposing the condition that γs = W s [29]. yield ˆ c “ 0. 8.e. From incipient crack initiation.005µinc 1{2 ? ˆ ˙1{2 h2 3 2µ1/2 c m l ph ` lq 4pσmax ´ σw q Rinc 8348 (19) . h is the same as l. However. 8338–8354 where σw is the critical stress below which crack does not initiate. W eq is the strain energy stored in the dislocations of an equivalent slipband and c is the crack length.005µinc l ph ` lq 8pσmax ´ σw q inc (18) Taking the square root of Equation (18). W eq can be expressed as Weq “ 4lγS (15) where γs is the surface energy of the crack. moreover. the Equation (10) can be simplified as  9Ws µm pµm ` µinc q 1 Ni “ 2 R µinc 8pσmax ´ σw q inc „ (12) For the fracture energy W s . it can be given by [28] 2 Ws “ ∆Kth { p2Eq (13) 3.05 (14) bh 2lµm where b is the magnitude of the Burgers vector. Modified Chan Model The Tanaka-Mura model is based on the assumption that all the dislocations in the slip band zone contribute to the crack initiation.Materials 2015. can be given by [30] ˆ ˙d cWeq 2l nc “ 0. and the number of dislocations that contribute to the crack initiation is greatly related to the crack size [29]. i.3.5. In this case..

the crack initiation life curves corresponding to different 12  c-values can be established.005µ ` lFGA q 4pσ3max2 ´mσw q Ni    0.992. all predicted results  agreement between the predicted and experimental data. Equation (19) can be rewritten as 1/2 1/2 1/2 agreement i For the better agreement with the experimental data.  the  fitting  correlation  coefficents  associated  local stress-life model. The established crack initiation life  corresponds to the crack length of FGA size. all predicted results reveal that the crack initiation life accounts for the vast majority of the total fatigue life.Materials 2015. Comparison of three crack initiation life models.005 variable α (0 < α < 1) [31]. reveal that the crack initiation life accounts for the vast majority of the total fatigue life.  lFGA  is  assumed  to  be  stress-life data and the measured sizes of inclusion and FGA. The established crack initiation life FGA − Rinc. c = R curves all exhibit the continuously descending trend. the crack initiation life curves corresponding to these  three models are plotted in Figure 11.. Based on a modified Chan model. the predicted result based on modified Chan model shows  betterthe better agreement between the predicted and experimental data. Obviously. 8.4. which is consistent with the variation trend of  experimental data.  Based  on  the  established  maximum   determined.  Figure 11. as shown in Figure 12. the crack initiation life curves corresponding to these According to Equations (5). Moreover. R= -1 Local stress life model Tanaka-Mura model Modified Chan model 850 max . Moreover. (12) and (22).977. it is assumed that the elliptic slip band zone is expanded Thus. Comparison of Model  According to Equations (5).958 with local stress‐life model.e.  and 0. MPa 800 750 700 650 600 6 10 7 8 10 10 Number of cycles to failure Nf . the predicted result based on modified Chan model shows the 0.958 and 0. 8.  From  the  statistical  consideration. 0. page–page  Materials 2015. the exponent of N ?  ˆ i can be generalized to a  ˙1{2 „ h2 3 2µ1/2 c µm pµm ` µinc q 1{2 variable α (0 < α < 1) [31]. the values of α and hFGA can be equivalent  to  RFGA  and  hFGA  is  a  fitting  parameter  just  like  α. The crack initiation life corresponding to different c-values can be obtained as stress‐life  data  and  the  measured  sizes  of  inclusion  and  FGA. (12) and (22).   