Temperature effect on insert tool life in dry machining

(1)

Temperature effect on insert tool life in dry machining

William Hagberg

Mechanical Engineering, master's level 2021

Luleå University of Technology

Department of Engineering Sciences and Mathematics

(2)

Abstract

This thesis work researched the temperature eﬀects the insert tool life. This was examined through physical laboratory testing with designed milling tools and temperature simulations in ANSYS. The designed milling tools altered the temperature in the insert through external parameters like design and material. Three milling tools were used in total: A milling tool created in an copper alloy that cools the inserts, a milling tool created in the material idun that increases the temperature in the inserts, and the milling tool Coro Mill 245 that was used as a reference. These tools were tested with two inserts, insert 1130 with a coating of PVD, and insert 4330 which has a coating of CVD.

The idun milling tool and the reference milling tool was already designed, but the copper milling tool was designed during the thesis work. The designed copper milling tool implemented a heat sink to cool the milling tool with convection and was manufactured in a copper alloy with high strength. The copper milling tool was 20 % colder than the reference milling tool when comparing the maximum temperature of the insert.

Through testing in dry milling of the three milling tools with diﬀerent inserts and a simulation in ANSYS of the temperature development, the conclusion was that insert 4330 had crater wear which increased with increased temperature. This crater wear can be the main factor in the insert tool life.

(3)

Acknowledgements

This report is my last step in the Master of Science education in Mechanical Engineering at the Lulea University of Technology.

I would like to thank my main supervisor from Sandvik Coromant Olof Larsson for his guidance and encouragement. I would also like to thank Linus Hasselrot and Ralf Lehto from Sandvik Coromant for their guidance in my work. Their knowledge and experience in the ﬁeld have been very helpful in the discussion and planning of the project.

I would like to thank Andreas Lundb¨ack as my supervisor from Lulea University of Technology with his guidance in my thesis work.

(4)

5 Discussion and Conclusions 30 5.1 Future Work . . . 30 A Wear measurement of insert 4330 with the reference milling tool i B Wear measurement of insert 4330 with the copper milling tool ii C Wear measurement of insert 4330 with the idun milling tool iii D Wear measurement of insert 1130 with the idun milling tool and the reference

milling tool iv

E Graph of the thermal camera results v

(5)

v_c

Designations Nomenclature

α Angle of engagement (^◦)

µ_air Dynamic viscosity of air (kg/(ms))

θ Cutting force angle in the verticle plane (^◦) ϕ Cutting force angle in the horizontal plane (^◦) κ Entering angle (^◦)

Emissivity (-)

ρ Density (kg/m³)

ρ_air Density of air (kg/m³) Ω Angular velocity (rad/s) a_e Radial depth of cut (mm) a_p Axial depth of cut (mm) D_C The cutter diameter (mm)

DC_ap Cutting diameter at cutting depth (mm)

E Young’s modulus (GPa)

F_C Speciﬁc cutting force (N/mm²)

F_T Specific cutting force in tangential direction (N/mm²) F_R Specific cutting force in radial direction (N/mm²) F_A Specific cutting force in axial direction (N/mm²) f_z Feed per tooth (mm/tooth)

h₀ Average outer convection (W/m²) k_air Thermal conductance of air (W/mK)

x_wall Wall length (mm)

m Mass (kg)

N u Average Nusselt number (-)

n Spindle speed (RPM)

Q Heat flow in the insert (W) Qsteadystate Steady state heate flow (W) qconduction Conduction heat flux (W/m²)

q_x Conduction heat ﬂux in the x-direction (W/m²) q_radiation Radiation heat ﬂux (W/m²)

Re_r Rotational Reynolds nymber (-) RP M Revolutions per minute (-)

T Temperature (^◦C)

T₁ Surface temperature (K) T₂ Surrounding temperature (K) tengagement Time in engagement (s) t_rotation Time for one rotation (s)

Cutting speed (m/min) v_f Table feed (mm/min) z Number of inserts (-)

x₁ x-value of the point of engagement (mm)

(6)

1 Introduction

1.1 Sandvik Coromant

As a part of the global industrial engineering group Sandvik, Sandvik Coromant is leading in manufacturing of tools and machinining solutions. They innovate new solutions that constantly moves the industry standards forward. With metalworking products for the car industry, aerospace, health technology and more. With the development of advanced machining technologies and systems, they will change the future of manufacturing. Sandvik Coromant owns more than 3100 patnents, they have over 7600 employees and is represented in 150 countries.

