Tribological behaviour and tool wear analyses in rough turning of large-scale parts of nuclear power plants using grooved coated insert

Tribology International 80 (2014) 58–70

Contents lists available at ScienceDirect

Tribology International journal homepage: www.elsevier.com/locate/triboint

Tribological behaviour and tool wear analyses in rough turning of large-scale parts of nuclear power plants using grooved coated insert B. Haddag a,n, H. Makich a, M. Nouari a, J. Dhers b a University of Lorraine, Laboratoire d’Energétique et de Mécanique Théorique et Appliquée, LEMTA CNRS-UMR 7563, Mines Nancy, Mines Albi, GIP-InSIC, 27 rue d’Hellieule, 88100 Saint-Dié-des-Vosges, France b AREVA NP, Usine de Chalon Saint Marcel, 71380 Saint-Marcel, France

art ic l e i nf o

a b s t r a c t

Article history: Received 17 March 2014 Received in revised form 16 June 2014 Accepted 22 June 2014 Available online 2 July 2014

The tribological behaviour and tool wear mechanisms in rough turning, using coated grooved insert, of a large-scale part, made of 18MND5 steel used for steam generators of nuclear power plants, have been analysed. Characterisation of the insert rake face, using various techniques (SEM, EDS and Profilometer), has revealed an intense thermomechanical loading with wear localisation at particular zones, and formation of an adhered thin tribo-layer. A numerical model has been developed to predict finely the tool wear as revealed by experimental observations. The main findings concern the prediction of contact discontinuities and where the wear is highly localised as well as thermomechanical conditions causing observed phenomena, like adhesion of the workmaterial on some zones on the rake face. Contact discontinuities are attributed to the complex geometry (presence of concave zones) of the coated grooved insert, designed especially to reduce the tool–chip contact area and to promote the chip breakage. Other cutting parameters, like the specific cutting force and chip morphology parameters are also well predicted. This is a contribution to understand wear mechanisms occurring on complex cutting inserts at high material removal rate (moderate cutting speed, but high depth of cut and feed rate). & 2014 Elsevier Ltd. All rights reserved.

Keywords: Rough turning Large-scale part Tribological behavior Wear localization

1. Introduction Various machining processes, such as turning, deep drilling and broaching, are used to realize large scale parts, having few meters, in the construction of nuclear power plants. The forged parts are machined following several steps to obtain specified workpiece geometry with required technical specifications. For instance, for shells composing steam generators of nuclear power plants, the rough turning process is the first machining operation, on the forged component, to remove a maximum of the workmaterial. High material flow rate is required to optimize the productivity. Vertical turning machines are usually used to support huge workpieces (e.g. parts of more than 5 m of diameter and height). The cutting inserts used in the rough turning operation undergo a strong thermomechanical loading, which generates an excessive tool wear and therefore reducing drastically the tool life. This leads to change several times cutting inserts, which necessarily needs stopping the cutting operation and hence affects the productivity. The analysis of the cutting process in rough turning, including the tool wear, is therefore justified to understand mechanisms of tool wear and further to improve the cutting operation. There is no

n

Corresponding author. Tel.: þ 33 3 29 42 18 21; fax: þ33 3 29 42 18 25. E-mail address: [email protected] (B. Haddag).

http://dx.doi.org/10.1016/j.triboint.2014.06.017 0301-679X/& 2014 Elsevier Ltd. All rights reserved.

criterion on cutting conditions to qualify the rough turning operation, since it depends on global dimensions of the workpiece, but it can be defined as first cutting steps where the objective is to remove the maximum of the workmaterial before subsequent finishing steps. In the case of rough turning of large scale parts (having several meters), like that studied here, it corresponds to depth of cut of at least 10 mm and feed rate of about 1 mm. However the cutting speed is very limited on the vertical lathe (less than 100 m/min) to reduce inertia effects and vibration due to the size of the workpiece. This allows obtaining an acceptable material flow rate. The cutting operation can take several hours. The turning process is widely studied in the literature. However few research works are dedicated exclusively to the rough turning operation. For instance, Diniz and Oliveira [1] have performed rough turning tests under dry and wet conditions. Their work aims to seek conditions in which dry cutting is satisfactory compared with the flood of fluid (called wet cutting) usually used. They conclude that if the tool material has a good wear resistance, dry cutting can be used with similar cutting conditions to those used with a flood of fluid. Kee [2] has developed constrained optimisation analyses and strategies for selecting the optimum cutting conditions for multi-pass rough turning operations on CNC and conventional lathes. Since in rough tuning the cutting tool undergoes strong thermomechanical loading involving excessive tool wear, particularly when machining hard materials, the choose

B. Haddag et al. / Tribology International 80 (2014) 58–70

Nomenclature vc f ap

κr λs γo

F Fc Ff Fp Kc K cc Kf c K pc Pc h hch

λh σ

fv

ρ

E

ν A B n C m

εp p ε_

cutting speed [m/min] feed rate [mm/rev] depth of cut approach angle [1] inclination angle [1] orthogonal rake angle [1] resultant cutting force [N] cutting force [N] feed force [N] passive force [N] specific resultant cutting force [MPa] specific cutting force [MPa] specific feed force [MPa] specific passive force [MPa] cutting power [W] uncut chip thickness chip thickness chip compression ratio true or Cauchy stress tensor [MPa] body force density [N/m3] material density [kg/m3] Young modulus [GPa] Poisson’s ratio initial uniaxial tension stress of the workmaterial [MPa] strain hardening parameter of the workmaterial [MPa] strain hardening exponent parameter of the workmaterial strain-rate sensitivity parameter of the workmaterial temperature sensitivity parameter of the workmaterial von Mises equivalent plastic strain von Mises equivalent plastic strain-rate

