Author’s Accepted Manuscript

The failure mechanisms of hot forging dies

Zbigniew Gronostajski, Marcin Kaszuba, Sławomir

Polak, Maciej Zwierzchowski, Adam
Niechajowicz, Marek Hawryluk

www.elsevier.com/locate/msea

PII: S0921-5093(16)30030-2
DOI: http://dx.doi.org/10.1016/j.msea.2016.01.030
Reference: MSA33217
To appear in: Materials Science & Engineering A
Received date: 10 October 2015
Revised date: 8 January 2016
Accepted date: 9 January 2016
Cite this article as: Zbigniew Gronostajski, Marcin Kaszuba, Sławomir Polak,
Maciej Zwierzchowski, Adam Niechajowicz and Marek Hawryluk, The failure
mechanisms of hot forging dies, Materials Science & Engineering A,
http://dx.doi.org/10.1016/j.msea.2016.01.030
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
 The failure mechanisms of hot forging dies

Zbigniew Gronostajski, Marcin Kaszuba, Sławomir Polak, Maciej Zwierzchowski,

Adam Niechajowicz, Marek Hawryluk.

Wrocław University of Technology, Poland

Corresponding author email address – zbigniew.gronostajski@pwr.wroc.pl

Abstract

This paper describes the phenomena taking place on the surface of the dies used for

hot forging. Because of this paper’s limited space only changes in the tool surface layer dur-

ing the forging of a gear wheel, as most representative, are presented. Similar changes were

observed in the case of the other two investigated closed die forging processes, i.e. the forging

of a cover and a yoke, respectively. The research was aided by FEM, which supplied a lot of

information about the forging conditions. The most intensive wear of the tools occurs in the

place of their longest contact with the material being forged, regardless of the number of pro-

duced forgings. The research has shown that the one of the most adversely factor affecting the

investigated forging process is thermomechanical fatigue which results in fine cracks quickly

developing into a network of cracks extending over the entire tool/forged material contact

surface. Also the abrasive wear of the investigated die is high due to the intensive flow of the

material in the presence of abrasive oxide particles and tools bits created by thermomechani-

cal fatigue. An attempt to model the abrasive wear using the Archard model is presented.

Key words: abrasive wear; thermomechanical fatigue; wear modelling; forging.

1. Introduction

1
 Hot forging tools have a rather short life which, depends on the forging process condi-

tions, the tool design, the heat treatment (proper for the tool material), the shape of the pre-

form, the workmanship and so on. In order to improve the durability of the tools used in hot

and warm forging processes one needs to accurately identify the changes, caused by various

failure mechanisms, taking place in the surface layer of the tools in the course of their service

[1]. About 15 years ago there was much research in this field, but its results were difficult to

interpret because of the inadequate technical capabilities. However, the degradation mecha-

nisms are variously interpreted and the problem is compounded by the fact that many of the

phenomena occur simultaneously [2]. There exists a general observation that tool fatigue

cracking is critical in cold forming while excessive abrasive wear, the plastic flow of the ma-

terial and thermal fatigue are critical in hot plastic forming [3]. The most difficult situation is

in worm forming since each of the phenomena can be equally critical. In such conditions the

tools must bear very high pressures (as in cold forming) and at the same time they must be

resistant to high temperature (as in hot forming). According to Lange [4] tool life at high

forming process temperatures depends on the abrasive wear in 70% of the cases. However,

this opinion is a gross simplification since the principal failure mechanism depends mainly on

the die operating conditions and may change in the course of the process. Moreover, other

mechanisms may predominate in the particular areas of the die. It seems that despite the con-

siderable increase in research capabilities, the problems relating to the description of tool fail-

ure phenomena have been neglected in recent years [5, 6]. The intensity of the physical phe-

nomena taking place on the surfaces of hot and warm forging dies depends on the forging

process conditions, the design, manufacture and heat treatment of the tools, the shape of the

blank and the preform, etc. The large number and variety of factors having a bearing on the

forging process, and their mutual interactions make this process difficult to analyse [7]. Re-

2
search so far has indicated that as regards the die’s surface layer, the wear mechanisms listed

below are decisive.

Abrasive wear results from material loss, mainly through the separation of material parti-

cles from the surface. This process arises when there are loose or fixed abrasive particles or

harder material protrusions (acting as local microblades) in the friction areas of the interacting

elements [8].

Adhesive wear occurs on the surface where local bonds form between the friction faces

and are then destroyed as metal particles break away or as metal is smeared on the friction

faces. Adhesive wear occurs in the case of low-velocity sliding friction under high unit pres-

sures in areas of actual contact if the particles of the two friction faces are brought close

enough together so that they come within the reach of molecular forces [9].