i In this study. respectively.4.e. In view of 8349 . From the statistical consideration. where the curve established by using modified Chan model three models are plotted in Figure 11. Herein. Therefore.005 semi-minor l inc  lFGA  hFGA  lFGA  4(max  w )  RincFGA  is assumed to be 1/2 lFGA 0. Obviously. the fitting correlation coefficents associated with experimental  data. the exponent  of N can  the experimental  inc  l  h  l  4(max  w )  Rinc i  0. Based on the established maximum where  hFGA  and  lFGA  are  the  semi‐minor  and  semi‐major  of  FGA.005µinc phFGA2 ` lFGA q 4pσmax ´1/2σw q R 1/2 where hFGA equivalent to RFGA and hFGA is a fitting parameter just like α. Comparison of three crack initiation life models.5. cycles 9 10   Figure 11. the need crack initiation life to induce a certain crack size is longer at low stress than at high stress. where the curve established by using modified Chan model  corresponds to the crack length of FGA size. Comparison of Model      lFGA  hFGA  lFGA  4(max   w )     Rinc    (22) (22) 3. Tanaka-Mura model and Modified Chan model are evaluated to be 0. Equation (19) can be rewritten as  m α Ni “ 0.  the  values  of  α  and  hFGA  can  be  determined. Therefore.  ?  ˆ ˙1{2 „ Equation (20) can be rewritten   h2 3 2µ1/2 c µ pµ ` µ q 1{2 m m Niα “ m inc FGA (21)      inc    hFGA 3 2 m  c inc (21) Ni   m m      and semi-major of FGA.  Herein. which is consistent with the variation trend of curves all exhibit the continuously descending trend. i. Tanaka‐Mura model and Modified Chan model are evaluated to be 0. The crack initiation life corresponding to different c‐values can be obtained as  ˜ ¸1{α ˆ ?  ˙ „ Ni “ h2FGA µm pµm ` µinc q 1{2α 1/2  3 2µ1/2 c1/2  1{2α m 1/  2 1/2    m  inc   lFGA phFGA   m inc  c Rinc  hFGA 0. and lFGA are the0. 8338–8354   m   m  inc    c  h2 3 21/2 m (19) to a N with   be generalized For the better data.5. Thus.005inc  l  h  l  4(max  w )  Rinc  Equation (20) can be rewritten In this study. i.977. respectively. It can be seen from the figure that the crack initiation life tends to increase with the increase of crack size at a certain stress.  0.. Furthermore. c = RFGA ´ Rinc .005µinc 1/2l ph ` lq 4pσmax ´ σw q Rinc 1/2 (20)   m   m  inc    c  h2 3 21/2  m (20)  N  to  the whole FGA.  900 Experimental data.005inc 3. it is assumed that the elliptic slip band zone is expanded to the whole FGA.992.

 its effective stress intensity factor range. For a circular crack with a radius  where ∆K stress intensity factor range at a crack tip.  It  can  be  seen  from  the  figure  that  the  crack  initiation life tends to increase with the increase of crack size at a certain stress. That is. hence. it reveals that the VHCF life is mainly consumed in the crack initiation  experimental data. Crack Growth Life  3. Furthermore. and (ii) long crack growth from FGA to fisheye. Crack Growth Life Combined with the previous studies [9]. 8.Materials 2015. Predicted crack initiation lives by using modified Chan model at different c-values. and 3 expressed as [32] da ˆ K eff˙ (23)  b 3   da dN ∆K “ b  E?effb  (23) dN E b where ΔKeff is the effective stress intensity factor range at a crack tip. as schematically shown in Figure 13.6. A to FGA. R= -1 Crack initiation life curves 850 max . the crack growth life at stage (i).  