of adequate inserts is important. So Serdyuk et al. [3] have analysed the influence of heat treatment parameters on wear mechanisms of T5K10 coated carbide insert, similar to that analysed in this paper, used for rough turning. The authors have stated that the tool life is altered by the presence of residual microporosity on cemented carbide structure. So the microporosity should be reduced in the fabrication process to improve the tool life in rough turning. Globally, the tool wear in turning process is widely investigated by experimental means. Grzesik and Zalisz [4] have investigated the wear mechanisms in dry, hard and finishing turning using ceramic tools (WC–Co), with varying only the feed rate. The wear mechanisms involve abrasion, fracture, plastic flow, material transfer and tribochemical effects which appear depending on the mechanical and thermal conditions generated on the wear zones. Ozcelik et al. [5] have focused on the impact of various vegetable based cutting fluids with extreme pressure during turning of AISI 304L. Authors have examined the performance of the cutting operation (cutting force reduction, surface roughness improvement) and the tool wear. Different wear types are revealed (e.g. fracture of the engaged cutting edge in dry cutting, and adhesion of the workmaterial on the rake face in wet cutting). Hakim et al. [6] have analysed the wear behaviour of some coated cutting inserts in hard turning of HSS. It was concluded that some coatings promote the formation of a tribolayer at the rake face, which slows the wear increase. Bhatt et al. [7] have studied wear mechanisms of WC coated and uncoated tools in finish turning of Inconel 718. It was concluded that abrasive and adhesive wear

ε_ 0 σ σn τf

m

τmax τ β

vs T T0 Tm Tt Tw

λ cp

α ηp ηf

ff h q_ p q_ c q_ -t q_ -w _ w, w a, b

59

reference equivalent plastic strain-rate von Mises equivalent stress [MPa] normal friction stress [MPa] shear friction stress [MPa] friction coefficient shear stress limit [MPa] von Mises equivalent shear flow stress [MPa] shear limit factor [MPa] sliding velocity at the tool–workpiece interface [m/s] temperature [1C] reference ambient temperature [1C] melting temperature [1C] tool temperature at the tool–workpiece interface [1C] workpiece temperature at the tool–workpiece interface [1C] thermal conductivity [W/m/1C] specific heat capacity [J/kg/1C] thermal expansion [mm/m/1C] plastic work conversion factor (Taylor–Quinney factor) frictional work conversion factor heat partition coefficient of the frictional heat heat transfer coefficient for the tool–workpiece interface [kW/m2/1C] volumetric heat generation due to plastic work [W/m3] heat conduction flux at the tool–workpiece interface [W/m2] heat flux going into the tool at the tool–workpiece interface [W/m2] heat flux going into the workpiece at the tool–workpiece interface [W/m2] wear depth [mm] and wear rate respectively [mm/s] wear law coefficients

were the most dominant wear mechanisms, controlling the deterioration and final failure of the WC tools. Lim et al. [8] have studied the performance of TiN coated and uncoated HSS inserts in turning. They have concluded that the benefits of TiN coatings on HSS tools may be easily realized over a wide range of cutting conditions when machining a hotrolled medium carbon steel (AISI 1045 equivalent). Authors have indicated that coatings offer the greatest advantage when operating under high speeds and feeds. Jianxin and Xing [9] have investigated the machining performance and wear mechanisms of two alumina-based ceramic cutting tools in continuous turning of hardened steel and nickel based alloy (Inconel 718). It was reported that the coating type has an impact on the amount of flank were in turning of hardened steel. While in continuous turning of Inconel 718, one coating type showed greatly improved flank wear resistance. Adhesion and abrasion wear were found to be the dominant wear mechanisms. Ezugwu et al. [10] have investigated the performance of uncoated carbide tools when rough turning Ti–6Al–4V alloy under flood cooling (conventional and high pressure cooling). Authors have highlighted the impact of fluid pressure on the surface roughness and the tool life improvement. Modelling the tool wear in the context of machining was also a concern of the scientific community. Both analytical and numerical analyses are proposed. Usui et al. [11,12] have proposed a phenomenological model to predict tool wear. The tool wear law is function of the tool–chip interface parameters (contact pressure, temperature and sliding velocity). Later, more physical models have been proposed. For instance, Molinari and Nouari [13] and

60

B. Haddag et al. / Tribology International 80 (2014) 58–70

Nouari and Molinari [14] have proposed a physical model to predict diffusive tool wear. The model has been applied for high cutting speed under orthogonal cutting configuration. With the development of finite element codes and their ability to simulate metal machining processes, wear evolution laws are implemented to predict the wear rate and hence the worn tool geometry. The laws are assessed with predicted quantities at the contact interface. The evolution of the tool geometry during machining is taken into account, as done by Yen et al. [15], Xie et al. [16], Filice et al. [17], and Lorentzon and Järvstrat [18] for a 2D cutting configuration, and Attanasio et al. [19,20] for a 3D cutting configuration. But these approaches are marginal, since it necessitates advanced and particular numerical developments to perform, particularly that related to remeshing the cutting tool, thermomechanical fields transfer and contact control when the tool rake face geometry evolves. Thus, Haddag and Nouari [21] recently developed a multisteps 3D FE modelling of the turning process to predict the tool wear as well as the heat diffusion in the cutting tool using interface thermomechanical fields. In Haddag et al. [22] uniform and non-uniform heat fluxes applied at the tool rake face are assessed to analyse the tool heating. The objective of this research paper is to analyse the tribological behaviour and tool wear in rough turning of a large-scale part of nuclear power plants using grooved coated insert. Experimental and theoretical analyses are performed. First, an experimental work to highlights the intense thermomechanical loading on the cutting insert in rough turning has been performed. The tool wear has been characterized using optical microscope, scanning election microscope and profilometer observations. Then, a finite element modelling has been developed to predict the intense thermomechanical loading at the tool–workpiece interface, including tool wear prediction. Finally, comparisons have been performed between experimental observations and numerical predictions of