Oxidative wear is the process of degradation of the surface layer of metal elements under

friction, due to the separation of the oxide coatings formed as a result of the adsorption of

oxygen in friction areas and the diffusion of oxygen into plastically and elastically deformed

metal microvolumes, accompanied by the formation of solid solution films. This kind of wear

occurs when the rate of formation of oxide coatings is higher than the rate of surface degrada-

tion caused by abrasion [10].

Thermomechanical fatigue wear is a kind of wear in which local loss of cohesion and the

resultant material loss are caused by material fatigue as a result of the cyclic action of contact

stress in the surface layers of paired friction elements, due to the accumulation of microstrains

in the surface layer as a result of the superposition of internal stresses and stresses generated

by external loads and thermal loads. Consequently, bits of metal break away from the base

[11, 12].

Another destructive factor, resulting in a large number of tools damaged or worn out in

forging processes, are permanent deformations of die impressions and punches. When design-

3
ing forging tools (which are particularly susceptible to plastic deformations) one must make

sure that equivalent stress σ does not exceed the material yield stress in any tool point [13].

The effects of the particular processes on the surface layer of dies and on their wear are

usually analyzed separately. There is no comprehensive detailed description of the physical

process of wear, which would take into account all the phenomena involved, whose intensity

varies depending on the forging parameters and the location on the die, which in turn deter-

mine contact time and temperature variation.

Today, computer applications such as CAD/CAM/CAE, mostly based on FEM, are used

in the design, analysis and optimization of metal forming processes, including the forging

process. Especially it is very convenient method to design new metal forming processes [14,

15]. The computer tools contribute to the development of new technologies for manufacturing

products with complicated shapes, especially for the automotive industry. A large number of

studies on the application of finite element methods to calculate abrasive wear volume, using

the model elaborated by Archard [16] have been carried out. Abachi et al. [17] determined

wear coefficient in Archard model for different points of the die surface. Kang et al. [18] pro-

posed a wear model incorporating thermal softening and used it to calculate the wear profile

of a warm forging die. Kim et al. [19] estimated, using FEM, die service life against plastic

deformation and wear during hot forging processes. Gronostajski et al. [20] proposed an

FEM-based expert system for predicting tool durability.

Considering that the research capabilities today are much greater than 20 years ago, the fi-

nite element method supplies a lot of information about process parameters, the scanning laser

system determines the quantity of tool wear in high resolution and microstructure examina-

tions are more detailed, the failure phenomena can be more thoroughly explored [21, 22, 23].

This paper presents detailed analyses of the mechanism of failure of hot forging tools, aided

by the finite element method supplying much information about, e.g., the temperature distri-

4
bution in the tools, the normal stresses on the tools, etc., which can be helpful in interpreting

the mechanism of failure of forging tools.

2. Description of investigated processes

Hot die forging processes consist of several operations. The first operation is usually open

die forging during which a preform is shaped and scale is removed. Since the tolerances of the

preform are much lower than those of the end product the wear of the tools is not so im-

portant. Generally, the tools have a simple shape without any sharp corners and the stresses

are much lower than in the next operations. Very often the tools are merely heat treated and

have no additional hardening layers. Generally, the dies are neither lubricated nor cooled in

this operation. In the next operations closed die forging takes place, which entails greater

pressures and more stringent requirements as to the forging’s dimensions, whereby it is vital

to reduce the wear of the dies by, e.g., nitriding or cooling them.

The authors carried out very extensive research into tool durability in different forging

companies. Three representative forging processes characterized by similar tool wear, i.e. the

forging of respectively a gear wheel, a cover and a yoke, were studied. Because of this paper’s

limited space mainly the changes in the tool surface layer for the bottom die insert in opera-

tion 2 of process 1, as most representative, are presented. Similar changes were observed in

the other closed die forging operations.

Forging of gear wheel

The process is carried out in three consecutive operations: 1) upsetting, 2) blocking forg-

ing and 3) finishing forging on a crank press with a press tonnage of 25 MN (Fig. 1a). The

tools used in the investigated process are made of tool steel H13 (steel 1.2344). After heat

treatment the tools for operations 2 and 3 are subjected to nitriding up to a hardness of 1100-

1200 HV. The nitrided layer is about 0.2 mm thick. The average tool life for the particular

operations is: operation 1 – about 30 thousand forging units, operation 2 – about 8 thousand

5
 units and operation 3 – 20 thousand units. The tools in operations 2 and 3 were lubricated

with a graphite suspension (4% graphite and 96% water). Photographs of the worn out tools

are shown in Fig. 1b-1d.

a)

b) c) d)
e)

Fig. 1. Forging of gear wheel: a) investigated process; worn out bottom tools after opera-

tion: b) 1- upsetting, c) 2 - blocking forging, d) 3 - finishing forging; e) forged wheel and f)

sketch of wheel.