1.2 Background

During the cutting process milling, material removal of the workpiece is accomplished by several small shear cuts. The milling tools have one or several cuts (also known as inserts) that rotate around the mill’s axis. A typical milling operation can be seen in Figure 1. Friction acts on the contact between the workpiece’s chip and insert, also between the workpiece and insert. The friction between the surfaces is converted from kinetic energy to thermal energy. Due to the cutting forces, the friction generates a large amount of heat on the cutting zone. The temperature on the insert can rise to several hundred degrees.

Figure 1: An illustration of a milling tool cutting a workpiece.

The inﬂuence of heat is said to have a great impact on the insert’s tool life. To what extent is of limited knowledge. It’s feasible that the temperature increases the abrasive wear of the insert, since the material properties of the insert diﬀers at various temperatures. But studies are needed to quantify the direct relationship between temperature and insert tool life.

The question that is to be answered with this theses work is: ”does temperature effects the insert’s insert tool life?”. This is answered by altering external parameters, e.g. tool design and tool material, the heat development in the insert differs. Tools with different designs and material is then used to compare with one another, to showcase if temperature has an effect on insert tool life.

Three tools will be used for comparison. The ﬁrst milling tool will be designed to cool the inserts.

The second milling tool is constructed in the material idun, which has low thermal conductivity and therefore acts as a thermal insulator for the insert, which increases the temperature of the insert.

The third tool, Coro Mill 245, is the commercially available milling tool at Sandvik Coromant. The third tool is used as a reference for the idun milling tool and the designed milling tool, therefore it’s

(7)

called the reference milling tool. The temperature is analyzed with ﬁnite element simulations of the milling tool and insert in Ansys. The simulations are compared to temperature measurements from physical laboratory testing. The testing also covers the insert’s tool life for the three milling tools mentioned.

1.3 Restraints

The temperature measurement from the physical laboratory testing is measured with a thermal camera. Thermocouples could be an alternative. But to install the thermocouples, they would require a time consuming process where spark-erosion cutting is needed, which doesn’t ﬁt the time frame of the thesis work. The thermal camera can give reliable measurements on the cutting tool, but not for the insert during machining.

Finite element method is used to simulate the temperature on the inserts. Many parameters aﬀect the temperature of the insert, and the temperature gradient is large. The simulations will only give a notion of the temperature.

A limited amount of tests are used. For each milling tool, 3 repetitions are carried out with insert 4330, an insert design from Sandvik Coromant. Each milling tool is fully equipped with 6 inserts, which means a total of 54 wear samples for insert 4330. One repetition of tests are carried out with insert 1130 for Idun and reference milling tool, a total of 12 wear samples.

1.4 State of Art

Papers and articles are presented in this section. All have studied topics that are within the scope of this thesis.

Article (E. Uhlmann, H. Riemer, D. Schr¨oter, F. Sammler, S. Richarz, 2017) present a internally cooled milling tool for dry milling. Heatpipes were applied to improve heat dissipation from the insert because of the heatpipes thermal conductivity. The heatpipes are connected to a heat sink that transfers the heat to its surroundings. Simulations of the heat distribution was made to optimize the prototype. Tests were also made in duplex steel, and the temperature was measured to verify and optimise the simulation results.

The technical report (D. Lundström, 2020) shows the temperature development of insert and tool during dry longitudinal turning. Finite element simulations were made in ANSYS for dry longitudinal turning in different materials and cutting data. The simulations were verified with testing. The testing was made with thermocouples inside of the insert to measure the temperature.

The simulation was also used to compare the results with milling. When simulating turning there is a constant heat flow into the insert, by applying a varying heat flow instead, the simulations can be compared with milling. The thermal development of the insert was analyzed for milling with different cooling methods and cutting data.

The doctoral thesis (V. Kalhori, 2001) has done ﬁnite element simulations of the mechanical cutting process. The thesis shows the development of a simulation that can predict, chip formation, cutting forces, temperature, pressure distribution on the tool-chip interface and the residual stresses in the workpiece. Two modelling processes were used to simulate the chip formation, a geometrical and a physical model.

The paper (A. Bejan, A. Fowler, G. Stanesce, 1995) shows the spacing between the cylinders of a

(8)

pin-ﬁn heat sink with laminar natural convection, so that the total heat transfer between the heat sink and the ambient is maximized. The volume of the heat sink is ﬁxed, so is the pins’ diameter and height, only the spacing between the pins varies. Numerical simulations and experimental measurements were made to predict and validate the spacing.