Workpiece 18MND5

Tool holder Insert SNMM250924RH Fig. 1. Rough turning operation on vertical lathe of a large-scale part (cylindrical shell) made of 18MND5 steel using SNMM250924RH coated insert. The cutting configuration is defined by angles κ r ¼ 751, λs ¼  61, γ o ¼  61.

tool wear zones as well as the chip formation. The nature of the tool–workmaterial contact has been discussed.

2. Tribological behaviour and tool wear analyses in rough turning 2.1. Description of the rough turning operation on a large-scale part A rough turning operation on a large-scale part has been performed in an industrial site of AREVA NP Company. It consists on the machining a large shell part, having cylinder form, as a component of steam generators of nuclear power plants. Fig. 1 shows the experimental setup of the rough turning operation. The cutting test has been performed on a vertical lathe supporting large workpieces (about 150 t for the shell part). On this type of machine, the cutting speed is limited to reduce inertia and vibration effects due to the size of the workpiece. However the depth of cut and feed rate are high enough to obtain an optimal material removal rate. The cutting configuration is defined by three angles: approach angle κ r equal 751, inclination angle λs equal 61, and orthogonal rake angle γ o equal  61. The cutting condition that has been taken for the cutting test corresponds to cutting speed vc ¼ 100 m/min, feed rate f ¼1 mm/rev and depth of cut ap ¼10 mm. As it can be deduced, the insert is engaged in the workmaterial with a large cutting area. The unformed chip area can be approximated by Ac ¼ f :ap . Under such cutting condition Ac is about 10 mm2, which is a large area compared to that found in conventional rough turning of workpieces having small or moderate dimensions. The objective is to remove in minimum time a maximum of workmaterial in the rough turning step. 2.2. Cutting insert and workmaterial A square form insert, shown in Fig. 2(a), designated by SNMM250924RH, have been used as a cutting tool. Its dimension characteristics are reported in Table 1. The insert is made of cemented tungsten carbide with cobalt as binder phase (WC–Co), and coated with KC8050 coating type (TiN–AL2O3–TiCN layers with TiN as the extern layer). The tool rake face geometry is shown in Fig. 2(b). The machined part is a component of a steam generator in nuclear power plants having a cylindrical shape, with more than 5 m of extern diameter and about 6 m of height. The workpiece is Table 1 Dimension characteristics of the SNMM250924RH insert, indicated in Fig. 2(a). D (mm)

L10 [mm]

S [mm]

Rε [mm]

D1 [mm]

25.4

25.4

9.53

2.4

9.12

Fig. 2. (a) Roughing coated insert SNMM250924RH used in the cutting test and (b) highlight of its rake face geometry.

B. Haddag et al. / Tribology International 80 (2014) 58–70

made of low alloy steel 18MND5, having a bainitic structure. Its chemical composition is reported in Table 2. The content of all additive elements is less than 5%. 2.3. Cutting process The chips morphology and cutting insert are examined after machining. Fragmented chips are generated during the cutting operation, as shown in Fig. 3(a). The average chip thickness is about 1.5 mm under the considered cutting condition. The global view of the insert after machining is shown in Fig. 3(b). 2.4. Tool wear characterisation The tool wear has been investigated using Optical Microscopy (OM), Scanning Electon Microscopy (SEM), Energy Dispersive X-ray Spectroscopy (EDS) as well as Optical Profilometry (OP). The OM observation of the tool rake face, as shown in Fig. 4(a), reveals several localized wear zones, and non-altered zones, which suggest that there are permanent contact discontinuities during the cutting. This is due, at first explanation, to the grooved face of the insert, where contact is lost at concave zones on the rake face. In order to analyse the nature of the wear on the rake face, SEM observations are performed for a better analysis of tool wear mechanisms. As shown in Fig. 4(b), there is several localized wear zones, where adhesion of the workmaterial or abrasive wear occurs. Three localised zones, noted Z1, Z2 and Z3, are identified in Fig. 4(b) for a thorough analysis. To confirm these wear mechanisms, EDS analyses are performed on these identified zones. SEM observations of Fig. 5 reveal that there are combined adhesion and abrasion wear mechanisms on the rake face. This suggests that the workmaterial has attained at lest half the absolute melt temperature under high contact pressure in some zones during the cutting process; this will be confirmed further Table 2 Chemical composition (wt%) of the 18MND5 steel.