The heating of the preform and the tools is precisely controlled. The temperature distribution

of the tools was measured by a thermovision camera and pyrometers while a sticking thermocouple

was used to measure the temperature of the billet (Fig. 2a). One should note that the measured tem-

peratures of the tools are approximate due to the fact that the emissivity coefficient was in a range of

6
0.9-0.95 and the tools were not shiny. The distribution of temperatures on the tools is presented in

Fig. 2b. The initial temperature of the billet is 1150-1170C (the forging temperature). Since the

tools in the second and third operation are lubricated with the graphite suspension their temperature

in the first, second and third operation is about 500C, 250C and 250C, respectively. The forgings

are manually carried from step to step. The time of one cycle is 14-17 seconds. The material of the

forging is steel 1.7147.

a)
b)

Fig. 2. Temperature of tools in operation 2: a) photo taken by thermovision camera and b)

temperature distribution measured by thermovision camera along line shown in Fig 2a.

3. Distribution of wear along tool cross section

Changes in the tool surface layer during the process were studied for the bottom die

insert (whose life had been found to be the shortest) in operation 2. The presented results are

the average from 3 sets of the tools.

A GOM ATOS II optical scanner was used to determine the degree of wear in the par-

ticular areas of the tools. The machine specifications are: 1 400 000 measured points per scan,

250x200x200 mm measuring areas, accuracy 0.05 mm. The scanner made it possible to scan

the active surface of the die before and after the forging process. The obtained images were

7
 compared. It was found that regardless of the number of forged units, the most intensive wear

occurred in the areas of the longest contact with the material being forged, i.e. in area 1 and

area 3 (on the flash land) (Fig. 3).

Region 2
Region 3

Region 1

Fig. 3. Distribution of wear after forging 4300 pieces.

Figure 4 shows the maximum wear of the die for the selected region (marked in Fig. 3)

in the cross section along the direction of highest wear after forging 500, 1850, 4300 and

6900 pieces.

8
Fig. 4. Maximum wear of die in selected region (marked in Fig. 5) of cross section.

4. FEM determination of die loading

In order to determine the intensity of the wear mechanisms in a forging die one must

know its working conditions, such as: hoop and radial stresses, changes in temperature, nor-

mal stresses, sliding distance, relative sliding velocity and contact time. The initial tempera-

tures of the billet and the die were measured, whereas the changes in temperature in the

course of the process were calculated by FEM. Also the hoop and radial stresses, normal

stress, the sliding distance, the relative friction velocity and the contact time were determined

by FEM.

The technical documentation of the forging process, including drawings of the tools, the press

specifications and the forging times measured for the particular operations, were used to build an

FEM model in the axially symmetric state of strain by means of the MSC.MARC software. In the

numerical model the tools were assumed to be deformable. The models built for the 1st and 2nd forg-

ing operation are schematically shown in fig. 5. The material data, i.e. thermal expansion, specific

heat, thermal conductivity, the Young modulus-temperature dependence and the Poisson ratio (as well

as the stress-strain dependence, the strain rate and temperature for the billet) were taken from the

MATILDA materials database. The stress-strain curves of the preform material were determined in

9
the upsetting test carried out using the GLEEBLE simulator. A preliminary FEM simulation showed

that the rate of strain changed from 0.1 to 100 s-1 and the temperature ranged from 600 to 1200C. In

the GLEEBLE tests the temperatures were: 750, 850, 1050, 1250C and the strain rates were: 0.1, 1,

10, 100 s-1. The dies were described using an elastic material model with an E-modulus of 2  105

MPa.

a) b)

Upper die Upper die

preform preform

bottom die bottom die

Fig. 5. Forging process models for: a) first operation and b) second operation.

In accordance with the initial temperatures measured in the real process, the following tem-

peratures were assumed in FEM: the preform – 1160C, the tools in the first operation – 500C (the

die is not lubricated) and the tools in the second and third operation – 250C. The coefficients of heat

exchange with the contact and with the environment were assumed to be 2500 and 35 W/(m2*K),

respectively. The model of the first and second operation consists of arbitrarily quadrilateral isopara-

metric elements designed for axisymmetric applications. The billet consists of 1780 and 3007 after

remeshing. The tools consist for first operation of 1680 elements upper tool and 1750 lower tool, for

second operation of 3922 elements upper tools and 3995 lower tool. A Coulomb friction model was

assumed with a coefficient of 0.35 determined by comparing the force obtained from modelling with

the real force measured on the press - 24MN.

