Numerical chip formation analysis during high-pressure cooling in metal machining [original]

Numerical chip formation analysis during high-pressure cooling in

metal machining

Eckart Uhlmann

a,b

, Enrico Barth

a,*

, Benjamin Bock-Marbach

, J¨

org Kuhnert

Technische Universit¨

at Berlin, Institute for Machine Tools and Factory Management (IWF), Chair of Machine Tools and Manufacturing Technology, Pascalstr. 8–9,

10587 Berlin, Germany

Fraunhofer Institute for Production Systems and Design Technology IPK, Pascalstr. 8–9, 10587 Berlin, Germany

Fraunhofer Institute for Industrial Mathematics ITWM, Fraunhofer-Platz 1, 67663 Kaiserslautern, Germany

ARTICLE INFO

Keywords:

Turning

Simulation

High-pressure cooling

ABSTRACT

Most metal turning processes utilize cutting fluids. Despite extensive experimental and analytical studies, the

mechanisms of chip formation under consideration of a cutting fluid are still not entirely understood. Due to

fluid-structure interaction, simulating wet cutting processes for an extended duration has not been feasible. The

primary objective of this study is to utilize a simulation approach to provide additional information about the

wet chip formation process in contrast to measurement methods, with a view to drawing conclusions. As

methodology the Finite-Pointset-Method (FPM) is employed to simulate the chip formation process for dry, flood

and specifically high-pressure cooling conditions during machining of carbon steel C45 as well as nickel-based

alloy Inconel 718. Due to the increased relative velocity between workpiece and cutting fluid with the use of

high-pressure cooling compared to flood cooling, numerical stability issues are present. Initially, the modeling

approach to handle high-pressure cooling conditions is described and validated by an impact test. Subsequently

the cutting simulation model is presented in detail and verified by measurements. The simulation results of stress,

temperature and plastic strain rate fields are used to elucidate the observed discrepancies between various

cutting fluid strategies in detail. These findings suggest explanations for the high efficiency of high-pressure

cooling such as a decline of hydrostatic stresses or activation of ductile damaging.

1. Introduction

1.1. Cutting fluid mechanisms of action

The role of cutting fluids in cutting processes is often unpredictable,

partially ambivalent and depends on various process parameters. For

instance, Stanford et al. [1] identified reduced process forces, tool wear

and process temperatures for cutting with flood cooling compared to dry

cutting during machining of steel. Conversely, Devillez et al. [2]

observed reduced cutting forces and enhanced surface qualities using

dry cutting rather than flood cooling. This complex behavior is caused

by diverse mechanisms of action and their superposition which are

currently the subject of scientific discussions [3].

Dynamic pressure, cooling and lubrication are generally accepted as

the primary mechanisms of action. The dynamic pressure p

of a cutting

fluid acts mechanically on the chip surface resulting in chip bending.

Denkena et al. [4] have analytically analyzed chip bending using a

bending beam model for the chip. The results indicate that a significant

chip deflection is achieved for low cutting fluid pressure of p

=10 bar.

Furthermore, the dynamic pressure can serve as chip breaker as

analytically calculated by Astakhov [5]. This specific effect has been

documented in various studies [6].

Due to friction and plastic deformation, heat is generated in the chip

formation area and is removed by the cutting fluid. Tanveer et al. [7]

have demonstrated that the tool temperatures can be reduced from T =

1200 ◦C to T =650 ◦C by a spray cooling compared to dry machining.

This mechanism of action is often mathematically described by the heat

transfer coefficient (HTC)

. However, the heat transfer depends on the

specific fluid flow parameters in combination with the solid material,

and the determination of the HTC

is complex. Consequently, a wide

range of literature values exist. Kops and Arenson [8] have determined a

value of

=2.5 kW/(m

K) using down cooling experiments. Luchesi

and Coelho [9] have obtained the HTC within the same range at

=3.8

kW/(m

K). In contrast, Gerstenberger [10] and Liu [11] have identified

* Corresponding author.

E-mail address: [email protected] (E. Barth).

Contents lists available at ScienceDirect

CIRP Journal of Manufacturing Science and Technology

journal homepage: www.elsevier.com/locate/cirpj

https://doi.org/10.1016/j.cirpj.2024.07.003

Received 21 March 2024; Received in revised form 18 June 2024; Accepted 13 July 2024

CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

Available online 30 July 2024

a significantly higher HTC partly exceeding

>100 kW/(m

K). Based

on the Nusselt number, Courbon et al. [12] even calculate an

>848

kW/(m

K).

The role of lubrication is also controversially discussed in literature.

Due to the high contact pressures, temperatures and sliding velocities,

penetration of a hydrodynamic film into the chip contact zone is

commonly doubted in research discourse. Analytical calculations based

on capillary effects suggest that penetration only occurs at low cutting

speeds [13]. Tribological changes during wet cutting are often attrib-

uted to the development of a chemical boundary layer or the Rehbinder

effect [3,14]. However, friction and particularly lubrication analysis

during the cutting process are complex. Consequently, analogical ex-

periments are frequently used to simplify the analysis and interpretation

of lubrication. In the study by Lakner and Hard [15], a tool is pressed

against a rotating shaft that suppresses the formation of chips which is

used to determine friction coefficients for dry and high-pressure wet

conditions. It has been determined that the effect of lubrication depends

on the workpiece material. In the case of steel 42CrMo4+QT, the

determined friction coefficients under dry conditions are reduced

compared to wet conditions. As explanation, the authors suspect thermal

material softening. In the literature, various experimental friction

principles are presented and utilized. A comparison between analogical

experiments using translational relative and rotatory relative motion has

revealed that the determined friction coefficients are similar for dry

conditions but differ for wet conditions [16]. Compared to translational

relative motion the use of a cutting fluid has been identified to distinc-

tively reduce the friction coefficient for rotatory relative motion. This

observed discrepancy is attributed by the authors to differences in the

causes of lubrication, such as the formation of a hydrodynamic film or

tribo-chemical reaction products. However, these examples illustrate

the complexity of friction and lubrication, even in simplified analogical

experiments.

1.2. Cutting fluid applications

Despite the fundamental gap in understanding of the mechanisms of

action, cutting fluids are widely used in metal cutting processes. Esti-

mations indicate that approximately 85 % of metal cutting processes are

conducted with cutting fluids [17]. According to Benedicto et al. [18],

the total consumption of cutting fluids is estimated at approximately m

=39.4 ⋅10

t for the year 2015. In contrast to existing guidelines for

sustainable production, the usage of cutting fluids results in environ-

mental and economical deficits. It is assumed that the cutting fluid

related costs of a product are between 2 % and 18 % [17,19]. In order to

enhance the traditional flood cooling application, a variety of other

cutting fluid strategies such as high-pressure cooling, minimum-quantity

lubrication or cryogenic cooling are presented in the literature. In a

review study, Bartolomeis et al. [20] have evaluated these strategies for

machining of Inconel 718 and categorized the impact on various aspects

of interest. The authors conclude that no strategy can meet high re-

quirements for sustainability, productivity, and surface integrity

simultaneously.

This paper focusses on the high-pressure cooling strategy. Pigott and

Colwell [21] conducted an initial analysis of this cutting fluid strategy.