61

with the finite element simulation in Section 4.2. In Z1, as shown in Fig. 6(a), iron is clearly highlighted with yellow colour of the observed feature, which is the main chemical element of the workmaterial. In Z1, Z2 and Z3, one can shows that magnification of some zones (see Fig. 5) reveals parallel lines which correspond to the trace left by the chip flow on the rake face; the direction of these lines is commonly known as the local chip flow direction. These flow lines are quasi perpendicular to the straight part of the cutting edge, since the engaged rounded edge is lower compared to the engaged straight part. The formation of observed stripes lines is at the origin of the crater wear development. Using the OP technique, the crater depth as well as the thickness of the adhered workmaterial layer can be estimated. Advanced manipulation allows quantifying the volume of the formed crater and adhered layer. As shown in Fig. 7, in Z1 and Z2, which are close to the rounded cutting edge, OP images indicate clearly the amount of adhered workmaterial layers and their form. The concave shape in these zones combined with high contact pressure and temperature lead to the adhesion or trapping of the workmaterial in form of a thin tribolayer. The thickness of

5mm

Z3

Z2

Z1

5mm

C

Mn

Si

Ni

Cr

Mo

0.18

1.4

0.17

0.72

0.17

0.5

Fig. 4. (a) OM and (b) SEM micrographs of the engaged cutting face (see Fig. 3(b)) of the SNMM250924RH insert after machining, highlighting several localized wear zones.

Cutting face

Fig. 3. (a) Typical generated chip, and (b) insert after rough turning, highlighting the engaged cutting face analysed in next sections.

62

B. Haddag et al. / Tribology International 80 (2014) 58–70

Fig. 5. SEM observations of three localized wear zones, highlighting wear types and the flow direction of the workmaterial on the rake face: (a) Z1, (b) Z2 and (c) Z3 are identified zones in Fig. 4(b). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

the adhered layer in Z1 is about 30 mm, and in Z2 is about 10 mm. By cons, in the Z3, where the chip leaves the rake face, there is combined wear mechanisms with low adhesion but more abrasion wear, which is the consequence of chip sliding under low contact pressure in this zone.

Fig. 6. EDS analyse of the three localized wear zones: (a) Z1, (b) Z2 and (c) Z3, revealing the adhesion of the workmaterial at the rake face mainly at Z1 and Z2, and dominant abrasive wear in Z3. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

B. Haddag et al. / Tribology International 80 (2014) 58–70

63

Fig. 7. OP analyses: 3D profiles and top view of the three localized wear zones, revealing adhesion of the workmaterial at (a) Z1 and (b) Z2, and dominant abrasion wear at (c) Z3.

3. Modelling of the rough turning operation In order to explain the observed tribological behaviour of the tool–workmaterial interface, and highlighting the origin of the observed adhesion of the workmaterial and abrasion wear in some zones of the rake face, a FE-based numerical model has been developed for the rough turning operation. The development of an efficient predictive numerical model should take into account the following aspects: (i) cutting configuration representing finely the tool cutting face geometry, (ii) adequate mesh of the cutting zone, (iii) adequate mechanical and thermal boundary conditions, and

(iv) representative constitutive equations of the workmaterial, the tool as well as the contact interface. The challenge is to predict finely observed physical phenomena, and especially the thermomechanical loading which induced it. The cutting process is modelled by representing the tool–workpiece couple with the consideration of the vicinity of the cutting zone. The physical properties of the workpiece and tool materials are given in Table 3. Note that the tool coating is taken into account implicitly through parameters involved in the contact interface equations, described in Section 3.2. However as reported by Koné et al. [23], since the simulated cutting time is very short, predicted contact interface

64

B. Haddag et al. / Tribology International 80 (2014) 58–70

Table 3 Mechanical and thermal properties of the workmaterial and tool.

Table 5 Tool–workpiece interface parameters.

Parameters

Workmaterial (18MND5)

Tool (WC–Co)

μ

β

ηf

ff

h [W/m2/1C]

ρ [kg/m3] E [GPa] v cp [J/kg/1C] λ [W/m/C] α [1C  1]

7800  0:125T þ 221:5 0.3 0:744T þ 225:6 0:0116T þ 15:87

11900 650 0.25 1260 59

–

0.6

1

0.5

45.103

2:10  9 T þ 10  5 1430 20

5:10  6  20

T m [1C] T 0 [1C]

pffiffiffi

τ ¼ σ = 3 is the plastic shear flow of the workmaterial. Since there

Table 4 Johnson–Cook flow stress parameters of 18MND5. A [MPa] 480

B [MPa] 456.7

n

C

0.29

0.016

m 0.91

ε_ 0 10

–3

T m [1C]

T 0 [1C]

1430

20

is a strong contact at the tool–workmaterial interface (high feed and depth of cut), it is considered that the sliding occurs when the shear friction stress τf reaches a certain amount of the shear flow stress τ of the workmaterial due to the strong contact pressure in the rough pffiffiffiturning test. This is controlled by the condition τf ¼ βτ ¼ 3βσ when sliding occurs. OM and SEM observations of Fig. 4 confirm the intense thermomechanical loading at contact zones. The friction at the tool–workpiece contact interface may generate a heat, which is evaluated by the following relation: q_ f ¼ ηf τf vs

fields (e.g. pressure, temperature, sliding velocity) is less affected by the presence of coatings due to their small thickness. 3.1. Constitutive equations During the chip formation, the workmaterial undergoes large deformations with strain rate and temperature effects. Hence the behaviour of 18MND5 mild steel, as a workmaterial, is described by a thermo-viscoplastic flow stress, given by the classical Johnson–Cook law [27] as follow:
  vp  kðTÞ ð1Þ Y ¼ g εvp h ε_ with

n g εvp ¼ A þ B εvp 8  vp    < 1 þ C ln ε__ vp ε0 ¼ h ε_ :1

where vs is the relative sliding velocity, τf is the friction stress given by Eq. (4), and ηf is the frictional work conversion factor. Assuming all the frictional work converted into heat, ηf ¼ 1 is often considered. The heat generated by viscoplastic strain at the secondary shear zone (chip surface under contact) creates a temperature gradient at the contact interface which is taken in the heat balance equations by introduction of a heat conduction flux q_ c . So the heat balance at the interface can be written as follows: q_ -t ¼ f f q_ f þ q_ c q_ -w ¼ ð1  f f Þq_ f  q_ c