The correctness of the modelling was confirmed by comparing the shape of the parts obtained

in FEM with the physical parts. The shapes of the forgings after the 1st and 2nd forging operation are

10
compared in fig. 6. The macrostructure in the cross section of the forging and the strain distribution in

the consecutive operations prove that the numerical model of the process reliably describes the plastic

flow of the material.

a) b) c) d)

Fig. 6. Shape of forged unit: a) macrostructure – Jacewicz test (macro etching procedure)
after 1st operation, b) distribution of plastic strains after 1st operation, c) macrostructure –
Jacewicz test after 2nd operation, d) distribution of plastic strains after 2nd operation. Chem-
ical composition of Jacewicz reagent: 38cm3HCl + + 12cm3H2S04(l,83) + 50 cm3 H20. Mate-
rial was ground. Etching at 60-70oC for 1-2h.

Since hoop and radial stresses are most critical for surface cracking therefore they

were determined (by FEM) as the first (Fig. 7).

a) b)

11
Fig. 7. Maximum stresses in tools: a) hoop stress and b) radial stress.

The maximum values of the determined stresses reach above 2000 MPa. The hoop stresses

are slightly higher than the radial ones, reaching their highest values in region 1, whereas in

region 3 they reach high values only on the very edge of the bridge. In region 2 the stresses

reach the level of 1100-1300 MPa.

Figure 8 shows the temperature distribution in the tools immediately after forging, de-

termined through modelling.

Fig. 8. Temperature of tools immediately after forging, obtained from FEM modelling.

Figure 9 shows the distribution of normal stress in the cross section, determined

through FEM modelling. The highest level of normal stress, amounting to about 1000 MPa,

occurs in region 1. In regions 2 and 3 normal stress amounts to about 600 MPa and 500 MPa,

respectively.

12
Fig. 9. Distribution of normal stress in cross section, obtained from FEM modelling.

Abrasive wear results from material loss, mainly through the separation of material

particles from the surface, and depends mainly on material hardness, normal pressure and

sliding distance. According to the Archard (1953) equation, lower hardness, higher pressure

and larger sliding distance generally result in greater abrasive wear. The sliding distance (Fig.

10) and the highest pressure (Fig. 9), determined by FEM indicate that the largest wear should

be in region 1. This does not apply to regions 2 and 3. The wear measured in region 3 is 10

times higher than in region 2 (Fig. 4) even though the normal pressure, the sliding distance

and the hardness (nitrided tools) in region 3 are similar as in region 2.

13
 Fig.10. Sliding distance in mm, determined by FEM.

5. Modelling of abrasive wear

According to the literature on the subject, the loss of die material at high temperatures

in the forming processes depends on the abrasive wear in 70% of the cases and can be mod-

elled by the Archard equation.

Consistently with the Archard model, it was assumed that the rate of wear at any moment of

forging can be written as:

(1)
̇

the wear in one cycle being:

(2)
∫

Finally, all the wear for any number C of forging cycles can be expressed by the formula:

(3)
∑∫

where: w is the value of wear [mm]), σ is normal stress [MPa], Vrel is the relative velocity of

displacement in contact [mm/s], H is the material’s hardness [MPa], and k is a dimensionless

wear coefficient.

14
 The change in real hardness measured along the surface after forging different num-

bers of pieces was taken into consideration. It is discussed and presented in Figs 21-23.

In the literature the wear coefficient is determined by comparing the analytical wear

profile and the measured profile of the worn out die. The average wear profile of the analysed

tools is presented in Fig. 11a while the wear calculated from the Archard model is presented

in Fig. 11b (the wear profile shown in Fig. 3). Since there is a large discrepancy between the

two curves coefficient k was not determined and only a dimensionless wear distribution along

the radius of die is presented.

In the case of the Archard model, wear in region 1 is the highest, whereas in regions 2

and 3 it is about 3 times lower. The distribution of normal stresses (fig. 9) and the sliding dis-

tance (fig. 10) are in good agreement with the flow of material (fig. 6), which proves that no

errors were made in the FEM calculations. Unfortunately, the calculated wear considerably

differs from the real wear.

a) b)

Fig. 11. Average wear profile of analysed tools: a) experiment and b) Archard model.

The abnormal results obtained from the Archard model can be explained by the fact

that the model does not take into account other parameters contributing to the failure mecha-

nism, such as relative sliding velocity or contact time.

15
 The relative sliding velocity determined by FEM (it is velocity between point of forg-

ing at which contact is being calculated and point of tools) is much higher in area 3 (on the

flash land) than in the rest of the die (Fig. 12) while the contact time is the longest in region 1.

1200

1000
region 1
relative sliding velocity [mm/s]

800 region 2
region 3
600

400

200

0
0.45 0.47 0.49 0.51 0.53 0.55 0.57 0.59
time [s]

Fig. 12. Relative sliding velocity.