the  crack  initiation  life  curves  corresponding  to  different   c‐values  can  be  established. the crack growth for the interior inclusion-FGA-fisheye failure in the VHCF regime can be divided into two stages: (i) small crack growth from inclusion to  FGA.growth  and (ii) long crackwas  growth from FGAthe  tocrack  fisheye. MPa 800 c = RFGA-Rinc c = 20 m 750 700 c =15 m 650 600 6 10 c =10m c =5m 7 8 10 9 10 Number of cycles to failure Nf . N1 . the crack growth for the interior inclusion‐FGA‐fisheye  Combined with the previous studies [9]. 8338–8354 900 Experimental data.6.  as  shown  in  Figure  12. Predicted crack initiation lives by using modified Chan model at different c‐values. ΔK π 2 Kinc intensity max factor Rinc   range. stage inside the FGA. 8. ∆K can be expressed (25) as For the inclusion. hence.   ˆ ˙3 ˆ ˙3{2 ˆ where  K inc / E b  da “b dN 2∆σmax πa ? πE b “b ∆Kinc ? E b a Rinc “b a Rinc ˙3{2 (26) ? where ∆Kinc {E b “ 1 is the threshold corner location with da/dN = b. the value of ΔK eff can be defined as [33]  eff is the effective radius a.  12. That is. the need  crack initiation life to induce a certain crack size is longer at low stress than at high stress. the crack growth rate for stage (i) can be given by  a 2 ∆Kinc “3 ∆σmax πR (25) 3/2 3/2 3 inc  2 max a  π  K inc   a   a  da (26)  b    b   R   b R    E band(25). as schematically shown in Figure expressed as [32]  crack growth rate equation was used to describe the crack growth behaviors in these two stages. the E crack b  growth inc  inc stage  In combination withdN Equations rate for (i) can be given by  (23) ? ˙3 ˆ1  is the threshold corner location with da/dN = b. it reveals that the VHCF life is mainly consumed in the crack initiation stage inside the FGA.  Materials 2015. In view of  the fact that the predicted crack nucleation lives associated with the FGA sizes are in good agreement the fact that the predicted crack nucleation lives associated with the FGA sizes are in good agreement  with thewith the experimental data. A crack  failure in the VHCF regime can be divided into two stages: (i) small crack growth from inclusion rate  equation  used  to  describe  growth  behaviors  in  these  two  stages. cycles 10   Figure Figure 12.  and  13. 3. is obtained as N1 “ « πE2 2p∆σmax q2 8350 1´ d Rinc RFGA ff (28) . the value of ∆Keff can be defined as [33] 2 K eff   a   (24)  ? 2 ∆Keff “ ∆σ πa (24) inc can be expressed as  For the inclusion. its effective stress inc  In combination with Equations (23) and (25). page–page  Based  on  a  modified  Chan  model. 13  ? ∆Kinc 2∆σmax πRinc ? 1“ ? “ E b πE b (27) Integrating Equation (26) from Rinc to RFGA . For a circular crack with a a.

 the crack growth  is shown in Figure 14.40 ˆ 103 1. Schematic of small and long crack growth behaviors. the crack growth life  at stage (ii). the growth rate in  stage (ii) is reduced by the fact “1/x3” in comparison with the growth rate in stage (i). the ratio of the predicted crack growth life to the experimental life.With  With the decrease of stress amplitude.36 ˆ 104 1. the  that the crack growth life is mainly consumed in the long crack growth stage. 