ð6Þ

with if if

vp ε_ 4 ε_ 0



T T0 kðTÞ ¼ 1  vp Tm T0 ε_ r ε_ 0

q_ c ¼ hðT w  T t Þ

m ð2Þ

where A, B, C, m and n are material parameters, ε_ 0 the reference equivalent plastic strain rate, T m and T 0 are, respectively, the material melting temperature and the reference ambient temperature, while εvp the von Mises equivalent viscoplastic strain ffi R vp R qffiffiffiffiffiffiffiffiffi vp 2 _ vp _ vp and ε_ its rate, related by εvp ¼ ε_ dt ¼ ε ε dt. : 3 The Johnson–Cook law parameters of the 18MND5 are identified form dynamic rheological tests, using Hopkinson bars device. Tensile and compression specimens are used. These tests are performed under low and high temperatures. The identified parameters are reported in Table 4. The viscoplastic deformation is accompanied by a heat generation resulting in the temperature rise in the workmaterial. It is described by the following relation: vp vp q_ p ¼ ηp σ : ε_ ¼ ηp σ ε_

ð3Þ

where ηp is the plastic work conversion factor (Taylor–Quinny factor). 3.2. Tool–workpiece interface behaviour The contact behaviour at the tool–workpiece interface is defined by the relationship between the normal friction stress σ n and the shear friction stress τf as follow: ( μσ n if μσ n o βτ τf ¼ ð4Þ βτ if μσ n Z βτ where

ð5Þ

μ is the friction coefficient, β is the shear limit factor and

ð7Þ

where q_ -t and q_ -w are, respectively, heat fluxes going into the tool and workpiece from the tool–workpiece interface, and T t and T w are respectively temperatures on the tool and workpiece surfaces at this interface, f f is the heat partition coefficient of the frictional heat and h is the interface heat transfer coefficient. The tool–workmaterial interface parameters are reported in Table 5.

3.3. Tool wear model To estimate the tool wear for the considered tool–workmaterial couple in the adopted cutting configuration and under considered cutting condition, the phenomenological Usui wear law [11,12], expressed as follow, is used: _ ¼ aσ n vs expð  b=TÞ w

ð8Þ

The objective is to capture localized wear zones, rather than giving a quantitative analysis of the tool wear. In the frame of FE modelling, the wear depth w is assessed by time integration of Eq. (8) using the following relationship, as performed recently by Haddag and Nouari [21]: Z wðx; tÞ ¼ aσ n ðx; tÞvs ðx; tÞexpð  b=Tðx; tÞÞdt ð9Þ where σ n , vs and T are local values of contact pressure, sliding velocity and temperature, respectively, computed at each FE nodal point x at the contact face of the cutting tool and at the end of each cutting time increment dt, while a and b are experimentally calibrated coefficients reported in Table 6.

B. Haddag et al. / Tribology International 80 (2014) 58–70

3.4. Finite element model To analyse finely the tribological behaviour and tool wear in the rough turning test a 3D FE model has been developed in Deform FE software [24]. This allows predicting the local thermomechanical loading accruing at the contact interface during the cutting process. Fig. 8(a) shows the cutting configuration as in the cutting test. The cutting tool is considered thermo-rigid, while the workmaterial is represented by the thermo-viscoplastic behaviour described in Section 3.1. The workpiece and tool are meshed with tetrahedron finite element of Deform library, a coupled linear displacement-temperature four nodes element, with about 240,000 for the insert and 100,000 for the workpiece at the initial configuration. With the insert penetration in the workmaterial, the workpiece is remeshed to form de chip and the finite elements number increases to capture geometrical non-linearities. The Table 6 Tools wear law parameters. a 10  5

b 1000

65

remeshing occurs automatically if some finite elements are distorted (if the code detects negative Jacobian). Also, remeshing is done every specified number of increments, independently of the finite elements distortion. For the developed FE model, remeshing is performed at specified interval of 10 time increments. It is possible to add other remeshing criteria based on the variation of state variables (like strain, strain rate, temperature…) over time increment, by specifying the limit not to be exceeded. However this may increase remeshing number, which may deteriorate at the end the quality of the expected solution. A minimum finite element size of about 100 mm is taken in the insert (at the engaged part of the rake face), and about 400 mm in the chip. These values are chosen after some numerical tests to find a compromise between result accuracy and reasonable CPU time. The computation time for the final FE model is about 21 days. Fig. 8(a) shows the FE model before chip formation, Fig. 8(b) with chip formation and Fig. 8(c) part of the insert, with mesh refinement of the engaged cutting face. To represent the cutting operation, mechanical and thermal boundary conditions are defined. For the mechanical one, the bottom face of the workpiece is fixed and the cutting speed vc is applied to the rigid tool. For the thermal one, at the initial

Fig. 8. FE model representing the rough turning test: (a) initial mesh, (b) remeshing the workpiece during the chip formation, and (c) zoom on the mesh of the engaged cutting face.