Since the early 1950s, the application has been refined and numerous

experiments have been documented. In contemporary practice, the

cutting fluid is directed to specific regions of the chip formation area

with cutting fluid pressures often utilized above p

=80 bar. The use of

high-pressure cooling has been repeatedly demonstrated to result in

improved tool life compared to flood cooling and dry machining [20,22,

23]. Furthermore, high cutting fluid pressures can effectively act as a

chip breaker [6,24]. The high-pressure cooling strategy is frequently

employed for difficult-to-cut materials. For face turning of Inconel 718,

Su´

areza et al. [25] have investigated the effects of high-pressure cooling

compared to conventional cooling. Their findings indicate that

high-pressure cooling results in reduced process forces, contact areas,

and process temperatures. Additionally, the dominant wear has changed

from flank wear during conventional cooling to notch wear during

high-pressure cooling due to higher temperature gradients along the

cutting edge. Ezugwu and Bonney [26] also have analyzed wet cutting of

Inconel 718 and have identified a 7-fold improvement in tool life for

high-pressure cooling compared to flood cooling. The authors observe

reductions in cutting force and contact length between cutting tool and

workpiece material, suggesting that penetration of cutting fluid into the

tool-chip interface provides efficient cooling and lubrication. In addi-

tion, a critical cutting fluid pressure limit has been identified beyond

which a significant positive improvement cannot be observed. Similarly,

Ayed et al. [22] have identified a critical cutting fluid pressure in the

context of titanium machining. Results from experimental studies indi-

cate that the utilization of multiple cutting fluid jets exerts a beneficial

impact. Tamil Alagan et al. [27] investigate the influence of

high-pressure cooling jets directed at both the rake and flank faces

during the turning of Inconel 718. Their findings indicate that the

additional high-pressure flank cooling has a considerable impact on

enhancing tool life. Another factor influencing the effectiveness of

high-pressure cooling has been demonstrated to be the cutting fluid it-

self. In a research study, three cutting fluid grades have been analyzed

for machining of Ti-6Al-4V [28]. In contrast to flood cooling, the cutting

fluid grades have a significant effect on tool performance. The authors

explain this phenomenon by the limited access to the chip-tool interface

during flood cooling, which consequently increases the relevance of

properties such as surface tension for each grade during high-pressure

cooling. The potential to act as a chip breaker has been observed in all

cutting fluid grades. In contrast to produced continuous chips using

flood cooling, the chips are broken. However, the typical C-shaped chip

morphologies differ between the cutting fluid grades which is attributed

to oxidations and flow disturbances. Additionally, the use of

high-pressure cooling has been shown to enhance steel machining pro-

ductivity. In a study conducted by Naves et al. [29], the stainless steel

AISI 316 has been machined under dry, flood cooling and high-pressure

cooling conditions. The results indicate that high-pressure cooling

significantly reduced adhered material which is attributed to improved

penetration of the cutting fluid into the chip-tool interface. Furthermore,

the study has demonstrated that increasing the cutting fluid concen-

tration has a positive effect on tool performance. Nevertheless, also

negative mechanisms for tools have been reported. A recent study by

Tamil Alagan et al. [30] indicates that cavitation wear occurs on the

flank surface of the tool when high cutting fluid pressures are used. Due

to specific flow conditions in combination with process temperatures,

cutting fluid vapor bubbles can form and damage the tool when

collapsing. To circumvent this wear mechanism, the authors suggest the

control of the process temperature below the Leidenfrost point or the

control of the fluid pressure above the vapor pressure. However, in

addition to the widely positive effects on productivity, negative impacts

on energy consumption and environmental issues are associated to

high-pressure cooling [20]. Nevertheless, the sustainability assessment

is subject to current scientific research. For instance, Cica and Kramar

[31] state that high-pressure cooling offers superior machineability as

well as economic, and environmental performance in comparison to

flood cooling.

The apparent benefits of high-pressure cooling are the subject to a

number of presumed explanations. Furthermore, a variety of factors

influencing the effectiveness of high-pressure cooling, such as the crit-

ical cutting fluid pressure, need to be taken into account regarding

sustainability aspects. Due to these challenges associated with cutting

fluids, in particular with high-pressure cooling, advanced methods to

adjust the cutting fluid strategies and improve the general understand-

ing are necessary.

1.3. Simulation of wet cutting

Simulations of dry chip formation support process analyses since

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

104

several decades. Various simulation methods as well as different soft-

ware packages are utilized, having differences in modeling character-

istics and qualities. For dry cutting the mesh-based Finite-Element-

Method (FEM) in different formulations is commonly used [32,33]. By

the Lagrangian formulation the FEM mesh is fixed with the material and

consequently element distortions occur during large deformations such

as chip formation during machining. To overcome this, adaptive

remeshing is frequently used [33]. Another opportunity is the usage of

the Arbitrary Lagrangian-Eulerian (ALE) formulation. Thereby, the

motion of the FEM mesh is independent of material deformation.

However, challenges in definition of the mesh motion scheme related to

a sufficient numerical discretization are present [33].¨

Ozel et al. [34]

have conducted a comparative study between simulations based on ALE

using software Abaqus and simulations based on Lagrangian formula-

tion with remeshing using software Deform-3D to assess the cutting of

Inconel 718. The authors have concluded that both frameworks are

suitable to model the dry cutting process. In this study, the most accurate

results in predicting the cutting force have been detected using the

Lagrangian formulation. Observed strains, stresses and temperatures

have been identified to be comparable between the two frameworks.

Nevertheless, it is repandly reported that the ALE shows restrictions in

predicting chip morphologies [35–37]. This issue can be addressed by

the use of the Coupled Eulerian-Lagrangian (CEL) formulation [38].

Mesh-less simulation methods offer an additional opportunity to model

large deformations during chip formation [32]. The

Finite-Pointset-Method (FPM), which discretizes the domain by a cloud

of numerical points represents a suitable simulation environment. These

points move in a Lagrangian formulation with the material flow but can

be adaptively managed during the simulation. In one of the first com-

parisons between FPM and FEM, comparable simulated cutting forces

and contact temperatures have been determined [39]. Differences be-

tween the two simulation methods have been identified in the prediction

of the shear angle variations relative to the rake angle. Furthermore,

FPM has demonstrated the capacity to predict both continuous and

segmented chips [39]. Grimm et al. [40] recently compare eight

different simulation frameworks such as FEM in different formulations

using software Abaqus, AdvantEdge, Marc and Deform-2D as well as

FPM using software MESHFREE for the same cutting application. As far

as possible, in all simulations the same model set up is used. All

frameworks suitably predict the cutting force F

and underestimate the

feed force F

. Significant differences in the simulated temperature fields

between the eight frameworks and even between different FEM software

have been found. These are partly explained by differences in chip for-

mation. Further differences in describing geometrical quantities have

been demonstrated. In comparison to other simulations, the FPM ex-

hibits advantages in the description of the contact length while showing

minor inaccuracies in the representation of chip thickness.

Adding a cutting fluid to the chip formation process poses challenges

to numerical modelling due to the presence of solid and fluid phases.

Some specific challenges include the different time step sizes due to

varying densities or the application of dissimilar numerical methods for

solving the related differential equations. Moreover, the large de-

formations that occur during the formation of the chip result in a vari-

ation of the domains of the solid and fluid phases which presents an

overall challenge to numerical discretization techniques. To overcome

this, various approaches have been presented so far. One such approach

is the simplified method proposed by Courbon et al. [12]. The dry chip

formation has been simulated with a pressure as well as a HTC added as

boundary conditions. The results indicate that pressure has a significant

effect on chip curvature while cutting force and contact length are also

influenced by cooling. However, this approach requires approximating

mechanisms of action and disregards any interaction between chip

formation and cutting fluid.

An idea to couple chip formation and a cutting fluid flow is to

simulate both phases with separate methods. Relevant information, such

as velocity and stress fields, needs to be transferred between both

methods and in order to be used as boundary conditions at the chip-fluid

interface. However, the organization of data transfer is complex due to

different time step size and numerical discretization. Therefore, in most

cases, only a one-way coupling has been presented so far. Oezkaya and

Biermann [41] have transferred a simulated dry chip to a fluid simula-

tion for a drilling process. This model can analyze the cutting fluid flow

while considering the chip shape. Nevertheless, it does not take into

account the back-coupling of fluid flow to chip formation. Similarly,

Helmig et al. [42] also have transferred a simulated dry chip to a flow

simulation and demonstrate the resulting temperature reduction in the

tool. However, the impact of cooling on chip formation is neglected

again.

Another possibility for solving the fluid-structure interaction during

wet chip formation is to use a monolithic approach where both fluid and

solid are solved within the same method and software. The CEL

formulation offers a possibility for handling fluid and solid phases. A

first semi-2D model for wet chip formation based on the CEL formulation

is presented by Klocke et al. [43]. The model highlights the functionality

of a cutting fluid as a chip breaker. Liu et al. [11] utilize the CEL

formulation to analyze dry and wet chip formation processes. Among

other things, stagnation points of the cutting fluid are identified which

act as a barrier and reduce the mechanical impact of the cutting fluid.

The FPM also permits the integration of fluid and solid simulation within

the same method. Due to the mesh-less approach, variations in the do-

mains of the solid and fluid phase are automatically captured by an

adaptive management of the numerical points. Furthermore, the method

has been specifically developed to accommodate intricate

fluid-structure interactions and as a consequence, a variety of mathe-

matical algorithms are involved in order to handle the challenges

required for wet cutting operations. Uhlmann et al. [44,45] present

models for wet chip formation based on the FPM considering flood

cooling. The models can predict the wet chip formation process and

provide initial numerical insights such as changes in temperature fields

when a cutting fluid is considered. Thus far, high-pressure cooling has

not been applied in the models.

1.4. Motivation and procedure of research

Cutting fluids are of great importance for metal cutting processes.

Different cutting fluid strategies are applied to improve the performance

of cutting fluids. However, the mechanisms of action of cutting fluids are

not fully understood and the design of wet cutting processes is often

based on experience and experiment. To improve the general under-

standing, in this paper a wet cutting simulation model for high-pressure

cooling based on the FPM is presented and utilized for detailed analyses

of wet chip formation process as the primary objective. Consideration of

high cutting fluid velocities v

in a cutting simulation model places

specific demands on the fluid-structure interaction. Therefore, in a first

step, an implicit modeling approach is validated by experiments. Af-

terwards orthogonal cutting experiments are described. Cutting process

parameters have been selected in order to investigate the varying in-

fluence of cutting fluid on chip formation. Different cutting fluid pres-

sure stages are analyzed to vary the impact of the dynamic pressure on

chip formation. The different workpiece materials hot formed C45

(1.0503) without a heat treatment and Inconel 718 (2.4668) in heat

treatment condition of solution annealed are considered to vary the heat

conduction within the chip and the mechanical strength. Subsequently,

the developed wet cutting simulation model is presented in detail. The

quality of the model is extensively analyzed by measured process pa-

rameters. After the verification, the model offers further numerical in-

sights into the process, such as temperature distribution. A discussion of

the observed findings concludes the study.