The article by (A. Malmelöv, A. Lundbäck, L. Lindgren, 2020) discuss how to reduce the computing time for simulations of additive manufacturing. The computation time had been reduced significantly.

Also, a material model of Inconel 625 was created for the simulations. The material model was tested and validated with experimental data. For the testing, three temperature points and distortion was measured, and compared with the simulations.

(9)

2 Theory

2.1 Heat Transfer

The second law of thermodynamics gives that a volume of molecules with an average of higher kinetic energy, will transfer the kinetic energy to a volume with lesser average of kinetic energy.

In other words, heat transfers from hot to cold.

Heat can be categorized into 3 categories: conductance, convection, and radiation.

Convection is heat transfer via moving gases or liquids, like the vapor from a boiling pot of water onto your hand. Conduction is the heat transfer through solid objects, like heat transfer through the pot handle to your hand. Radiation is the emission of electromagnetic waves, which can transfer the energy through vacuum, for example the UV rays from the sun.

In milling, conduction, convection and radiation are all factors that aﬀect the process, radiation less so. When the system is in equilibrium it has reached Steady State. It occurs when the heat input and output is equal, like that of wet turning, where steady state is reached after less than a second. In dry milling it never occur since the heat input isn’t constant. But using Steady State equations will give insight on the heat gradient of the inserts. (D. Lundstr¨om, 2020)

The basic equations for steady state heat transfer are as follows:

For conduction, the diﬀerential form of Fourier’s law of thermal conduction is given in Equation 1, where the heat ﬂux is proportional to the temperature gradient

qconduction = −kOT. (1)

Where qconduction is the heat ﬂux, k is the thermal conductivity and −OT is the negative local temperature gradient.

Conductivity acts between the insert and the bodies in which it is in contact with, screw, holder and shim. The heat generated is typically in the range of 50 W to 100 W , distributed on a cutting

2 2

surface that is 0.5 mm to 1 mm .

Consider a 50 W heat input on a 1 mm surface, this gives a heat flux of 50 M W/m2 ². To get a magnitude of the gradients, evaluate the one-dimensional case seen in Equation 2, where qx is the heat flux in the x-direction. The variable 4xwall is the length of the wall that the heat is conducted through and 4T is the temperature difference over that length

q_x= k 4T . (2)

4x_wall

Given a 1 mm thick carbide wall, carbide’s thermal conductivity of around 100 W/mK, the temperature drop 1 mm from the cutting zone will be 500 ^◦C.

This is an assumption that does not describe the three dimensional case, but gives the idea of the magnitude of the temperature gradient near the cutting zone. In the 3D case, the temperature gradients will only be extreme at a distance much smaller than 1 mm. Since heat ﬂux will decay fast and so will the gradients.

The heat removal rate by means of convection from a solid surface to a surrounding ﬂuid or gas is

(10)

described by Newtons law of cooling. Convection in dry milling of steel is turbulance ﬂow.

The cutting zone has the highest temperatures, which means that the cutting zone has the the highest heat removal, since the temperature diﬀerence is the greatest. This is relevant when analyzing the radiation.

The emitted power from the surface is given in Stefan-Boltzmann radiation law in Equation 3,

q_radiation= σ(T ₂⁴− T₁⁴). (3)

Where qradiation is the radiated power per unit area, σ (Stefan-Boltzmann constant) is 5.67∗10⁻⁸, is the emissivity, and T1 is the surface temperature of the object in Kelvin, and T2 is the temperature of the environment surrounding the object. Radiation is only relevant near the cutting zone. For lower temperatures Stefan-Boltzmann constant is too low to have an aﬀect. The temperature is in the fourth degree, so in higher temperatures the radiation plays a big role. T needs to be in the region of 700 ^◦C or more to have an impact.

Only a small region on the surface of the insert reaches those temperatures. Therefore the total amount of emitted energy is small. Which means the radiation is neglected in this thesis work.

The temperature diﬀers across the surface of the insert. Which means that the expression for heat removal is a surface integral. Numerical methods are needed to calculate the temperature distribution. Therefore, the Finite Element Method is needed, see Section 2.4t.

2.2 Heat sink

A heat sink is a component that dispenses heat from an electronic or mechanical device, to a ﬂuid medium, e.g. water or air, via convection. Regulating the temperature of the device so it operates at a given temperature, often used in computers to cool its CPUs, GPUs and other mechanical components.

The design of the heat sink varies. But all of them are designed with a high surface area to volume ratio to increase the contact with the cooling medium. Air turbulence, material properties, protruding design, and surface treatment affect the performance of the heat sink. The way the heat sink is mounted to the device also affect the performance. Thermal paste is used to increase thermal conductivity between the contact of the device and heat sink. The thermal paste fills the air pockets between the two imperfect surfaces which can be seen in Figure 2.