The contact time was determined by mathematical modelling and through real process

measurements. The recorded average time of a single forging operation amounted to 1.06 s

and it was divided into contact time before deformation (the time during which the preform

stays in the impression before forging), deformation time and contact time after deformation

(the time spent by the forging in the impression before it is removed). The times for the ana-

lyzed regions are presented in table 1.

Table 1. Times for analyzed regions.

Region Contact time before defor- Deformation time Contact time after de-
mation formation
1 0.5 s 0.06 s 0.5 s
2 0s 0.05 s 0.5 s
3 0s 0.02 s 0.5 s

Considering that in region 1 the relative sliding velocity is low (Fig 12) and the contact

time is long (Table 1), one can expect intensive adhesive wear in this region under high unit

pressures (Fig. 9).

16
 During hot forging only some of the scale accumulates on the surface of the tools and on

the preform, most of it is removed. Some of the spalling scale remains on the tools, acting as

an abrasive in the next cycles of hot forging and thereby intensifies the wear of the tools. Fig-

ure 13 shows an image taken by a high-speed thermovision camera, in which a large number

of hot scale bits flaking off the tools are visible.

a) b)

Fig. 13. Image of scale bits, taken by: a) high-speed thermovision camera and b) normal
camera during operation 2.

Fig. 14. Cross section of scale collected from forge material.

Content of Fe or O [%]

Fig. 15. Oxygen and iron content in sample along line marked in Fig. 14

17
 The scale was found to contain from 45/55% to 5/95% of FeO. Thus it can be concluded that

the scale collected from the forging contains various iron oxides, including Fe2O3 and Fe3O4

(Figs 14 and 15). The nonuniform distribution of the elements in the part dealing with meas-

urements was due to the picking up of the background signal from void places by the EDX

detector. In such places the stronger signal from the heaviest elements (in this case, iron) is

picked up ( pick in the Fig.15).

a) b)

Fig. 16. Multilayered scale collected from forging material steel 1.7147.

Also the hardness of the oxide layers was measured. The hardness of the oxide layer is

greater than that of the hardened steel, amounting to approx. 550 -700HV. The size of the

various particles present in the scale ranges from 1.0 to about 50.0 m (Fig. 16 shows the

laminar structure of the scale collected from the forging). Considering the brittleness and

great hardness of the scale and its presence in the three operations of the hot forging process,

one can conclude that the scale intensifies the abrasive wear of the dies.

The presented results clearly show that the Archard model is not suitable for describ-

ing the wear of all hot forging tools. It yields good agreement with the actual wear mainly for

a simple profile. This has been reported by many scientists, but each of them obtained a dif-

ferent value of coefficient k. Sometimes the difference amounted to one order of magnitude.

This can be ascribed to the fact that different failure mechanisms occurred, which could not

18
be described by one model. This problem is particularly apparent in the case of a complicated

impression. All the mechanisms, such as thermomechanical cracking, adhesion, abrasive wear

and plastic deformation occur concurrently from the very beginning of the forging process

and they can be more or less intensive in particular conditions and places.

6. Description of changes in surface layer

After forging respectively 550, 1850, 4300 and 6900 units of gear wheel the tools were

examined. Figure 17 shows photographs of tool area 1 after the respective number of forged

units. Signs of wear are visible already after 550 forgings.

Changes in the surface layer in the selected areas were examined under a TESCAN

VEGA 3 electron scanning microscope. Already after the forging of 550 units a primary net-

work of cracks typical of thermomechanical fatigue appears in area 1 (fig. 18a), where the

hoop, redial stresses and temperature are maximum. As the number of forged units increases,

material weakening occurs that is the hardness of the surface layer decrease due to the cyclic

thermal loading of the tools, caused by the alternate heating and cooling of the die. The

change of hardness is presented in the figs 21-23. As a result, grooves develop along the fa-

tigue cracks which appeared in the initial forging cycles (fig. 18b). The grooves formed after

the forging of 1850 units run mainly radially, i.e. in the direction of the largest flow of the

material being deformed and maximum of hoop stresses . In this region the largest sliding

distance was calculated by FEM (17 mm in Fig. 10). Simultaneously a secondary network of

cracks (propagating mainly across the grooves) develops at the beginning of the process due

to the large radial stresses in the region 1 (Fig. 7)

19
a) b)

c) d)

Fig. 17. Photos of analyzed region 1 after forging: a) 550 pieces, b) 1850 pieces, c) 4300

pieces and d) 6900 pieces.

a) b)

c) d)

Fig. 18. Network of thermo-mechanical cracks in area 1 after forging: a) 550 pieces, b) 1850

pieces, c) 4300 pieces, d) 6900 pieces.