2. Fluid-structure interaction for high-pressure cooling

To evaluate the modelling of pressure interaction between solid and

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

105

fluid, an impact test is utilized and simulated. Therefore, a dynamometer

of type 9257B from Kistler Holding AG, Winterthur, Switzerland, is

installed in a turning lathe TNX 65 from TRAUB Drehmaschinen GmbH

&Co.KG, Reichenbach, Germany. In order to provide a flat surface, a

plate made of plexiglass is fixed on the dynamometer. A jet of cutting

fluid is then directed vertically onto the plate. Fig. 1a) illustrates the

experimental setup. The distance between nozzle outlet and flat surface

is l =10 mm. Flood cooling and high-pressure cooling supply systems

are analyzed related to the cutting processes. For flood cooling, a nozzle

with a diameter of d =6.5 mm in combination with a cutting fluid ve-

locity of v

cf,FC

=7.7 m/s is conducted. For high-pressure cooling, a

nozzle of diameter d =1.2 mm is used. Two high-pressure stages HP1

and HP2 are considered in order to analyze the cutting fluid velocities of

cf,HP1

=30 m/s and v

cf,HP2

=68 m/s. Adrana AY 401 from Houghton

Deutschland GmbH, Dortmund, Germany, is used as the oil component

in an emulsion with a water concentration of c

H2O

=92 % and oil

concentration of c

oil

=8 %. All experiments are repeated three times.

The transfer of dynamic pressure from the jet into the plate is based

on the impulse deflection. To verify this effect in the simulation, Fig. 1b)

shows the related model in a y-z-section view. The impact plate is

simplified as one body made out of steel and modeled with reduced

dimensions around the impact area to save computation time. To ensure

comparability with the cutting simulation model, identical mechanical

models are used for elastic and plastic deformation. At the bottom and

side surfaces the impact plate is clamped. The cutting fluid flows verti-

cally out of the nozzle with the specified cutting fluid velocities. Material

properties of the used cutting fluid were previously measured in a study

[46]. At the interface between fluid and solid, the implicit

surface-surface coupling method is used. This includes a conventional

wall-boundary condition within the velocities and stress tensors are

automatically transferred between both phases by the used software

MESHFREE from Fraunhofer ITWM, Kaiserslautern, Germany. In pre-

vious cutting simulation models, the fluid-structure interaction was

explicitly modeled. Explicit and implicit treatment of boundary condi-

tions between interacting phases are significantly different within the

FPM framework. For explicit boundary conditions, the physical quan-

tities of a phase at the interacting surface at time t =n+1 are calcu-

lated based on the physical quantities of the opposite phase at time

t=n. After solving the associated system of equations, the user-defined

boundary conditions are proven for consistency regarding a penetration

of both phases for example. This is corrected by the solver using an

artificial correction velocity and corresponding pressure adjustment. In

contrast, for implicit boundary condition, the physical quantities at the

time t =n+1 take into account the physical quantities of the opposite

phase at the time t =n+1. This implies that the boundary condition is

incorporated into in the system of equations to be solved. Consequently,

an artificial correction is not required, but the mathematical consistency

of the system of equations can be compromised. Mathematical details

can be found in the user manual of the software [47]. However, for

high-pressure stages the explicit approach leads to numerical in-

stabilities, requiring the described numerical adjustments. As an addi-

tional stabilization technology, the simulated cutting fluid velocity is

constrained at v

cf,max

=3.5 ⋅v

, which serves to suppress numerical

instabilities.

In Fig. 2a) the distribution of transferred mechanical stresses

into

the impact plate are exemplarily shown. Around the jet impact area, the

cutting fluid and impact plate are numerically discretized with h

G,min

=1.8 mm for fluid velocity of v

cf,FC

=7.7 m/s and with h

G,min

=0.18 mm for both high-pressure stages. The minimum smoothing

length h

G,min

is the major parameter to describe the numerical resolution

of a FPM model [48]. Variations in the numerical discretization are

based on the different fluid velocities. The resulting reaction forces F

are evaluated at the clamped surfaces of the impact plate. Assuming a

total deflection of the jet at the impact plate, the reaction forces can be

analytically calculated by F

⋅

/4 ⋅d

⋅ Δv

, where

represents

density. Fig. 2b) shows the measured, analytically calculated and

simulated reaction forces. All simulated forces correspond to the

analytical values with a maximum deviation below 3 %. For flood

cooling, the measured forces also match the analytical values. In case of

high-pressure cooling, the measured forces are increased compared to

simulation and analytical results. In case of high-pressure stages HP2

with v

cf,HP2

=68 m/s, a difference of 9 % is detected between experi-

mental and analytical values. This discrepancy is presumed in mea-

surement inaccuracy of the corresponding high flow rate of the cutting

fluid resulting in a deviation in calculated cutting fluid velocities.

Compared to the analytical values, the transfer of the dynamic pressure

from a fluid into the solid is correctly simulated and the modeling

approach is evaluated as valid. However, beside the dynamic pressure,

the transfer of shear stresses at the wall is not analyzed with this setup

and remains uncertain.

3. Orthogonal cutting experiments

To validate cutting simulations, orthogonal cutting experiments are

conducted at turning lathe TNX 65. Slotted workpieces resulting in a

width of cut of b =1 mm made of C45 and Inconel 718 are processed.

Uncoated carbides of type H13A from Sandvik, Sandviken, Sweden, in

the ISO-geometry SPUN 120312 are used to simplify the simulation

setup. This workpiece tool combinations have been used in several

studies [49–51]. In combination of the insert with the tool holder, this

setup results in a rake angle of γ

=0◦and clearance angle of

=11◦as

well as a symmetry plane in the direction of width of cut. The cutting

edge radius is approximately r

=5µm. Cutting velocities of v

=40, 70

and 100 m/min are carried out, with lower velocities chosen to increase

the impact of the cutting fluid as it has been shown in literature. A

constant feed of f =0.05 mm is used for all experiments.

In accordance with the previous impact test experiment, an identical

emulsion with an oil concentration of c

oil

=8 % is utilized. Two high-

pressure stages as well as flood cooling and dry conditions are consid-

ered. To enable high-pressure stages, the turning lathe is extended by an

additional fluid cycle using the pump LMP1000GWK219O02CI from

SKF Lubrication Systems Germany GmbH, Berlin, Germany. In addition,

a self-developed cutting fluid supply system for high-pressure cooling is

required to facilitate the orthogonal turning process, taking into account

Fig. 1. Impact test; a) photo during experiment using high-pressure supply system; b) sketch of related model.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

106

a symmetry plane, see Fig. 3. The cutting fluid flows out of a nozzle with

a diameter d =1.2 mm towards the overhead direction, i. e. the root of

the chip. The two high-pressure stages HP1 and HP2 are adjusted by

varying the power of the pump according to 28 % and 85 %. Prior to

cutting fluid supply system, the cutting fluid pressure is measured at

HP1

=14 bar and p

HP2

=87 bar. This settings result in the cutting fluid

velocities of v

cf,HP1

=30 m/s and v

cf,HP2

=68 m/s and the corre-

sponding flow rates of ˙

cf,HP1

=2.1 l/min and ˙

cf,HP2

=4.6 l/min. In

comparison to the works by Su´

arez et al. [25] and Ezugwu et al. [28], the

cutting fluid velocities are approximately halved. The utilized cutting

fluid supply system for high-pressure cooling has been adjusted in order

to result in a simple free fluid jet. Therefore, the free jet breakup regime

has been evaluated in accordance with the Ohnesorge principle [52]. It

is not anticipated that atomization disintegration occurs, as this phe-

nomenon is not yet possible to consider in simulations. For flood cooling,

the cutting fluid flows out of a nozzle of diameter d =6.5 mm with the

fluid velocity of v

cf,FC

=7.7 m/s and the corresponding flow rate of ˙

cf,

=15.4 l/min.

The experiments are repeated three times to ensure reproducibility.

During experiments the process forces are measured by the dynamom-

eter of type 9121 from Kistler Holding AG. The experiments are termi-

nated once a constant force level is reached. Chips are collected,

embedded in epoxy resin, ground and etched. Subsequently, the chip

thickness h’is measured using a digital microscope. Details of the raw

data preparation methods are described in a primary study [53].

Fig. 2. Simulation results of the impact test; a) mechanical stresses on the impact plate; b) reaction forces F

Fig. 3. Setup of wet orthogonal turning experiments with detached view of developed high-pressure supply system.

Fig. 4. Schematic simulation model for wet orthogonal turning.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

107

4. Wet cutting simulation considering high-pressure cooling

4.1. Basic setup

Fig. 4 shows the wet orthogonal turning model which is a further

development of the primary work [44,54]. The model is implemented in

software version MESHFREE β2023.12.1 using the FPM. Geometry and

kinematic process parameters are adjusted based on the experiments.