66

B. Haddag et al. / Tribology International 80 (2014) 58–70

configuration the reference ambient temperature T 0 is imposed for the tool–workpiece couple. While during the chip formation, at the tool–workpiece interface equations of the contact behaviour (see Section 3.2) are considered. Note that the heat exchange with the environment is neglected, since the simulated cutting time is very short (few microseconds) to obtain a significant effect on the chip formation process.

The theoretical chip width can be approximated by: w  ap = sin ðκ r Þ

The chip compression ratio, defined as the ratio of the chip thickness hch by the uncut chip thickness h (see e.g. Kouadri et al. [26]), has been assessed as follow:

λh ¼ 4. Results analysis and discussion Two complementary analyses of the rough turning operation are performed. One concerns the cutting process to validate the FE model, in which the chip morphology and the cutting force are analysed. The other concerns the tribological behaviour and tool wear analyses, through observed and predicted localized wear on the rake face of the insert. 4.1. Cutting process analysis The cutting process is analysed through the chip morphology and the cutting force. The generated chip in the rough turning operation is quasi-continuous with small amount of segmentation. During the turning test the chips fragmentation occurs when chips reach a certain cutting length (about few centimetres), due to the increase of their curvature during cutting. Fig. 9 shows a comparison between experimental and predicted chips, where the continuous chip is well predicted. However the chip fragmentation is not reproduced. Fig. 10(a) shows different views of the cutting process, where the contact at the tool–chip interfaces is clearly discontinuous. Such information is not accessible for observation by experimental means, like using CCD camera to follow the chip formation process. In addition, the simulation of the cutting process indicates that the chip flow direction on the rake face is quasi-perpendicular to the straight part of the engaged cutting edge, as shown in Fig. 10(b). This is what is expected, since the engaged rounded edge is lower compared to the engaged straight part. As reported in Atlati et al. [25] and Kouadri et al. [26], the chip morphology often carries the signature of the correct behaviour chosen in the simulation. Assessed parameters are the chip width, the chip thickness and the well-known chip compression ratio.

ð10Þ

hch hch ¼ h f sin ðκ r Þ

ð11Þ

Assessed chip morphology parameters are reported in Table 7. The chip thickness (hch ) and the chip compression ratio (λh ) are overestimated by the FE model, while the chip width (w) is well predicted. Fig. 11 shows the comparison between experimental and predicted cross-section of the chip. The analysis of cutting forces from the simulation shows that the feed force and the passive force are of the same order; see in Fig. 12(a) variations with the cutting time of specific cutting forces. These two components (F f and F p ) are about one fifth the cutting force F c , as it can be deduced for Fig. 12(b). The low oscillation of the specific cutting forces during the chip formation confirms the quasi-continuous shape of formed chips (see Fig. 9). The average values of specific cutting forces are reported in Table 8. Note that cutting forces and specific cutting forces are related as follow: F ¼ K c  f  ap F c ¼ K cc  f  ap F f ¼ K f c  f  ap F p ¼ K pc  f  ap

ð12Þ

There is no a direct measurement of cutting force components. In order to compare with the predicted one and hence validates the FE model, average values of simulated cutting force components are used to calculate the specific cutting force K c . The experimental value of the specific cutting force K c is deduced from the measured cutting power P c as follow: Kc ¼

Pc f  ap  V c

ð13Þ

The comparison between experimental and numerical values of K c is reported in Table 9, where the difference is less than 6%.

Fig. 9. (a) Experimental and (b) predicted chip shapes.

67

edge

B. Haddag et al. / Tribology International 80 (2014) 58–70

ht cu Straig

Chip flow direction

tting

to straight cutting edge

Nose radius edge

Fig. 10. Simulated chip formation process where (a) contact discontinuities, and (b) chip flow direction are clearly highlighted.

3500 2

Specific cutting forces [N/mm ]

Table 7 Comparison between experimental and predicted chip morphology parameters.

Chip width, w [mm] Chip thickness, hch [mm] Chip compression ratio, λh

Experimental

Numerical

Error (%)

11 1.5 1.55

11.3 1.8 1.86

2.7 20 20

3000 2500 2000

Kc Kcc Kfc Kpc

1500 1000 500 0

0

0.005

0.01 0.015 0.02 Cutting time [s]

0.025

0.03

R a t i o s o f s p e c if i c c u t t i n g f o r c e s

1.2 1 0.8 Kc/Kc Kcc/Kc Kfc/Kc Kpc/Kc

0.6 0.4 0.2 0

0

0.005

0.01 0.015 0.02 Cutting time [s]

0.025

0.03

Fig. 12. Predicted (a) specific cutting forces and (b) ratios of specific cutting forces.

4.2. Tool wear analysis

Fig. 11. (a) Experimental and (b) predicted chip thickness and chip width.

The contact at the tool–chip interface is discontinuous and the wear is highly localized at particular zones on the cutting face, as

68

B. Haddag et al. / Tribology International 80 (2014) 58–70

reported in Section 2.4 dedicated to the tool wear characterisation. This is due to the particular geometry of the rake face (grooved face), which presents concave zones, and although to the intense thermomechanical loading between the chip and tool (high contact pressure and temperature, due to the high engaged uncut area). Also, the contact is non-uniformly distributed. The contact discontinuity induces a stress concentration in small zones, as shown in Fig. 4(b), which support the major loading, leading naturally to the wear localization.