20
 The highest degree of wear (determined by means of the optical scanner), amounting to

about 0.9 mm, occurs in area 1 after forging 4300 units. But after 6900 units the value of wear

increases to 1.19 mm. Then the appearance of the surface changes, the network of primary

and secondary cracks disappears and surface irregularities resembling dunes appear. This is

probably caused by the local spalling (the adhesion processes) of the surface layer and the

weakening of the material, leading to plastic deformation and the closure of the network of

cracks (Fig. 18c). After forging 6900 pieces almost all the surface cracks disappear and nu-

merous grooves appear (Fig. 18d). After a small number of forgings (550 units) the cracks run

perpendicularly to the surface, which is evidence of plastic deformation (Fig. 19a). Small

plastic deformations resulting in the bending of the cracks in the surface layer are visible in

the cross section after forging 1850 units (Fig. 19b). This means that large plastic defor-

mations appear only after the nitrided layer is removed, which, as hardness tests and structural

examinations show, takes place after the forging of about 2000 units. The fact that the bend-

ing of the cracks increases after a greater number of units are forged indicates that the share of

plastic deformations in the wear mechanism increases with the number of forgings (fig. 19c).

a) b) c)

Fig. 19. Plastic deformation in area 1 after manufacture of: a) 550 units, b) 1850 units, c) 4300

units.

21
 Due to low-velocity sliding friction and long contact time in region 1 adhesion wear has

the form of spalling pits (Fig. 20).

Spalling
pits

Fig. 20. Spalling pits after forging 1850 pieces.

Thermomechanical cracking, adhesion wear, plastic deformation and abrasive wear

were observed in region 1, but it seems that at the beginning of the process (up to 2000 pieces

forged) the principal failure mechanism is thermomechanical cracking (adhesion, plastic de-

formation and abrasive wear intensify later). Abrasive wear is intensified by presence of abra-

sive oxide particles and tools bits created by thermomechanical fatigue.

Since changes in the surface layer of the forging tools depend on the hardness, thickness

and durability of the nitrided layer the evolution of the above parameters in region 1 was in-

vestigated. In region 1, where wear is most intensive, the hardness of the surface layer de-

creased almost everywhere from 800 HV to 400 HV after forging 1850 units (Fig. 21). Such a

decrease in hardness indicates the softening of the nitrided layer due to the cyclic thermal

loading of the tools, caused by the alternate heating and cooling of the die surface.

22
 Fig. 21. Distribution of hardness in area 1 in direction perpendicular to surface.

In region 2, where contact with the hot material is shorter than in region 1, the stability

of the nitrided layer is only slightly better (Fig. 22).

Fig. 22. Distribution of hardness in area 2 in direction perpendicular to surface.

Since the nitrided layer disappeared at a smaller number of forgings (about 1500 pcs)

the worst situation occurs in region 3 (Fig. 23).

Fig. 23. Distribution of hardness in area 3 in direction perpendicular to surface.

23
 In area 2, where the degree of wear is much lower than in areas 1 and 3, the network of

cracks after 550 cycles is less intensive (fig. 24a). After forging 1850 units a very fine sec-

ondary network of cracks develops due to much lower hoop, radial stresses (about 1200 MPa)

and temperatures (about 400oC) than in the area 1. There are no grooves characteristic of area

1, forming on the primary network, but material spalling occurs (fig. 24b). An EDS analysis

revealed that a very brittle layer of oxides forms the very fine network of cracks.

a) b)

Spalling pits

c) d)

Fig. 24. Network of thermo-mechanical cracks in region 2 after producing: a) 550 pieces, b)

1850 pieces, c) 4300 pieces, d) 6900 pieces.

In area 2 after 4300 and 6900 cycles, measurements did not show a higher degree of

wear than for the lower number of cycles. The network of cracks is similar to that for 1850

forged units. Oxides were found to be present only in the cracks (Fig. 24c). After forging

24
6900 units the pattern and size of the cracks are very similar to those after forging 4300 units

(Fig. 24d). Since no plastic deformation was observed the dominant wear mechanisms in this

region are thermomechanical cracking and some material spalling resulting from adhesion.

The network of cracks in area 3 after the forging of 550 units and 1850 units has a simi-

lar rectangular shape, due to large hoop stresses and to the radial material flow (Figs 25a and

b). No grooves typical of area 1, developing along the fatigue cracks, occurred on the bridge

after 550 and 1850 forged units. Only shallow grooves, but not running along the cracks as in

area 1, are observed there (Figs 25c and d). This is probably caused by the lower temperature

of the tools due to their shorter contact with the forged material and to uniform intensive abra-

sive wear.

a) b)

c) d)

Fig. 25. Network of thermo-mechanical cracks in region 3 after producing: a) 550 pieces, b)

1850 pieces, c) 4300 pieces, and d) change in bridge initial radius after 6900 pieces.