The tool is modeled mechanically as a fixed rigid body, taking into ac-

count heat conduction. Thermal material parameters of the tool are

obtained from the literature [55]. The workpiece moves horizontally

with the cutting velocity v

towards the tool. A fixed boundary condition

is applied at the bottom surface and a minor part of the front surface

while the side surfaces are modeled as free surfaces. Due to the specific

meshfree approach, the FPM allows to activate and deactivate solid

bodies [45]. This feature is used for the workpiece material to save

calculation time. As a result, the initial active workpiece length of l

w,a

=1 mm can be used instead of the full workpiece length of l

=2 mm

for C45 and l

=3 mm for Inconel 718. The difference in workpiece

length is reasoned in different cutting path required for a thermal

convergence. For a further reduction of the computation effort, the

width of cut is reduced to b

sim

=0.7 mm. The process forces are then

scaled to the factor b/b

sim

during evaluation. Elastic and thermal ma-

terial properties for both workpiece materials are taken from literature

[10,56,57]. The plastic material behavior is described by the

Johnson-Cook model which depends on the plastic strain

, plastic

strain rate ˙

pand the temperature T, see Eq. 1. Used model parameters A,

B, C, T

, n and ˙

0are based on literature and are listed in Table 1. The

damage term D describes the effect of ductile damaging during chip

formation. For Inconel 718, the model proposed by Sievert et al. [58] is

utilized. According to Klocke et al. [59], Ning et al. [60] and Bergs et al.

[61] a damage model is not necessary for C45 due to reduced ductility.

An ideal thermal contact using a Dirichlet boundary condition is applied

between the tool and workpiece. Free surfaces are exposed to a free

convection with

con

=0.02 kW/(m

K) to the surroundings. Interior

surfaces are modeled as adiabatic.

F=(A+B⋅

pn)(1+Cln[˙

0])(1−[T−Tr

Tm−Tr]m)(1−D)(1)

The cutting fluid flows into the simulation domain at the inflow with

the cutting fluid velocity v

related to the experiments. In order to save

computation time, a cutting fluid outflow box is applied, outside of which

the cutting fluid is deactivated for the computation. The material prop-

erties of the emulsion have been determined in a previous study [46]. The

-turbulence model with an initial turbulence of Tu=5 % at the inflow

is utilized. All three phases of workpiece, tool and cutting fluid are solved

by the incompressible and implicit solver vp- of MESHFREE. Despite the

implicit time integration between solid and fluid phases, the motion of

numericalpointsiscalculatedexplicitlybytheusedsolver.Consequently,

the Courant–Friedrichs–Lewy-(CFL)-criterion must be applied to the

entire domain in each time step. During wet cutting, the cutting fluid

velocity is increased in comparison to velocity of the workpiece material

and thus the cutting fluid defines the time step size Δt. As the cutting fluid

velocityv

is increased through the useof high-pressure cooling, thetime

step size decreases. In the case of an explicit coupling between fluid and

solidphases, thisdrop isso significantthatthe modelisunable to convert.

This issue is resolved through the establishment of the implicit

surface-surface coupling method between both phases, described in

Chapter 2. In contrast to the primary works, the wet cutting simulation

model presented in this study can consider high-pressure stages in

numerically stable fashion. The discretization of the model is adaptably

managed. In the chip formation area, the tool and workpiece have a

minimum smoothing length of h

G,min

=12 µm. With growing distance

from the chip formation area, the smoothing length is increased up to h

max

=60 µm. A constant smoothing length of h

G,min

=25 µm is used for

the cutting fluid. Time integration is performed using adaptive time step

control with an additional limit to keep the step size below Δt

≤0.001 ms. The cutting model source files are published in a Mendely

Data repository [63].

4.2. Physical fluid-structure interaction

The physical interaction between chip and cutting fluid is based on

the three primary mechanisms of action dynamic pressure, lubrication

and cooling. In Chapter 2 the approach used to consider the dynamic

pressure of a cutting fluid is separately described and verified. Modeling

friction and lubrication specifically during cutting is a complex process.

Findings form literature under wet and dry conditions for the tool and

workpiece material combination used are limited. To simplify the

analysis, a friction experiment based on plate-on-cylinder-contact is

utilized, which is inspired by Lakner and Hardt [15].Fig. 5a) illustrates

the measuring principle. The uncoated carbide tool is pressed with the

preload force F

pre

against a cylinder made of workpiece material. During

experiments, the workpiece rotates with the contact velocity v

con

and

the forces in normal direction F

as well as tangential direction F

are

measured. Analyzed contact velocities v

con

are related to conducted

cutting velocities v

in Chapter 3. The technological setup of this prin-

ciple is realized in the turning lathe TNX 65, see Fig. 5b). A slotted shaft

with a width of cut b =2 mm is used as a workpiece. To measure the

process forces, a dynamometer of type 9121 from Kistler Holding AG is

mounted between the tool revolver and the tool holder. Prior to mea-

surements, the preloaded forces of F

pre

=100 or 200 N are adjusted by

moving the revolver. The analysis covers both dry and wet experiments.

The wet experiments are conducted under flood cooling conditions.

Using high-pressure supply does not provide additional insights in this

specific experiment because the kinetic energy dissipates when hitting

the workpiece in front of the contact.

The apparent friction coefficient µ

app

is in general a sum of the

friction coefficient µand the plastic deformation component µ

plast

, see

Eq. 2 [64,65]. To ensure a simplified evaluation, plastic deformations

must be minimized as far as possible. Therefore, contact stresses

are

calculated using Hertzian-contact and monitored during evaluation

procedure. Beside peaks in force signals, the contact stresses are

generally below the yield stress and thus µ

plast

=0 is assumed. Conse-

quently, the friction coefficient is calculated using the formula

µ=F

. The cutting process also involves stagnation areas with

plastic friction. In the cutting simulation model, this phenomenon is

Table 1

Used Johnson-Cook model parameter based on literature [58,62].

Material A B n C ˙

Unit MPa MPa 1 1 1/s ◦C◦C 1

C45 439 476 0.214 0.0181 1.0 1460 20 0.848

Inconel 718 450 1700 0.65 0.017 0.001 1300 20 1.8

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

108

addressed by the hybrid friction model and is not separately analyzed

with the presented friction test.

µapp =FT

=µ+µplast (2)

In Fig. 6 measured friction coefficients for dry conditions µ

dry

and

wet conditions µ

wet

are shown. The error bars represent the minimum

and maximum values of the repetitions for each parameter combination.

In most trials, the range of variation is below Δµ <0.1. The differences

between the preload forces are mostly within the range of variation. For

pre

=200 N, the range of variation is reduced compared to F

pre

=100 N. The average dry friction coefficient of Inconel 718 is µ

dry

=0.49, which is slightly increased compared to C45 with a mean value

of µ

dry

=0.4. Various friction coefficients have been presented and

utilized for cutting simulations in the literature. For C45 in combination

with uncoated carbide, Puls et al. [66] measured apparent friction co-

efficients between µ

app

=0.23 and 0.61, which is in the range of the

results of the present study. Zemzemi et al. [64] have found apparent

friction coefficients ranging from µ

app

=0.26 to 0.43 for the combina-

tion of Inconel 718 with uncoated carbides. The apparent friction co-

efficients are determined during plastic deformation, which can be a

reason for observed differences to this experiment. The wet friction

coefficients µ

wet

are significantly lower than the dry friction coefficients

dry

. This indicates that hydrodynamic lubrication acts during the fric-

tion tests in which the cutting fluid is applied directly in front of the

contact. This is in contrast to cutting. During cutting, juvenile surfaces

are produced and the cutting fluid can only penetrate from the chip

edges into the chip formation area.

Different assumptions are necessary to model lubrication due to a

lack of research in understanding lubrication. In Fig. 7 the rake face of

the cutting tool is divided into three sections. Due to contact conditions

during cutting, a penetration of cutting fluid into Sections 1 and 2 is

doubted in literature [13,14,67]. Consequently, dry friction is modeled

using the hybrid friction model, which is initiated by Zorev [68]. In

Section 3, lubrication is considered by applying the wet friction coeffi-

cient µ

wet

which is reduced in comparison to the dry friction coefficient

dry

. Penetration of the cutting fluid can be taken into account by the use

Fig. 5. Friction experiment; a) measuring principle; b) setup in turning lathe with close-up view of contact area.

Fig. 6. Measured dry and wet friction coefficients; a) C45; b) Inconel 718.

Fig. 7. Sections of lubrication model at rake face.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

109

of the wet contact length l

wet

. Analytical approaches of the penetration

following Smith et al. [69] or Godlevskiy et al. [70] use a wide range of

assumptions and are evaluated as general estimations. Consequently,

the wet contact length is simplified as l

wet

=0.1 ⋅l

. The extent of Sec-

tions 1 and 2 is a result of the actual contact length l

and contact

conditions. This model is mathematically described by Eq. 3.

The model is parameterized using the measured friction coefficients

from Fig. 6 with a preload force F

pre

=200 N. The contact velocities v

con

from friction experiment are correlated to the cutting velocities v

cutting experiments so that friction coefficients depending on v

are

modeled. In analogy to Puls et al. [66], Klocke et al. [59] and Malakizadi

et al. [71] for both workpiece materials the shear friction coefficient of

=1 is used. For dry conditions, the wet contact length is set to be l

wet

=0⋅l

to disable the lubrication effect.

In this lubrication model, the cause of lubrication such as hydrody-

namic lubrication or chemical interaction is secondary while the effect

on cutting is the primary focus. The model includes necessary assump-

tions but allows consideration of lubrication in a simple approach. In the

future, an advanced lubrication model will be elaborated considering

friction analyses during cutting based on Arrazola et al. [72].

The cooling capacity of the cutting fluid and its modeling approach

using a HTC have been analyzed in previous studies [54,73]. In the

following the applied approach is summarized. The transfer of heat from

a solid to a fluid is approximated by Newton’s law of cooling, see Eq. 4.