8338–8354 In view of the difference between the small and long crack growth behaviors. In contrast.53 ˆ 105 3. the proportion of the crack  σmax .31 ˆ 105 .44 ˆ 104 1. the proportion of the crack growth life in the total fatigue life tends to increase. This means that the crack growth life is mainly consumed in the long crack growth stage. This means  from inclusion to FGA is shorter than that for the long crack growth from FGA to fisheye.32 ˆ 104 1. At a relatively low stress level.36 ˆ 1. the crack growth life at stage (i). it reveals that the VHCF life is mostly consumed in the crack  initiation stage. growth life in the total fatigue life tends to increase.Integrating Equation (26) from Rinc to RFGA.90 ˆ 104 1. the growth rate in stage (ii) is reduced by the fact “1/x3 ” in comparison with3 the growth 3/2 rate in stage (i). In contrast. is obtained as  N1  E 2 2(max )2  Rinc 1  RFGA      (28) Materials 2015.97 ˆ 104 2.40 ˆ 105 3. the life for the small crack growth from inclusion to FGA is shorter than that for the long crack growth from FGA to fisheye. N1.51 ˆ 105 3. Predicted results of crack propagation life.15 ˆ 1. Schematic of small and long crack growth behaviors. Thus. where the total crack growth lives 5 only  scattered  in  the range range of of 11 ˆ ×  10 1055–4  the  decrease  of  stress  amplitude. 8. Conversely.  life only accounts for a very tiny fraction of the total fatigue life.40 ˆ 105 2. 14  Table 2. On the other hand.65 ˆ 105 2. Np can be expressed as d Rinc ff E 2 «  d Rinc πE2 Rinc Rinc N  N  N  1  26  27    p 1 2 Np “ N1 ` N2 “ 26 2(max ) 21 ` RFGA´ 27 RFisheye RFisheye   RFGA 2p∆σ q2 (31) (31) max   Figure 13. the total crack growth life. N /N p f considered as a function of is shown in Figure 14.45 ˆ 105 2. where the total crack growth lives are  The predicted crack growth lives are presented in Table 2. N2 . Np can be expressed as  Combined with Equations (28) and (30).  ratio of the predicted crack growth life to the experimental life. is obtained as Rinc Rinc 27E ff   «d  d  2 2 27πE RFisheye 2( max ) Rinc RFGA R inc  N2  N2 “ (30) (30) ´ RFGA RFisheye 2p∆σmax q2 Combined with Equations (28) and (30).  cycles. Conversely.41 ˆ 104 1. N p/Nf considered as a function of σmax. The predicted crack growth lives are presented in Table 2.  the  the are only scattered in the –4 ×  ˆ10 10 5cycles. It can be found that with the increase of stress.31 ˆ 105 2. On the other hand.41 ˆ 105 2. is obtained as  dN 27 Rinc x FGA to RπE (29) Integrating Equation (29) from RFGA to RFisheye 2 .27 ˆ 105 2. At a relatively low stress level.08 ˆ 105 N p (cycles) 3. It can be found that with the increase of stress. the life for the small crack growth  predicted crack growth life has an increase tendency. the total crack growth life. the crack da b  2 max a  b  a  (29) growth rate for stage (ii) is given as  3        dN ˆ x  E? b ˙  27  ˆ  Rin˙c 3{2 3 da b 2∆σmax πa b a ? “ 3 “ Integrating Equation (29) from R Fisheyeb .11 ˆ 105 3. σa (MPa) 700 675 650 625 625 600 600 575 550 525 σmax (MPa) 842 837 781 783 792 761 725 714 672 645 N 1 (cycles) N 2 (cycles) 103 105 7.01 ˆ 104 9.31 ˆ 105 3.10 ˆ 105 2. N 2.54 ˆ 105 2.  Figure 13. predicted crack growth life has an increase tendency. it reveals that the VHCF  life is mostly consumed in the crack initiation stage.19 ˆ 105 2. the crack  growth rate for stage (ii) is given as  In view of the difference between the small and long crack growth behaviors. the crack growth life only accounts for a very tiny fraction of the total fatigue life.34 ˆ 105 3.23 ˆ 105 1.33 ˆ 104 8351 3. Thus. the crack growth life at stage (ii).80 ˆ 105 2.