Table 8 Average values of specific cutting forces [MPa]. K cc

Kf c

K pc

Kc

2539

512

426

2625

Table 9 Comparison between experimental and predicted values of the specific cutting force K c [MPa]. Experimental

Numerical

Error (%)

2790

2625

6

The numerical modelling is of major interest to explain the origin of tribological phenomena occurring at the contact interface, since the direct measurement of thermomechanical fields at the interface is not obvious. The developed FE-based model has allowed prediction of interface thermomechanical quantities, which are not accessible to measure using experimental means, like contract pressure, sliding velocity and temperature at the tool–workmaterial interface. Fig. 13 shows predicted fields at a certain cutting time, when the chip formation is stabilized. As shown all fields are discontinuous. At contact zones the temperature is sufficiently high to cause the adhesion of the workmaterial on the rake face, particularly in Z1 and Z2 (see Fig. 4(b)), although the insert is coated and has a good wear resistance. This confirms SEM observations of Fig. 5 and EDS analysis of Fig. 6. In order to confirm thermomechanical conditions leading to the workmaterial adhesion on the rake face, Fig. 14 presents the evolution of the temperature field beyond half the absolute melting temperature of the workmaterial (i.e. T 4 12 T m ). As stated in some research work, e.g. as reported by Qi and Mills [28], beyond about 12 T m conditions of adhesion wear may occur. In addition, as reported early by Saka et al. [29], it is emphasized that beyond about half the absolute melting temperature of materials thermal aspects at the contact interface in sliding systems become important particularly at high sliding speed. Fig. 14 shows that the first zone where the cutting temperature exceeds 1=2T m is Z1

Fig. 13. Predicted (a) temperature, (b) contact pressure, (c) sliding velocity and (d) wear depth fields at the tool rake face.

B. Haddag et al. / Tribology International 80 (2014) 58–70

69

Cutting time Fig. 14. Evolution of the temperature field beyond half the melting temperature at the tool rake face.

Fig. 15. Predicted tool wear (worn geometry) on the rake face indicating localised wear zones.

(close to the rounded cutting edge), followed by Z2 and than Z3. Indeed, as highlighted with OP technique (see Fig. 7), the amount of adhesion wear is higher in Z1, then in Z2 and lastly in Z3. The tool wear is predicted with the Usui tool wear law described in Section 3.3 (see Eq. (9)). Since the simulated cutting time is very short (few milliseconds) using the FE model, a phenomenological prediction of the tool wear is obtained (i.e. simulated wear on the initial tool geometry), so no update of the tool rake face geometry has been done during the chip formation. A quantitative prediction of the tool wear, as done for instance by Yen et al. [15], Xie et al. [16], Filice et al. [17], and Lorentzon and Järvstrat [18] for a 2D cutting configuration, and Attanasio et al. [19,20] for a 3D cutting configuration, necessitates adapted technique of remeshing of the tool rake face geometry, contact control when the rake face geometry changes, etc. A qualitative prediction of the tool wear in the simulated cutting time (about 30 ms) is sufficient to highlight wear localisation zones and the intense thermomechanical loading at the contact interface leading to observed wear types. Fig. 15 shows the worn geometry after few milliseconds of cutting time as predicted by the FE model. The qualitative prediction of the tool wear corresponds closely to that observed by SEM technique. Indeed, as shown in Fig. 16, there is a close correspondence between FE prediction and SEM observation of the tool wear at the rake face. In contact zones, high interface temperature (close to the melt temperature of the workmaterial) and contact pressure

Fig. 16. Close correspondence between FE prediction and SEM observation of localised wear zones on the rake face.

70

B. Haddag et al. / Tribology International 80 (2014) 58–70

(more than 3 GPa) are predicted, which explain the adhesion of the workmaterial on the rake face in the zone near the rounded cutting edge and abrasive wear in zones where the chip leaves the rake face. 5. Conclusions Analyses of the tribological behaviour of the tool–workmaterial interface and associated tool wear mechanisms in rough turning of a large-scale part, made of 18MND5 mild steel, with SNMM250924RH grooved coated insert have been performed in this paper. The concluding remarks are as follow: (1) Literature review shows that there is little research works dedicated to the analysis of the rough turning operation on large-scale parts. Hence an interesting original study of this operation has been proposed, with focus on the tribological behaviour and tool wear mechanisms analyses. (2) Using different characterisation techniques, it is revealed that the tool–chip interface presents contact discontinuities. This is due to the particular tool rake face geometry, designed especially to reduce the contact area and hence to facilitate the chip flow. In some contact zones, adhesion of the workmaterial is occurred, due to the combined effect of a high contact pressure and a low sliding velocity, which induce temperature rise exceeding half the absolute melting temperature of the workmaterial. The loading localisation is the result of the contact area reduction, so small contact zones on the rake face support the major loading exerted by the chip. (3) Analysis of the cutting process shows that the chip morphology is well reproduced by the FE model as well as the specific cutting force, which validated the developed model. (4) To highlight the origin of the particular tribological behaviour at the contact interface and wear localization on the rake face, the developed FE model has allowed the prediction of thermomechanical fields, not accessible by direct measurements, which confirm the intense loading and wear localisation on small contact zones. The contact features are well reproduced as observed in SEM images. (5) Finally, the FE-based approach can be used for the design of cutting tools, by studying the tribological behaviour at the tool–chip interface. The only drawback at the moment is the high computation time that prevents to achieve a parametric study within reasonable time.