⋅(Ts−Tf)(4)

Here, the heat flux ˙

q is calculated as the product of the HTC

and the

temperature difference between the solid temperature T

and the fluid

temperature T

. The HTC values used in this paper have been determined

through jet cooling experiments [54]. In these experiments, a preheated

disk has been cooled down from temperature T

≈400 ◦C to 25 ◦C using

a cutting fluid jet. The HTC is dependent on various parameters such as

fluid flow properties, material properties or the temperature difference.

To ensure comparable conditions with wet cutting, the same workpiece

materials for the disk and same cutting fluid have been applied in the jet

cooling experiments. Further, the cutting fluid flow properties have been

related to flood cooling. The measured time-dependent temperature

field of the disk during jet cooling was used as input to solve an inverse

heat-transfer problem in 1D to obtain temperature-dependent HTC

). In Fig. 8 the determined and implemented HTC

) are shown,

which are in the upper range of literature. C45 exhibits a significantly

higher HTC compared to Inconel 718, which is believed to be due to its

higher thermal conductivity. During cutting processes, the temperature

range can exceed the maximum preheatable temperature of T

≈400 ◦C,

as reached in jet cooling experiments. Consequently, a constant HTC

) must be assumed above T

≈400 ◦C. Using HTC is a simple method

to consider the complex mechanism of cooling compared to complex

wall models which require high resolution of boundary layers. To verify

the modeled HTC, the jet experiment has been simulated using the FPM.

A maximum deviation of 15 % between simulation and experiment has

been observed [54]. However, the simulation has indicated deficits such

as neglection of evaporation or reduced accuracy in spatial temperature

distribution.

5. Results and discussion

5.1. Model verification

Simulated and measured cutting process forces and chip thicknesses

h’are presented for C45 in Fig. 9 and for Inconel 718 in Fig. 10. The

error bars indicate the range of minimum and maximum values obtained

from experiments. Simulated dry cutting forces for C45 match the

measurements with a maximum deviation of D

=10 % for v

=100 m/min. Additionally, for flood cooling of C45 the simulation

model is able to predict the experimental cutting forces with deviations

below D

<14 %. The prediction accuracy for Inconel 718 is lower

when using dry and flood cooling conditions compared to C45. This is

suspected to result from developing tool wear by the use of uncoated

carbides. Using high-pressure stages results in a decrease in both

experimental and simulated cutting forces. However, this decrease is

Fig. 8. Implemented HTC

) based on previous study [54].

(3)

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

110

overestimated for both materials, the cause of which is discussed below.

The simulations underestimate the feed forces compared to measure-

ments. Under dry conditions, cutting C45 results in a maximum devia-

tion in feed forces of D

=43 % and cutting Inconel 718 up to D

=57 %. Underestimated feed forces are a common issue in cutting

simulations and there are various possible explanations, such as

neglecting wear or material anisotropy [74]. In many cases, the pre-

dicted dry chip thicknesses are within the range of experimental

reproducibility for Inconel 718. For C45, there is a slight reduction in the

predicted chip thickness compared to the experiments.

To simplify the model verification overview, the mean deviation D is

calculated using Eq. 5, where the index sim represents the simulated

results and exp represents the experimental measured values.

D=100/3(

Fc,sim

Fc,exp

−1

+

Ff,sim

Ff,exp

−1

+

h’c,sim

h’c,exp

−1)(5)

Fig. 11 shows the mean deviation D for the conducted parameter

variation. The dry and wet processes are estimated to have a mean de-

viation D below 30 % for C45. High-pressure stage HP2 results in mean

deviations above 35 %. In general, the mean deviation D for Inconel 718

is increased but mostly below 35 %, indicating a satisfactory accuracy

for this difficult-to-cut material. Various combinations of material and

friction parameters have been evaluated during development of the

model. The set parameters result in the lowest mean deviations over all

parameter variations. Overall, the accuracy is sufficient for such a

complex process.

5.2. Numerical insights in chip formation

In addition to total accuracy, the tendencies between dry and wet

conditions are of particular interest. Various tendencies are consistently

observed in both experiment and simulation. To comprehend the rea-

sons for observed tendencies, the simulation provides additional insights

into the chip formation process. The following analysis is divided into

two parts, one for each workpiece material.

C45

Changes in chip thicknesses h’between cooling strategies are within

the range of experimental reproducibility for C45 which is consistent

with minor changes in the simulations, see Fig. 9. In Fig. 12 the exper-

imental chip morphologies and different simulated field variables under

dry condition, flood cooling and high-pressure stage HP2 are presented.

Under HP2, experimental and simulated chips show increased curvature

compared to the other conditions. In this specific simulation setup, the

cutting fluid flow is oriented towards the root of the chip and therefore

the dynamic pressure does not bend the chip. The bending of the chip

Fig. 9. Model verification for C45.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

111

Fig. 10. Model verification for Inconel 718.

Fig. 11. Mean deviations of simulation model; a) C45; b) Inconel 718.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

112

under simulation of HP2 is a result of different temperatures on the

upper and bottom chip surfaces. Due to a longer process time in

experiment compared to simulations, longer chips are produced. In

consequence, the dynamic pressure can lead to chip bending, which

explains the spiral chips in experiments using HP2.

Experimental and simulated results indicate that the process forces

are widely unaffected by dry and flood cooling for C45, see Fig. 9. In

particular, high-pressure stage HP2 results in a reduction of experi-

mental and simulated process forces. During the experiment for v

=100 m/min the cutting force is reduced by approximately 21 % be-

tween dry and HP2 and feed force by approximately 45 %. Using high-

pressure stages in simulations leads to an overestimation of this decline

with a reduction of cutting force by 60 % and feed force by 60 %.

However, the trend in process forces is identical in simulation and

experiment over all cooling conditions, which has different causes.

Both flood cooling and HP2 decrease the temperature level in chips

compared to dry processes, see Fig. 12 b). The temperature decrease is

significantly increased using HP2 compared to flood cooling. The high

mass flow using HP2 effectively removes heat from the chip, resulting in

a temperature reduction in the primary shear zone as well. The chip

thicknesses are only slightly reduced from dry to the wet conditions.

However, this leads to slightly increased plastic strain rates ˙

as shown

in Fig. 12 c).

Both the reduced temperature and increased plastic strain rates

should result in increased flow stresses

following the Johnson-Cook

material model. In Fig. 13, relevant stress fields are shown. Contrary

to this expectation, the current flow stress level is slightly reduced for

HP2 compared to dry and flood cooling. Only at the interface between

the chip and the tool is there a more significant reduction in flow

stresses, which is one explanation for the decrease in feed forces.

Additionally, the described chip bending results in a reduction in the

contact length l

using HP2 compared to the other conditions. The

transfer of stresses in the feed direction, respectively the feed force, is

limited by the hybrid friction model and is therefore directly linked to

the reduced flow stress and contact length, see Eq. 3. Both effects seem to

be more relevant compared to lubrication for feed force reduction using

HP2. Due to the minor differences in feed forces between dry and flood

cooling both in experiment and simulation, the marginal effect of

lubrication is detected. Between flood cooling and HP2, lubrication is

identically modeled in the simulation and therefore presumed to be of

minor importance for feed force decline using HP2.

However, these effects cannot explain the decrease in cutting forces

using HP2 compared to the other conditions in experiment and simu-

lation. The process forces result directly out of the mechanical stress

tensor

. The mechanical stress tensor is composed of the hydrostatic

stress

and the deviatoric stress approximated by the von Mises stress

. In the area of the secondary shear zone, the hydrostatic stress

level

is significantly lower using HP2 compared to the other conditions, see

Fig. 13 b). The differences in this component of the stress tensor explains

the decrease of the cutting force in simulations. In accordance with the

flow stress, the reduction of von Mises stress is less noticeable. As a

result, the marginally changed chip formation, represented by reduced

chip thicknesses as well as increased chip bending, requires less hy-

drostatic stresses to be formed. In accordance with these findings, a

reason for the observed overestimation in the reduction of simulated

process forces is presumed to the idealized modeled cutting fluid flow

towards the root of the chip. During real cutting, the accumulation of

chips can deflect the flow, thereby reducing the dynamic pressure effect

of the cutting fluid.