Figure 2: The trapped air is replaced with thermal paste

Heat sinks are usually made out of copper or aluminium due to the materials’ high thermal conductivity. See Figure 3 for reference, where the heat sink is installed to a component. Fins are used to create a large surface area. As the medium passes through the ﬁns it dispenses the heat from the heat sink to the medium via forced convection.

(11)

Figure 3: An illustration of the principle behind a heat sink

Splayed pin fin heat sinks are alternatives to the standard pin fin heat sink. Instead of vertical pins, the splayed pin fin heat sink futures pins which axles are placed at an angle, gradually bending the pins from one another, see Figure 4. This allows the spacing between the pins to increase without affecting the surface area. With splayed pin fin heat sink the cooling medium will pass through the heat sink more efficiently.

The splayed pin ﬁn heat sink’s advantages are most potent at lower speeds of the cooling medium, since there is less power to penetrate the array of pins. At lower medium speeds and in natural convection, the splayed pins can reduce the heat sink’s thermal resistance with 30 % compared to the vertical pins. Which means that splayed pin ﬁn heat sinks are usually used under these circumstances.

Figure 4: A typical splayed heat sink in copper

For cylinders in rotation, a boundary layer will form on the rotating body due to the no-slip condition at the body surface as illustrated in Figure 5. At lower values of rotational Reynolds number, there is laminar ﬂow. As the rotational Reynolds number rises, the medium becomes transitional and then turbulent. The limit for laminar ﬂow is at a Reynolds number around 40 to 60.

Figure 5: The convection of a cylinder with angular velocity

The diameter of the milling tools is 80 mm and the workpiece will be in steel, which means a lower cutting speed. This will result in a lower rotational Reynolds number. Because of the lower Reynolds number there will be an advantage to use the splayed pin-ﬁn heat sink.

(12)

2.3 Insert Wear

There are a number of wears that can eﬀect the insert. They depend on cutting speed, workpiece material, temeprature and more. Some of the wears that can eﬀect the insert are described in the following sections.

2.3.1 ﬂank wear

Flank wear occurs on the flank side of the insert/tool due to loss of cutting tool material as a result of the flank side sliding against the newly cut workpiece material surface. The wear normally starts at the edge line and develops perpendicularly to the edge line downward on the flank side.

Flank wear is one of the most common wear types, because the tool is always in contact with the workpiece material during machining. Flank wear is mainly an abrasive mechanism, ﬂank wear generally increases in workpiece materials with a high proportion of hard constituents, such as carbines and slag inclusions. Occurrence of crater wear can also accelerate the ﬂank wear.

2.3.2 crater wear

Crater wear can occur during almost all cutting/machining operations, but is most common when there is a large and continuous engagement/contact, whereby high thermal and mechanical loads are generated. In contrary short, intermittent contact usually generates only limited crater wear.

Wear consisting of a slow growing crater in combination with ﬂank wear is relatively predictable.

The crater development can be divided into diﬀerent steps, with the following being applicable to coated cemented carbide tools. Before the actual crater reaches the substrate there is a gradual wearing of the coating. Since the coating is worn through, the actual crater wear occurs in the cemented carbide, the crater gradually becoming deeper. As the crater grows, the tool geometry becomes increasingly positive, and eventually edge breakage can occur due to the crater breakthrough.

A number of wear mechanisms can cause crater wear, for example chemical wear, abrasive wear and wear due to thermo-mechanical load.

Wear of the rake face can be divided into two zones: the sticking zone and the sliding zone. Closest to the edge line, the underside of the chip does not slide, but remains more or less ﬁxed against the insert’s surface. In this zone sliding occurs inside the chip. The further from the edge, the more chip slides against the insert’s surface for as long as the chip is in contact with the rake face.

abrasive wear, and/or elastic deformation of the coating surface with subsequential microfracture of the deformed coating occur in the sliding zone.

The wear development is aﬀected by the coating, the substrate and the chip breaker geometry.

Crater wear can be delayed by using cutting fluid. A chemical vapor deposition (CVD) coating generally produces a balanced combination of flank and crater wear, while with physical vapor deposition (PVD) coatings the crater wear is usually dominant. This is mainly due to the PVD coating having thinner coating thickness on the rake face compared to the flank side. For CVD coatings the thickness on the rake face and flank are almost the same.