and reflects the inclusion-FGA-fisheye induced failure be used well. Furthermore. the predicted result is satisfactory and the above-established model involves the crack initiation and growth processes.  the  VHCF  life  prediction  model  associated  with  the   Based on Equations (22) and (31).  used well.53 × 105  3. this model is still needed to verify by more  mechanism.97 × 104  2.31 × 105  3. 8338–8354 525  792  761  725  714  672  645  1.33 × 104  2.41 × 104  1. cycles   Figure 15. the VHCF life prediction model associated with the inclusion‐FGA‐fisheye failure can be obtained as  inclusion-FGA-fisheye failure can be obtained as N pre  Ni  N p Npre “ Ni ` Np 1/  1/2  1/2  2 1/2   c      m   m  inc    h  3 2 FGA m ¸  ? 1 { α 1{2α ˜  ˆ „    ˙ 1/2  w )   Rinc  c 1{2α lFGA h 2FGA hFGA  lFGA  4( 3 max2µ µm pµm `0µ.11 × 105  3. Relationship between σ max and Np/Nf.  it  should  be  noted  that  herein  proposed  VHCF  life  prediction  model  is boundaries. However.44 × 104  1.65 × 105  2. As long as the material exhibits this failure mechanism. As long as the material exhibits this failure mechanism.80 × 105  2.31 × 105  900 850 max . On the whole.  On  the  whole. this model is still needed to verify by more experimental data of diverse materials. the proposed model can be experimental data of diverse materials. 8. Comparison of predicted and experimental lives. from the viewpoint of reliability. 8. However. Total Fatigue Life Model Based  on  Equations  (22)  and  (31). Comparison of predicted and experimental lives.625  600  600  575  550  Materials 2015.45 × 105  2.005µinc 2 l ` l 4pσ ´ σ q R (32) max w FGA FGA FGA inc  Rinc d Rinc ff «E    1 26 27 c   πE2 2(max ) 2  Rinc RFGA RRinc  Fisheye  ` ´ 27 1 ` 26 2 RFGA RFisheye 2p∆σmax q where Npre is the predicted total fatigue life.  Furthermore.7. the proposed model can  the crack initiation and growth processes.08 × 105  2. Conclusions  8352 Main conclusions obtained in this study are summarized as follows:  (1) The  Cr‐Ni‐W  gear  steel  exhibits  the  constantly  decreasing  S‐N  property  without  traditional  fatigue limit. life.  Figure 15. Materials 2015. the  According agreement  is  good  within  the  factor‐of‐three  where Npre is the predicted fatigue to fairly  the comparison between the predicted boundaries.41 × 105  2.27 × 105  2.31 × 105  2.  . from the viewpoint of reliability. page–page  15  9 Predicted life Npre . it should be noted that herein proposed VHCF life prediction model involves failure mechanism. Relationship between σmax and Np /Nf . Total Fatigue Life Model 3. and the fatigue strength corresponding to 109 cycles is around 485 MPa. the agreement is fairly good within the factor-of-three acceptable.005 m (32) inc qinc    “ ph q 0. 4. 3. According to the comparison between the predicted and  experimental  lives  shown  total in  Figure  15.7.  the  result  is  satisfactory  and  the  above‐established  model  is  and experimental lives shown inpredicted  Figure 15. cycles 10 8 10 7 10 Factor of 3 6 10 6 10 7 10 8 10 9 10 Experimental life Nexp .36 × 104  1.90 × 104  1. MPa 800 750 700 650 600 -3 -2 10 -1 10 10 Np / Nf   Figure 14.  Figure 14.34 × 105  3. and reflects the inclusion‐FGA‐fisheye induced  acceptable.