Inconel 718

Due to different material properties, Inconel 718 behaves slightly

different. The experimental process forces appear to be a bit higher

Fig. 12. Section views of chips for C45; a) experimental chips; b) simulated temperature T fields; c) simulated plastic strain rate ˙

fields.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

113

when using flood cooling compared to dry machining, see Fig. 10. This

trend is consistent with simulations. In accordance to C45, increasing

the cutting fluid velocity by the use of high-pressure stage HP2 leads to a

reduction of cutting forces compared to flood cooling both in experiment

and simulation. The maximum reduction of cutting forces for Inconel

718 is observed for v

=40 m/min with 20 % in experiment and 40 % in

simulation. Further, this decline between flood cooling to HP2 is more

significant for feed forces in experiment and simulation. The maximum

reduction in feed forces has accordingly occurred at v

=40 m/min with

37 % for experiment and 71 % for simulation. This effect can be also

observed for v

=100 m/min but is decreased with cutting force

reduction between flood cooling and HP2 of 10 % in experiment and

26 % in simulation. Both, the experimental and simulated chip thick-

nesses for Inconel 718 tend to decrease slightly with increasing cutting

fluid velocity, see Fig. 10. Due to the practical relevance, in Fig. 14 chip

morphologies and important field variables are represented for v

=100 m/min. The experimental chip morphologies seem to become a

wavier shape with increasing cutting fluid velocity, which is not directly

detected in simulation results. The lower temperatures in the chip using

HP2 compared to flood cooling and dry conditions can also be observed

for Inconel 718. In contrast to C45, another effect is taken into account

during chip formation using HP2. Small stripes with low flow stresses

using HP2 indicate an activation of ductile damage due to the specific

chip formation conditions. As a result, the flow stresses are significantly

reduced in certain areas of the chip. This can be seen as a reason for

decrease in process forces using HP2 compared to the other conditions

during cutting of Inconel 718. Further, a wavy or segmented chip for-

mation, respectively shear bands, are a well-known characteristic for dry

high-speed cutting and are based on ductile damaging [58,75]. Conse-

quently, these simulation results indicate that the positive effect of

high-speed cutting and high-pressure cooling has the same cause in

activation of ductile damaging in this specific case. The hydrostatic

stress

is less affected by the high-pressure stage compared to C45

which is presumed to be based on the generally increased level of

compared to C45. However, due to ductile damage, the interpretation of

the chip formation of Inconel 718 is more complex compared to C45. In

the future, the reasons for the altered chip formations and activation of

ductile damaging will be analyzed numerically by selectively consid-

ering the primary mechanisms of action in the simulation model.

Additionally, a temperature boundary layer in cutting fluid with

temperatures T

>100 ◦C around the chip surface is significantly

reduced using HP2 compared to flood cooling, see Fig. 14 b). Due to the

water-based emulsion, temperatures above T

>100 ◦C suggest an

evaporation of cutting fluid, which is not modeled. However, the dif-

ferences between cooling strategies indicate that an insulating damping

barrier can be avoided by the use of high cutting fluid velocities.

Nevertheless, cavitation, as reported in the literature by Tamil Alagan

et al. [30], is not modeled and therefore cannot be evaluated so far.

6. Summary and conclusion

The primary objective of this paper is to utilize a simulation model

for the analysis of chip formation for dry, flood cooling and high-

pressure cooling conditions. Firstly, the research procedure is based on

a concise review of pertinent literature. To enable high-pressure cooling

stages in the simulation, the implicit surface-surface coupling method is

Fig. 13. Simulated section views of chips with fields for C45; a) flow stress

fields; b) hydrostatic stress

fields; c) von Mises stress

fields.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

114

used for the fluid-structure interaction and the transfer of dynamic

pressure from fluid into a solid is verified by an impact test. In order to

analyze wet cutting, orthogonal turning processes are performed to vary

the influence of cutting fluid on the chip formation mechanisms. Based

on the experiments, a wet cutting simulation model using the FPM is

presented. In this model the three primary mechanisms of action of

cooling, lubrication and the dynamic pressure are quantifiably consid-

ered based on experiments.

The simulation model is verified by experimentally determined

process forces as well as chip thicknesses. A systematic underestimation

of feed forces is identified as a phenomenon that is also known to occur

in other simulation methods. Furthermore, reductions in process forces

using high-pressure cooling compared to other conditions are over-

estimated, which is presumed to be due in part to a deflection of the

cutting fluid flow during real cutting. However, in most cases the mean

deviation D between simulation and experiment is below 35 % and

consequently the presented simulation model is evaluated as acceptably

accurate.

In particular, several tendencies between different cutting fluid

strategies can be observed. In contrast to experiments, the simulation

model provides further insights into the chip formation process.

Consequently, different simulated field variables are used to analyze the

reasons for the observed trends. In conclusion of this additional infor-

mation, the following explanations can be given for chip formation

under the use of high-pressure cooling:

C45

•Chip bending can result solely from different temperatures on the

upper and bottom chip surfaces.

•Lower contact lengths and reduced flow stresses seem to be more

relevant for feed force reduction compared to lubrication.

•Small changes in chip formation led to a decrease in hydrostatic

stress which explains the reduced cutting forces for high-pressure

cooling compared to other conditions.

Inconel 718

•Ductile damaging takes place using high-pressure cooling, resulting

in a reduction of the flow stress.

Fig. 14. Section views of chips for Inconel 718; a) experimental chips; b) simulated temperature T fields; c) flow stress

fields; c) hydrostatic stress

fields.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

115

•Differences in temperature boundary layer of the cutting fluid using a

high-pressure stage compared to flood cooling are interpreted as a

reduced cutting fluid evaporation.

These findings confirm the achievement of the primary objective of

the study. However, the conclusions drawn are specific to a particular

cutting application. Consequently, the model will be used and further

refined for additional analysis in the future. For instance, an advanced

lubrication model will be developed. By selectively considering the

primary mechanisms, the role of each primary mechanism of action will

be analyzed in detail. Nevertheless, wet cutting simulations provide

comprehensive data sets on the interaction between chip formation and

cutting fluid, which are highly non-linear and complex to analyze. To

support the conversional result analyzes, methods of artificial intelli-

gence are intended to be used in the future. Further, the model will be

extended by an evaporation model for the cutting fluid to understand the

effects of evaporation as well as cavitation. For high-pressure stages, a

self-developed supply system has been employed to facilitate a simple

experiment and simulation setup. However, the full potential of high-

pressure cooling is not exploited because the jet is not injected be-

tween rake face and chip. A commercial supply system will be also used

as a subsequent step for a more industrially relevant application and

conclusions. Furthermore, this may facilitate the optimization of cutting

fluid strategies based on simulations. Due to the numerous issues that

have arisen, the source files of the developed simulation model are

available in a Mendeley repository [63].

CRediT authorship contribution statement

Eckart Uhlmann: Writing –review &editing, Supervision, Re-

sources, Project administration, Funding acquisition. Enrico Barth:

Writing –review &editing, Writing –original draft, Visualization,

Validation, Software, Methodology, Investigation, Formal analysis, Data

curation, Conceptualization. Benjamin Bock-Marbach: Writing –re-

view &editing, Software, Formal analysis. J¨

org Kuhnert: Writing –

review &editing, Supervision, Project administration, Funding

acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence

the work reported in this paper.

Acknowledgements

The authors express gratitude towards the Deutsche For-

schungsgemeinschaft DFG (German Research Foundation) for financial

support of this research. This work is funded by DFG-project “Multi-

phase modelling of cutting fluid and its aerosols in cutting simulations

using the Finite Pointset Method (FPM) for analysing the mechanisms of

action”, project number 439626733.

References

[1] Stanford M, Lister PM, Morgan C, Kibble KA. Investigation into the use of gaseous

and liquid nitrogen as a cutting fluid when turning BS 970-80A15 (En32b) plain

carbon steel using WC–Co uncoated tooling. J Mater Process Technol 2009;209(2):

961–72. https://doi.org/10.1016/j.jmatprotec.2008.03.003.

[2] Devillez A, Le Coz G, Dominiak S, Dudzinski D. Dry machining of Inconel 718,

workpiece surface integrity. J Mater Process Technol 2011;211(10):1590–8.

https://doi.org/10.1016/j.jmatprotec.2011.04.011.

[3] Brinksmeier E, Meyer D, Huesmann-Cordes AG, Herrmann C. Metalworking

fluids—Mechanisms and performance. CIRP Ann - Manuf Technol 2015;64(2):

605–28. https://doi.org/10.1016/j.cirp.2015.05.003.

[4] Denkena B, Kr¨

odel A, Ellersiek L. Influence of metal working fluid on chip

formation and mechanical loads in orthogonal cutting. Int J Adv Manuf Technol

2022;118(9-10):3005–13. https://doi.org/10.1007/s00170-021-08164-2.

[5] Astakhov VP. High-pressure supply of metalworking fluids. In: Astakhov VP,

Joksch S, editors. Metalwork Fluids (MWFs) Cut Grind Oxf, Camb, Phila, N Delhi:

Woodhead Publ 2012:201–90.

[6] Yılmaz B, Karabulut S

¸, Güllü A. A review of the chip breaking methods for

continuous chips in turning. J Manuf Process 2020;49:50–69. https://doi.org/

10.1016/j.jmapro.2019.10.026.

[7] Tanveer A, Marla D, Kapoor SG. A thermal model to predict tool temperature in

machining of Ti–6Al–4V alloy with an atomization-based cutting fluid spray

system. J Manuf Sci Eng 2017;139(7). https://doi.org/10.1115/1.4036123.

[8] Kops L, Arenson M. Determination of convective cooling conditions in turning.

CIRP Ann - Manuf Technol 1999;48(1):47–52.

[9] Luchesi VM, Coelho RT. Experimental investigations of heat transfer coefficients of

cutting fluids in metal cutting processes: analysis of workpiece phenomena in a

given case study. Proc Inst Mech Eng, Part B: J Eng Manuf 2012;226(7):1174–84.

https://doi.org/10.1177/0954405412442459.

[10] Gerstenberger R. Netzfreie Zerspanungssimulation mit der Finite-Pointset-

Methode. Dissertation, Technische Universit¨

at Berlin. In: Uhlmann E., editor.