25
 In area 3 after 4300 and 6900 cycles, measurements showed a higher degree of wear

than after the lower number of cycles. The tool geometry changed, the flash land width and

the initial radius decreased and only remnants of the network of cracks which had developed

at the smaller number of cycles were visible (Fig. 26).

Fig. 26. Distribution of wear along cross section of region 3 of die after producing: a)

550 pieces, b) 1850 pieces, c) 4300 pieces and d) 6900 pieces.

Almost no change in the distribution of wear was found from 550 to 1850 pieces,

while a great change has taken place from 1850 to 4300 pieces was observed. It could be

explained by intensity of wear changes in the course of the forging process, which is

presented in the Fig. 27 expressed by the volume loss determined from the changes in

the die volume in the course of the forging process. The changes were determined on the

basis of forging scans taken at every 1000 pieces.

26
 Fig. 27. Distribution of die wear (changes in volume) as function of number of forged
pieces.

An interesting observation emerges from the Fig. 27. The intensity of wear of the ana-

lysed die is small at the beginning of its use, markedly increases after forging about 2000

pieces and after forging about 6000 pieces the wear intensity decreases. In the publica-

tion which they are preparing the authors explain this by the fact that at the beginning of

the process (after a small number pieces have been forged) the degree of wear is low

owing to the existence of the hard nitrided layer. Once the latter is removed, the intensity

of wear increases. Since wear leads to the loss of tool material, whereby the die impres-

sion becomes larger, the pressures towards the end of the life of the tools (after forging a

large number of pieces) are much lower than at the beginning of the forging process and

so the wear intensity decreases. The authors are preparing a publication which on the

basis of real wear measurements and mathematical modelling confirms this hypothesis.

In the region 3 plastic deformation (the bending of cracks) could be observed even at an

early stage in the process (after 550 units Fig. 28), but only in a thin outer layer (50

mµ), under which the cracks were perpendicular to the surface.

27
 Fig. 28. Plastic deformation (bending of cracks) in region 3 after forging 550 pieces.

In spite of the similar normal stress, sliding distance and temperature in area 3 and 2, a

much larger material loss is observed in area 3. This can be ascribed to abrasive wear intensi-

fied by broken scale, tools bits created by thermomechanical fatigue which are moving from

canter to outside of die and the higher sliding velocity. The formation of grooves intensifies in

the presence of hard oxides which constitute an abrasive (a product of high-temperature oxi-

dation).

Conclusion

Thermomechanical fatigue of the tools, quickly resulting in a network of fine cracks, has

been found to be the most adverse factor for the investigated hot forging tools. The further

propagation of the cracks depends on the forging process parameters, the interaction between

the die and the forging and the rate of flow of the material, and it usually leads to the devel-

opment of a secondary network of cracks on the entire contact surface. Bits of scale flaking

off the forging and the tools clearly intensify abrasive wear, which leads to the formation of

grooves running along the direction of material flow within the network of primary cracks.

This is usually followed by the spalling of the secondary network of cracks, resulting in fur-

28
ther material weakening, the formation of a characteristic undulated surface and the plastic

deformation of the surface layer, and ultimately in the closure of the cracks.

FEM supplies a lot of information, such as stress distribution, temperature changes, nor-

mal stresses, sliding distance, relative sliding velocity and contact time, which can be helpful

in interpreting the failure mechanism of forging tools.

Three regions of wear were distinguished in the analyzed tools. Different dominant failure

mechanisms were observed in each of the regions. In region 1, where the highest temperature,

normal stresses, long contact time and low sliding velocity occur, the failure mechanisms are

thermomechanical cracking, adhesion wear, plastic deformation, abrasive wear, but at the

beginning of the process (up to 2000 pieces forged) the principal failure mechanism is ther-

momechanical cracking (adhesion and plastic deformation intensify later). In region 3 abra-

sive wear occurs due to the highest sliding velocity and the large amount of broken scale and

tools bits created by thermomechanical fatigue which are moving from canter to outside of

die. In region 2 where there are no such extreme conditions as in regions 3 and 1, only con-

siderable thermomechanical cracking without large material loss was observed.

In the light of the presented research results, the generally accepted view that abrasive

wear is the dominant mechanism in hot forging is highly debatable. It was probably adopted

in order to simplify the complex process of wear and to use the Archard model to describe

almost every case of wear. As a result, the model is often heavily modified. The coefficients

used in such models are often matched by comparing reality with the model and have no

physical basis whatsoever.

All the failure mechanisms, such as thermomechanical cracking, abrasive and adhesion

wear and plastic deformation, occur simultaneously from the very beginning of the process

and in given conditions they can be less or more intensive.

Acknowledgments

29
 This research was carried out as part of National Science Centre project NCN

2011/01/B/STB/02056.

References

1. T. Altan, T, Cold and Hot Forging Fundamentals and, Application, ASM Internation-

al, Ohio, 2005.