. A. Struct. Mater. Z. Lifetime measurements for random loading in the very high cycle fatigue range. Sakai.. Zhang. [CrossRef] Yang. [CrossRef] Lai. Danninger. Fatigue Fract. Matsumura. Lund. Mater. 769–780. Hui. A 2006. 195–200. C. A 2015. 9. Sci. Nagata. 25. 427. On the critical inclusion size of high strength steels under ultra-high cycle fatigue. Int. Conflicts of Interest: The authors declare no conflict of interest. Mater. Q. 123–147. Mater. [CrossRef] Li. J. Fatigue 2012.. Wang. 167–174. [CrossRef] Sohar. [CrossRef] Ochi.. Zhao. J.. Brodvanski. E. 37. M. 823–823. Fatigue 2008. Deng. Mechanism of fatigue failure in ultralong life regime. Sato.. 25. W. Qian.. Int.N. 2. H.T.Y. and the fatigue strength corresponding to 109 cycles is around 485 MPa.L.. Mater. Li. K. Eng. 765–773.Q. 45.. Sci.. [CrossRef] 8353 . 67. Struct. H. while the predicted crack growth life is only accounts for a tiny fraction. Y. 3. Z.. High-cycle rotating bending fatigue property in very long-life regime of high-strength steels. Eng. H. 781–790. Author Contributions: All authors contributed extensively to the work presented in this paper. [CrossRef] Murakami.. 30. 12. The inclusion-FGA-fisheye induced failure becomes the main failure mechanism of Cr-Ni-W gear steel in the VHCF regime. 2002. Struct. B.M.. 25. Ishihara. Gabelli. 10. W. Mater. Murakami. The predicted crack initiation life occupies the most majority of total fatigue life. Extremes 1999. Kopnarski. S.. Subsurface inclusion-induced crack nucleation and growth behaviors of high strength steels under very High cycle fatigue: Characterization and microstructure-based modeling. Eng. L. Li.R. J. P. Mayer. models and statistical implications.G. 1137–1149..X. K.D. Weiss. C.Materials 2015. G. 2009. Fatigue 1986. Essential characteristics and influential factors for very-high cycle fatigue behavior of metallic materials. 735–746. Beretta. [CrossRef] Hong. 24. 2012. J. 641. Tanaka-Mura model and modified Chan model were used to evaluate the crack initiation life. Lu. Rydén... 8338–8354 4. Yoshida. Eng. 2002.J. Y.Y.Y. Eng. 2002. G.. 6. 838–841. 5... L... T. 8. Lu. Betzwar-Kotas.S. Gierl.. Y. Mechanism of fatigue crack initiation and propagation in the very high cycle fatigue regime of high-strength steels. Reuscher. 11. I. The local stress-life model. Mater. B. Fatigue Fract. Weng. Scr. 155–168. Oguma.. 8. [CrossRef] Shiozawa. J. Grad. Y. S.. By using the finite element analysis of representative volume element. 4. Tschegg. Z. Acta Metall Sin.. N. Int... 7.. Struct. The established VHCF life prediction model involving crack initiation and growth for inclusion-FGA-fisheye induced failure is acceptable based on the good agreement between the predicted and experimental results.. Fatigue Fract. T. S. 10–20. Sun. T. J. Eng. E. Strandell.. [CrossRef] Sakai. S.A. K. 8. 2001.. Conclusions Main conclusions obtained in this study are summarized as follows: (1) (2) (3) (4) (5) (6) The Cr-Ni-W gear steel exhibits the constantly decreasing S-N property without traditional fatigue limit. S.. S-N curve characteristics and subsurface crack initiation behaviour in ultra-long life fatigue of a high carbon-chromium bearing steel. Y... and the predicted results based on modified Chan model are better. Y. T.G.. The fatigue limit of bearing steels-Part I: A pragmatic approach to predict very high cycle fatigue strength. 2. Zhang.. Yokoyama. References 1. Stanzl. N. Acknowledgments: This research was supported by the National Natural Science Foundation of China (No. Masaki. A. 51305027). Fatigue Fract. A. Characteristic S-N properties of high-carbon-chromium-bearing steel under axial loading in long-life fatigue. the local stress around the inclusion tends to increase with the increase of elastic modulus difference between inclusion and matrix. Small defects and inhomgenities in fatigue strength: Experiments. Gigacycle fatigue behavior of a high chromium alloyed cold work tool steel. A. Kerscher.