Berichte aus dem Produktionstechnischen Zentrum Berlin. Stuttgart: Fraunhofer

IRB; 2015.

[11] Liu H, Meurer M, Schraknepper D, Bergs T. Investigation of the cutting fluid’s flow

and its thermomechanical effect on the cutting zone based on fluid–structure

interaction (FSI) simulation. Int J Adv Manuf Technol 2022;121(1-2):267–81.

https://doi.org/10.1007/s00170-022-09266-1.

[12] Courbon C, Sajn V, Kramar D, Rech J, Kosel F, Kopac J. Investigation of machining

performance in high pressure jet assisted turning of Inconel 718: a numerical

model. J Mater Process Technol 2011;211(11):1834–51. https://doi.org/10.1016/

j.jmatprotec.2011.06.006.

[13] Childs T. Friction modelling in metal cutting. Wear 2006;260(3):310–8. https://

doi.org/10.1016/j.wear.2005.01.052.

[14] Yan P, Rong Y, Wang G. The effect of cutting fluids applied in metal cutting

process. Proc Inst Mech Eng, Part B: J Eng Manuf 2015;230(1):19–37. https://doi.

org/10.1177/0954405415590993.

[15] Lakner T, Hardt M. A novel experimental test bench to investigate the effects of

cutting fluids on the frictional conditions in metal cutting. J Manuf Mater Process

2020;4(2):45. https://doi.org/10.3390/jmmp4020045.

[16] Volke P, Courbon C, Krumme E, Saelzer J, Rech J, Biermann D. Frictional

behaviour of coated carbide tools and AISI 316L when using translational and

rotatory relative movement considering dry and lubricated conditions. CIRP J

Manuf Sci Technol 2024;51:36–46. https://doi.org/10.1016/j.cirpj.2024.03.011.

[17] Ministry of the Environment, Climate Protection and the Energy Sector Baden-

Württemberg, Germany. Betrieblicher Umweltschutz in der Metallbearbeitung.

Stuttgart; 2015.

[18] Benedicto E, Carou D, Rubio EM. Technical, economic and environmental review

of the lubrication/cooling systems used in machining processes. Procedia Eng

2017;184:99–116. https://doi.org/10.1016/J.PROENG.2017.04.075.

[19] Sanchez JA, Pombo I, Alberdi R, Izquierdo B, Ortega N, Plaza S, et al. Machining

evaluation of a hybrid MQL-CO

grinding technology. J Clean Prod 2010;18(18):

1840–9. https://doi.org/10.1016/j.jclepro.2010.07.002.

[20] Bartolomeis A, de, Newman ST, Biermann D, Shokrani A. State-of-the-art cooling

and lubrication for machining inconel 718. J Manuf Sci Eng 2021;143(5). https://

doi.org/10.1115/1.4047842.

[21] Pigott R., Colwell A.T. Hi-Jet System for Increasing Tool Life. SAE Technical Paper

1952(520254). 〈https://doi.org/10.4271/520254〉.

[22] Ayed Y, Germain G, Ammar A, Furet B. Tool wear analysis and improvement of

cutting conditions using the high-pressure water-jet assistance when machining the

Ti17 titanium alloy. Precis Eng 2015;42:294–301. https://doi.org/10.1016/j.

precisioneng.2015.06.004.

[23] Kramar D, Krajnik P, Kopac J. Capability of high pressure cooling in the turning of

surface hardened piston rods. J Mater Process Technol 2010;210(2):212–8.

https://doi.org/10.1016/j.jmatprotec.2009.09.002.

[24] Pimenov DY, Bustillo A, Wojciechowski S, Sharma VS, Gupta MK, Kunto˘

glu M.

Artificial intelligence systems for tool condition monitoring in machining: analysis

and critical review. J Intell Manuf 2023;34(5):2079–121. https://doi.org/

10.1007/s10845-022-01923-2.

[25] Su´

arez A, L´

opez de Lacalle LN, Polvorosa R, Veiga F, Wretland A. Effects of high-

pressure cooling on the wear patterns on turning inserts used on alloy IN718. Mater

Manuf Process 2017;32(6):678–86. https://doi.org/10.1080/

10426914.2016.1244838.

[26] Ezugwu EO, Bonney J. Effect of high-pressure coolant supply when machining

nickel-base, Inconel 718, alloy with coated carbide tools. J Mater Process Technol

2004;153-154:1045–50. https://doi.org/10.1016/j.jmatprotec.2004.04.329.

[27] Tamil Alagan N, Hoier P, Zeman P, Klement U, Beno T, Wretland A. Effects of high-

pressure cooling in the flank and rake faces of WC tool on the tool wear mechanism

and process conditions in turning of alloy 718. Wear 2019;434-435:102922.

https://doi.org/10.1016/j.wear.2019.05.037.

[28] Ezugwu EO, Da B, Silva R, Bonney J, Costa ES, Sales WF, et al. Evaluation of

performance of various coolant grades when turning Ti-6Al-4V alloy with uncoated

carbide tools under high-pressure coolant supplies. J Manuf Sci Eng 2019;141(1).

https://doi.org/10.1115/1.4041778.

[29] Naves V, Da Silva MB, Da Silva FJ. Evaluation of the effect of application of cutting

fluid at high pressure on tool wear during turning operation of AISI 316 austenitic

stainless steel. Wear 2013;302(1-2):1201–8. https://doi.org/10.1016/j.

wear.2013.03.016.

[30] Tamil Alagan N, Hoier P, Beno T, Klement U, Wretland A. Coolant boiling and

cavitation wear –a new tool wear mechanism on WC tools in machining Alloy 718

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

116

with high-pressure coolant. Wear 2020;452-453:203284. https://doi.org/10.1016/

j.wear.2020.203284.

[31] Cica D, Kramar D. Machinability investigation and sustainability analysis of high-

pressure coolant assisted turning of the nickel-based superalloy Inconel 718. Proc

Inst Mech Eng, Part B: J Eng Manuf 2023;237(1-2):43–54. https://doi.org/

10.1177/09544054221092939.

[32] Markopoulos AP, Karkalos NE, Papazoglou E-L. Meshless methods for the

simulation of machining and micro-machining: a review. Arch Comput Methods

Eng 2019;27:831–53. https://doi.org/10.1007/s11831-019-09333-z.

[33] Rodríguez JM, Carbonell JM, Jons´

en P. Numerical methods for the modelling of

chip formation. Arch Comput Methods Eng 2020;27(2):387–412. https://doi.org/

10.1007/s11831-018-09313-9.

[34] Ozel T, Llanos I, Soriano J, Arrazola P-J. 3D finite element modelling of chip

formation process for machining Inconel 718: comparison of FE software

predictions. Mach Sci Technol 2011;15(1):21–46. https://doi.org/10.1080/

10910344.2011.557950.

[35] Ducobu F, Rivi`

ere-Lorph`

evre E, Filippi E. Numerical contribution to the

comprehension of saw-toothed Ti6Al4V chip formation in orthogonal cutting. Int J

Mech Sci 2014;81:77–87. https://doi.org/10.1016/j.ijmecsci.2014.02.017.

[36] Zhang Y, Outeiro JC, Mabrouki T. On the selection of johnson-cook constitutive

model parameters for Ti-6Al-4V using three types of numerical models of

orthogonal cutting. Procedia CIRP 2015;31:112–7. https://doi.org/10.1016/j.

procir.2015.03.052.

[37] Soliman HA, Shash AY, El Hossainy TM, Abd-Rabou M. Investigation of process

parameters in orthogonal cutting using finite element approaches. Heliyon 2020;6

(11):e05498. https://doi.org/10.1016/j.heliyon.2020.e05498.

[38] Ducobu F, Rivi`

ere-Lorph`

evre E, Filippi E. Application of the Coupled Eulerian-

Lagrangian (CEL) method to the modeling of orthogonal cutting. Eur J Mech - A/

Solids 2016;59:58–66. https://doi.org/10.1016/j.euromechsol.2016.03.008.

[39] Uhlmann E, Gerstenberger R, Kuhnert J. Cutting simulation with the meshfree

finite pointset method. Procedia CIRP 2013;8:391–6. https://doi.org/10.1016/j.

procir.2013.06.122.

[40] Grimm C, Menze C, Saelzer J, Sauer F, Eisseler R, M¨

ohring H-C, et al.

Reproducibility analysis for different numerical models and experimental setups in

dry orthogonal cutting of AISI 4140 steel. Procedia CIRP Des Conf 2024(Acepted

2024;15:04.

[41] Oezkaya E, Biermann D. A new reverse engineering method to combine FEM and

CFD simulation three-dimensional insight into the chipping zone during the

drilling of Inconel 718 with internal cooling. Mach Sci Technol 2018;22(6):

881–98. https://doi.org/10.1080/10910344.2017.1415933.

[42] Helmig T, Peng B, Ehrenpreis C, Augspurger T, Frekers Y, Kneer R, et al. A coupling

approach combining computational fluid dynamics and finite element method to

predict cutting fluid effects on the tool temperature in cutting processes. J Manuf

Sci Eng 2019;141(10):1–6. https://doi.org/10.1115/1.4044102.

[43] Klocke F, D¨

obbeler B, Peng B, Lakner T. FE-simulation of the cutting process under

consideration of cutting fluid. Procedia CIRP 2017;58:341–6. https://doi.org/

10.1016/j.procir.2017.03.235.