2. E. Summerville, K. Venkatesan, C. Subramanian, Wear processes in hot forging press

tools, Mater. Design. 16 (1995) 289-294.

3. S. Babu, D. Ribeiro, R. Shivpuri, Material and Surface Engineering For Precision

Forging Dies, The Ohio State University1999.

4. K. Lange, L. Cser, M. Geiger, J.A.G, Kals, Tool life and tool quality in bulk metal

forming, Proc. Inst. Mech. Eng. 207 (1993) 223-239.

5. B.A. Behrens, Finite element analysis of die wear in hot forging processes, CIRP Ann-

Manuf. Techn. 57 (2008) 305-308.

6. E. Feyzullahoğlu, S. Karabay, Processing damages of material components of aerial

conductors and their tribological behaviors under dry friction, Arch. Civ. Mech.

Eng.14 (2014) 682-690.

7. R. Ebara, K. Kubota, Failure analysis of hot forging dies for automotive components,

Eng. Fail. Anal. 15 (2008) 881-893.

8. O. Barrau, C. Boher, R. Gras, F. Rezai-Aria, Analysis of the friction and wear behav-

iour of hot work tool steel for forging, Wear. 255 (2003) 1444-1454.

9. Xu, Wujiao, Li, Wuhua, Wang Yusong, Experimental and theoretical analysis of wear

mechanism in hot-forging die and optimal design of die geometry, Wear. 318 (2014)

78-88.

30
10. Z. Gronostajski, M. Hawryluk, M. Kaszuba, M. Marciniak, A. Niechajowicz, S. Polak,

M. Zwierzchowski, A. Adrian, B. Mrzygłód, J. Durak, The expert system supporting

the assessment of the durability of forging tools, Int. J. Adv. Manuf. Tech. DOI

10.1007/s00170-015-7522-3, 2015.

11. K. Gargul, B. Boryczko, Removal of zinc from dusts and sludges from basic oxygen

furnaces in the process of ammoniacal leaching, Arch. Civ. Mech. Eng. 15 (2015)

179-187.

12. Deepak Kundalkara, Mukund Mavalankara, Asim Tewarib, Effect of gas nitriding on

the thermal fatigue behavior of martensitic chromium hot-work tool steel, Mater. Sci.

Eng. A. 651 (2016), 391–398.

13. Ch. Choi, A. Groseclose, T. Altan, Estimation of plastic deformation and abrasive

wear in warm forging dies, J. Mater. Process. Tech. 212 (2012) 1742-1752.

14. Z. Pater, A. Gontarz, J. Tomczak, T. Bulzak,Producing hollow drive shafts by rotary

compression, Arch. Civ. Mech. Eng. 15 (2015) 917-924.

15. D. Pociecha, B. Boryczko, J. Osika, M. Mroczkowski, Analysis of tube deformation

process in a new pilger cold rolling process, Arch. Civ. Mech. Eng. 14 (2014) 376-

382.

16. J.F. Archard. Contact and rubbing at surfaces, J. Appl. Phys. 24 (1953) 981-988.

17. Abachi, Siamak, Akkok, Metin, Gokler Mustafa, Ilhan, Wear analysis of hot forging

dies, Tribol. Int. 43 (2010) 467-473.

18. JH. Kang, IW. Park, JS., Jae, SS. Kang, A study on a die wear model considering

thermal softening (I):construction of the wear model, J. Mater. Process. Tech. 96

(1999) 53-58.

31
19. Z. Gronostajski, M. Kaszuba, M. Hawryluk, M. Zwierzchowski, A review of the deg-

radation mechanisms of the hot forging tools, Arch. Civ. Mech. Eng. 14 (2014) 528-

539.

20. D.H. Kim, H.C. Lee, B.M. Kim, K.H. Kim, Estimation of die service life against plas-

tic deformation and wear during hot forging processes, J. Mater. Process. Tech. 166

(2005) 372-380.

21. Gronostajski Z., Hawryluk M., Krawczyk J. , Marciniak M., Numerical Modelling of

the thermal fatigue of steel WCLV used for hot forging dies, Eksploat. Niezawod. 15

(2013) 129-133.

22. G. Winiarski, A. Gontarz, Z. Pater, A new process for the forming of a triangular

flange in hollow shafts from Ti6Al4V alloy, Arch. Civ. Mech. Eng. 15 (2015) 911-916

23. Ryo Matsumoto, Masaaki Otsu, Michiaki Yamasaki, Tsuyoshi Mayama, Hiroshi

Utsunomiya, Yoshihito Kawamura, Application of mixture rule to finite element anal-

ysis for forging of cast Mg–Zn–Y alloys with long period stacking ordered structure,

Mater. Sci. Eng. A, 548 (2012) 75-82.

32