Acknowledgments The authors wish to acknowledge the direction of AREVA Company for their acceptance and contribution to perform experimental tests in their manufacture site of Le Creusot (France). References [1] Diniz AE, Oliveira AJ. Optimizing the use of dry cutting in rough turning steel operations. Int J Mach Tools Manuf 2004;44(10):1061–7.

[2] Kee PK. Development of constrained optimization analyses and strategies for multi-pass rough turning operations. Int J Mach Tools Manuf 1996;36 (1):115–27. [3] Serdyuk YD, Semizhon OA, Prokopiv NM, Petasyuk GA, Kharchenko OV, Omel’chuk TV. The influence of thermal compression treatment parameters on quality characteristics and wear mechanisms of T5K10 carbide inserts in rough turning. J Superhard Mater 2011;33(2):120–8. [4] Grzesik W, Zalisz Z. Wear phenomenon in the hard steel machining using ceramic tools. Tribol Int 2008;41:802–12. [5] Ozcelik B, Kuram E, Cetin MH, Demirbas E. Experimental investigations of vegetable based cutting fluids with extreme pressure during turning of AISI304L. Tribol Int 2011;44:1864–71. [6] El Hakim MA, Abad MD, Abdelhameed MM, Shalaby MA, Veldhuis SC. Wear behaviour of some cutting tool materials in hard turning of HSS. Tribol Int 2011;44:1174–81. [7] Bhatt A, Attia H, Vargas R, Thomson V. Wear mechanisms of WC coated and uncoated tools in finish turning of Inconel 718. Tribol Int 2010;43:1113–21. [8] Lim CYH, Lim SC, Lee KS. The performance of TiN-coated high speed steel tool inserts in turning. Tribol Int 1999;32:393–8. [9] Jianxin D, Xing A. Wear behaviour and mechanisms of alumina-based ceramic tools in machining of ferrous and nonferrous alloys. Tribol Int 1997;30:807–13. [10] Ezugwu EO, Bonney J, Da Silva RB, Machado AR, Ugwoha E. High productivity rough turning of Ti–6Al–4V alloy, with flood and high-pressure cooling. Tribol Trans 2009;52:395–400. [11] Usui E, Shirakhashi T, Kitagawa T. Part 3: Analytical prediction of three dimensional cutting process. Trans ASME J Eng Ind 1978;1:33–8. [12] Usui E, Shirakashi T, Kitagawa T. Analytical prediction of cutting tool wear. Wear 1984;100(1–3):129–51. [13] Molinari A, Nouari M. Modeling of tool wear by diffusion in metal cutting. Wear 2002;252:135–49. [14] Nouari M, Molinari A. Experimental verification of a diffusion tool wear model using a 42CrMo4 steel with an uncoated cemented tungsten carbide at various cutting speeds. Wear 2005;259:1151–9. [15] Yen YC, Söhner J, Lilly B, Altan T. Estimation of tool wear in orthogonal cutting using the finite element analysis. J Mater Process Technol 2004;146:82–91. [16] Xie LJ, Schmidt J, Schmidt C, Biesinger F. 2D FEM estimate of tool wear in turning operation. Wear 2005;258:1479–90. [17] Filice L, Micari F, Settineri L, Umbrello D. Wear modelling in mild steel orthogonal cutting when using uncoated carbide tools. Wear 2007;262: 545–54. [18] Lorentzon J, Järvstrat N. Modelling tool wear in cemented-carbide machining alloy 718. Int J Mach Tools Manuf 2008;48:1072–80. [19] Attanasio A, Ceretti E, Rizzuti S, Umbrello D, Micari F. 3D finite element analysis of tool wear in machining. CIRP Ann—Manuf Technol 2008;57:61–4. [20] Attanasio A, Ceretti E, Fiorentino A, Cappellini C, Giardini C. Investigation and FEM-based simulation of tool wear in turning operations with uncoated carbide tools. Wear 2010;269:344–50. [21] Haddag B, Nouari M. Tool wear and heat transfer analyses in dry machining based on multi-steps numerical modelling and experimental validation. Wear 2013;302(1-2):1158–70. [22] Haddag B, Kagnaya T, Nouari M, Cutard T. A new heat transfer analysis in machining based on two steps of 3D finite element modelling and experimental validation. Heat Mass Transfer 2013;49(1):129–45. [23] Koné F, Czarnota C, Haddag B, Nouari M. Finite element modeling of the thermomechanical behavior of coatings under extreme contact loading in dry machining. Surf Coat Technol 2010;205:3559–66. [24] Deform™–3D Version 10.1. Scientific Forming Technologies Corporation; 2010. [25] Atlati S, Haddag B, Nouari M, Znasni M. Analysis of a new Segmentation Intensity Ratio SIR to characterize the chip segmentation process in machining ductile metals. Int J Mach Tools Manuf 2011;51:687–700. [26] Kouadri S, Necib K, Atlati S, Haddag B, Nouari M. Quantification of the chip segmentation in metal machining: application to machining the aeronautical aluminium alloy AA2024-T351 with cemented carbide tools WC–Co. Int J Mach Tools Manuf 2013;64:102–13. [27] Johnson GR, Cook WH. Fracture characteristics of three metals subjected to various strains, strain rates, temperatures and pressures. Eng Fract Mech 1985;21:31–48. [28] Qi HS, Mills B. On the formation mechanism of adherent layers on a cutting tool. Wear 1996;198:192–6. [29] Saka N, Eleiche M, Suh NP. Wear of metals at high sliding speeds. Wear 1977;44:109–25.