[44] Uhlmann E, Barth E, Seifarth T, H¨

ochel M, Kuhnert J, Eisentr¨

ager A. Simulation of

metal cutting with cutting fluid using the Finite-Pointset-Method. Procedia CIRP

2021;101:98–101. https://doi.org/10.1016/j.procir.2021.02.013.

[45] Uhlmann E, Barth E, Quellhorst A, Seifarth T, Kuhnert J, Nabbout K, et al. Fully

coupled wet cylindrical turning simulation using the Finite-Pointset-Method.

Procedia CIRP 2021;102:43–8. https://doi.org/10.1016/j.procir.2021.09.008.

[46] Uhlmann E, Barth E, Nabbout K, Sommerfeld M, Bock-Marbach B, Kuhnert J, et al.

Cooling Lubricant Boundary Conditions for wet Turning Simulations. Procedia

CIRP 2023;118:384–9. https://doi.org/10.1016/j.procir.2023.06.066.

[47] MESHFREE. Online Documentation for MESHFREE. [June 18, 2024]; Available

from: 〈https://meshfree.pages.fraunhofer.de/docu/html/MESHFREE.InputFiles.

USER_common_variables.BoundaryConditions.LIQUID__BC__.BCON_CNTCT.ind_

v1___BCON_CNTCT___.html〉.

[48] Drumm C, Tiwari S, Kuhnert J, Bart H-J. Finite pointset method for simulation of

the liquid–liquid flow field in an extractor. Comput Chem Eng 2008;32(12):

2946–57. https://doi.org/10.1016/j.compchemeng.2008.03.009.

[49] Peng B, Bergs T, Schraknepper D, Smigielski T, Klocke F. Development and

validation of a new friction model for cutting processes. Int J Adv Manuf Technol

2020;107(11-12):4357–69. https://doi.org/10.1007/s00170-019-04709-8.

[50] Peng B, Bergs T, Klocke F, D¨

obbeler B. An advanced FE-modeling approach to

improve the prediction in machining difficult-to-cut material. Int J Adv Manuf

Technol 2019;103(5-8):2183–96. https://doi.org/10.1007/s00170-019-03456-0.

[51] Klocke F, D¨

obbeler B, Peng B, Schneider S. Tool-based inverse determination of

material model of Direct Aged Alloy 718 for FEM cutting simulation. Procedia CIRP

2018;77:54–7. https://doi.org/10.1016/j.procir.2018.08.211.

[52] Ohnesorge WV. Die Bildung von Tropfen an Düsen und die Aufl¨

osung flüssiger

Strahlen. Z für Angew Math und Mech 1936;16(6):355–8. https://doi.org/

10.1002/zamm.19360160611.

[53] Uhlmann E, Barth E, Seifarth T, H¨

ochel M, Kuhnert J, Eisentr¨

ager A. Data for a

simulation of metal cutting with cutting fluid using the Finite-Pointset-Method.

Data Brief 2021;38:107339. https://doi.org/10.1016/j.dib.2021.107339.

[54] Nabbout K, Sommerfeld M, Barth E, Uhlmann E, Bock-Marbach B, Kuhnert J.

Cooling Capacity of Oil-in-Water Emulsion under wet Machining Conditions.

Procedia CIRP 2023;117:74–9. https://doi.org/10.1016/j.procir.2023.03.014.

[55] Deppermann M. Werkstückerw¨

armung bei trockener Drehbearbeitung

[Dissertation]. Aachen; 2017.

[56] Hannosch¨

ock N. W¨

armeleitung und -transport. Berlin, Heidelberg: Springer Berlin

Heidelberg,; 2018.

[57] Spittel M., Spittel T. Steel symbol/number: C45/1.0503. In: Warlimont H, editor.

Materials Metal Forming Data of Ferrous Alloys - deformation behaviour.

Heidelberg, Berlin: SpringerMaterials; 2009, p. 1–5.

[58] Sievert R, Hamann A, Noack H-D, L¨

owe P, Singh KN, Künecke G. Simulation der

Spansegmentierung einer Nickelbasislegierung unter Berücksichtigung thermischer

Entfestigung und duktiler Sch¨

adigung. In: T¨

onshoff HK, Hollmann F, editors.

Hochgeschwindigkeitsspanen metallischer Werkstoffe. Weinheim: Wiley-VCH;

2005. p. 446–69.

[59] Klocke F, Lung D, Buchkremer S. Inverse Identification of the Constitutive Equation

of Inconel 718 and AISI 1045 from FE Machining Simulations. Procedia CIRP 2013;

8:212–7. https://doi.org/10.1016/j.procir.2013.06.091.

[60] Ning J, Liang SY. Inverse identification of Johnson-Cook material constants based

on modified chip formation model and iterative gradient search using temperature

and force measurements. Int J Adv Manuf Technol 2019;102(9-12):2865–76.

https://doi.org/10.1007/s00170-019-03286-0.

[61] Bergs T, Hardt M, Schraknepper D. Determination of Johnson-Cook material model

parameters for AISI 1045 from orthogonal cutting tests using the Downhill-Simplex

algorithm. Procedia Manuf 2020;48:541–52. https://doi.org/10.1016/j.

promfg.2020.05.081.

[62] Storchak, Rupp, M¨

ohring, Stehle. Determination of johnson–cook constitutive

parameters for cutting simulations. Metals 2019;9(4):473. https://doi.org/

10.3390/met9040473.

[63] Uhlmann E, Barth E, Bock-Marbach B, Kuhnert J. Cutting simulation model for the

article Numerical chip formation analysis during high-pressure cooling in metal

machining. Mendeley Data 2024:V1. 〈https://doi.org/10.17632/xxt96gs88k.1〉.

[64] Zemzemi F, Rech J, Salem WB, Dogui A, Kapsa P. Identification of friction and heat

partition model at the tool-chip-workpiece interfaces in dry cutting of an inconel

718 alloy with cbn and coated carbide tools. Adv Manuf Sci Technol 2014;38(1).

https://doi.org/10.2478/amst-2014-0001.

[65] Rech J, Arrazola PJ, Claudin C, Courbon C, Pusavec F, Kopac J. Characterisation of

friction and heat partition coefficients at the tool-work material interface in

cutting. CIRP Ann 2013;62(1):79–82. https://doi.org/10.1016/j.cirp.2013.03.099.

[66] Puls H, Klocke F, Lung D. A new experimental methodology to analyse the friction

behaviour at the tool-chip interface in metal cutting. Prod Eng 2012;6(4-5):

349–54. https://doi.org/10.1007/s11740-012-0386-6.

[67] Godlevskiy V. Mechanisms of action of metalworking fluids in metal cutting. In:

Astakhov VP, Joksch S, editors. Metalwork Fluids (MWFs) Cut Grind Oxf, Camb,

Phila, N Delhi: Woodhead Publ 2012:1–22.

[68] Zorev N.N. Inter-relationship between shear processes occurring along tool face

and shear plane in metal cutting. In: Shaw MC, editor. International Research in

Production Engineering; Proceedings of the International Production Engineering

Research Conference. New York: ASME; 1963, p. 42–49.

[69] Smith T, Naerheim Y, Lan MS. Theoretical analysis of cutting fluid interaction in

machining. Tribology Int Tribology Int 1988;21(5):239–47.

[70] Godlevski VA, Volkov AV, Latyshev VN, Maurin LN. The kinetics of lubricant

penetration action during machining. Lubr Sci 1997;9(2):127–40. https://doi.org/

10.1002/ls.3010090203.

[71] Malakizadi A, Hosseinkhani K, Mariano E, Ng E, Del Prete A, Nyborg L. Influence of

friction models on FE simulation results of orthogonal cutting process. Int J Adv

Manuf Technol 2017;88(9-12):3217–32. https://doi.org/10.1007/s00170-016-

9023-4.

[72] Arrazola PJ, Ugarte D, Domínguez X. A new approach for the friction identification

during machining through the use of finite element modeling. Int J Mach Tools

Manuf 2008;48(2):173–83. https://doi.org/10.1016/j.ijmachtools.2007.08.022.

[73] Nabbout K., Sommerfeld M., Bock-Marbach B., Kuhnert J., Barth E., Uhlmann E.

Heat Transfer Study for Oil-in-Water Emulsion Jets Impinging onto hot Metal

Surface. In: Proceedings of the 8th World Congress on Mechanical, Chemical, and

Material Engineering (MCM’22). Prag, Tschechien; 2022.

[74] Sadeghifar M, Sedaghati R, Jomaa W, Songmene V. A comprehensive review of

finite element modeling of orthogonal machining process: chip formation and

surface integrity predictions. Int J Adv Manuf Technol 2018;96(9-12):3747–91.

https://doi.org/10.1007/s00170-018-1759-6.

[75] Wertheim R. High Speed Cutting. In: Chatti S, Laperri`

ere L, Reinhart G, Tolio T,

editors. CIRP Encyclopedia of Production Engineering. Berlin, Heidelberg: Springer

Berlin Heidelberg; 2019. p. 871–93. https://doi.org/10.1007/978-3-662-53120-4.

E. Uhlmann et al. CIRP Journal of Manufacturing Science and Technology 53 (2024 ) 103–117

117