Hill. K.O.. Struct. Eng. Lei. A Theory of Fatigue Crack Initiation at Inclusions. [CrossRef] Venkataraman.. 11. A. Fract. H. Fatigue 2010. 1555–1565. USA. Int. 28. G. H. Phys. Y. Metal Fatigue in Engineering. Y.. pp. K.. M. [CrossRef] Stephens.. S. Metall. 25. J. Int. 117–123.R. A Wiley Interscience Press: New York. [CrossRef] Wang.. [CrossRef] Paris. USA. Fatigue 2009. [CrossRef] Hertzberg. T. R. 1978. J.. 19. [CrossRef] Liu.S. Nakamura.W. Y. Mater. Mater. Trans. Proc. J. [CrossRef] Anderson. A quick and simple method for orienting cubic single crystals from Laue back-reflection photographs. 21. Tagawa. Miyamoto.0/). Y. R. 661–662. J. 2008. 28.. Metall. NY. Fatigue Fract. Cerullo. Sci. 1991. 29. Int. 74. Oguma. 1351–1357.. Fracture Mechanics. 9. [CrossRef] Cheng.. [CrossRef] Mayer. Lei. 34.. pp. 21. CRC Press: Boca Raton. 24. American Society for Testing and Materials (ASTM) Press: Philadephia. Schuller.. 19–27.J. Z. K. Eng. [CrossRef] Hoeppner. FL. 32. Pergamon Press: Fukuoka.. [CrossRef] Murakami. H. Mura. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons. Zhao. Proc. Fundamental and Application.. Sci. 333–338.J. USA. 18. 134–158. Shinohara. Calculation of slip band spacing. 31. Bathias. Z. Li. 236–244. N.S. 356. Effects of inclusion size and stress ratio on fatigue strength for high-strength steels with fish-eye mode failure.. Crystallogr. Rolling contact fatigue life model incorporating residual stress scatter. Stress Intensity Factor Handbook. [CrossRef] Sun. Fatemi. Y.. T. Elastic properties of reinforced solids: Some theoretical principles. Bacher-Hochst. Notch effect on fatigue strength reduction of bearing steel in very high cycle regime. PA. R. 1993. Prediction of the S-N curves of high-strength steels in the very high cycle fatigue regime. W. N.. Fatigue 1999. Choi. S.S. 84–84. 35. Mech. T. B. 2012. 1982. M.W. 509–551. J. Acta Metall. C.. Fatigue 2006. T. Int. Kawagoishi. Mater.Y. 357–372. 23. Donald. Li.. Switzerland. Henry. 20.D. 1993. Mech.. 2010. P. Effect of inclusion on subsurface crack initiation and gigacycle fatigue strength. Chen. Y. H. J.R. Service load fatigue damage-a historical perspective. 1991. 638–647. Mura. 15. 80–83. Fatigue Testing of Weldments. 17.W. A 2003. Liu. An examination of load shedding during fatigue fracture. Tsuru. 43–58. 33. Xie. 50. A cumulative damage model for fatigue life estimation of high-strength steels in high-cycle and very-high-cycle fatigue regimes.. Eng. [CrossRef] Akiniwa. A microstructure-based fatigue-crack-initiation model. Chan. 35–46..G. Int.. J. licensee MDPI. pp. Basel. Fatigue 2013. 8338–8354 13. 8. A.. 2.Materials 2015. 242–249. J. Y. Miyata. Z. T. [CrossRef] Chapetti. Japan.. 1572–1577. Tanaka. C. Int. Chung.Q. Eng.X. Trans. Haydn. 15. 8354 .I. pp.. 2014.. J.. 1. Tada. 24.. Q. Fatigue 2002.... Stephens. Hong. 32. Laird. Int. Sub-surface fatigue crack growth at alumina inclusions in AISI 52100 roller bearings. 22. T. C. Very high cycle fatigue properties of bainitic high carbonchromium steel.Q. Pecorini. Mater. Hong. A. [CrossRef] Sun.. J. Zhao. [CrossRef] © 2015 by the authors. B. 13. Mater.. D. 30. 14. R. A T. J. A 2003. C. 16.K. I. Q.B. Furtner.L. Solids 1963. 31. H. [CrossRef] Tanaka... 137–137.. Xie.C. Y. R. A. C. 2631–2638. Int. Application of minimum energy formalism in a multiple slip band model for fatigue. 27. The effect of vacuum-like environment inside sub-surface fatigue crack on the formation of ODA fracture surface in high strength steel.D. Yang. 26. Appl. Issler. Y. Ultra-long cycle fatigue of high-strength carbon steels part II: Estimation of fatigue limit for failure from internal inclusions. 48. J. 1269–1274. 2001. J. 2121–2129.