The position of the crater on the rake face depends on the feed rate. A low feed rate can cause the

(13)

crater to develop very close the the edge.

2.3.3 Comb Cracks

Cracks perpendicular to the edge line is called comb cracks. Comb cracks are a form of thermo-mechanical wear mechanism that arise due to thermal cycling of the cutting edge. Common in wet, interrupted machining. It is also common for CVD coated inserts machining in steel.

2.3.4 Chipping

Chipping consists of minor damage to the edge line. The difference between chipping and fracture is that with chipping, the insert can still be used. The exact line between chipping and fracture cannot be defined, as different applications can tolerate different extents of chipping. Tool material can be ripped out between comb cracks, which is another form of chipping.

2.4 Finite Element Method

The ﬁnite element method (FEM) is a numerical method to solve engineering problems. When analytical solutions are not an option due to the complexity of the geometry or material properties, FEM is an alternative. FEM uses a numerical technique and solves diﬀerential equations. It is used in the industry for developing products, and optimization of products. Solid mechanics, thermal calculations, eigenvalues, electromagnetic potential are some of the applications of FEM.

The finite element method is often divided into pre-processing, simulation and post processing. For pre-processing the model is simplified by its geometry and irrelevent components are removed. This reduces the simulation time. When the model is simplified, the mesh is applied to the simulation model. The regions of interest have a finer mesh, and the regions that has less of an affect on the solution have a coarser mesh. This is also for reducing the simulation time. All the loads, constraints and contacts are set in the pre-processing.

The second section of the finite element method is simulation, where the computer solves differential equations. The simulation time is dependent on many factors. The numerical stability of the simulated model, the chosen time step, the amount of cores in the CPU used, are some of the parameters that affect the time to simulate.

The last section is post processing, where the chosen simulation results are examined through fringe plots or graphs. Deformation, stress and temperature are a few of many output results which can be analyzed. After analyzing the results, the simulated model can be altered if necessary.

2.5 Product Development Process

This project has been carried out with the help of the product development process. This product development process consists of a series of steps and activities to ensure the quality of the product.

Implementing this process oﬀers opportunities for quality assurance, better coordination and planning strategy, easier management and possibilities for improvement. Many development methods are used in the industry, but for this thesis the method described by (Ulrich & Eppinger, 2016) has been used.

(14)

3 Metod

3.1 Product Development

The product development follows (Ulrich & Eppinger, 2016) as much as possible, however some parts are left out. For example, identifying customer needs aren’t necessary since this thesis is for research purposes. The development process used is mentioned in the following sections.

3.1.1 Phase 0 - Planning

The planning is about researching and gathering enough information. And to evaluate if the project is feasible. If the project is feasible, a plan for the project is created.

Time and Resource Planning

A plan for the project is necessary. It’s created so that each development process is prioritized correctly and doesn’t use too much time. Then, the project will be ﬁnished within its time frame.

The projects time schedule can be defined with a Gantt Chart. It is a common tool for visualizing the project different tasks. Every task is represented by a bar whose position and length represent the start, duration and end of the specific activity. The chart gives a broad picture of the project as a whole.

3.1.2 Phase 1 – Concept Development

For the the concept development, the requirements of the product is identiﬁed. Solutions to fulﬁll those requirements are created with concept generation, where ideas are created to solve the problems. The ideas arrives from sessions of brainstorming and through analysis. The concepts are then evaluated to see which ones have the most potential. The concept gets chosen through an elimination process, where the concepts are stacked up against each other and eliminated one by one.

3.1.3 Phase 2 – System-Level Design

This phase of the development process contains preliminary designs of the components and a preliminary plan for the production process. To make sure that the concept really works, and to identify what’s problematic to design.

3.1.4 Phase 3 – Detail Design

The detail design phase is where all components is designed with exact, geometry, measurements and tolerances. The material is chosen for all components as well. If suppliers are needed for certain or all components, orders are placed, and a planning of the production process is made. All documentation required to make the ﬁnished product is made, that includes: drawings, assembly plan and simulations that proves requirement fulﬁllment.

When designing a product, several methods are possible, see (Autodesk.Help, 2016). All have their advantages and limitations. But the most common ones are ”Top-down”, ”Bottom-up” and

”Middle-out modeling”. The latter one has been used when designing the cooled milling tool. This means that the all modeled components revolve around a single component, the base, and all the components are designed in retrospect of the base component.

(15)

3.1.5 Phase 4 – Testing and Reﬁnement

Testing and refinement is the last phase. The product is tested to see how it performs, its strength and weaknesses; if necessary changes in the design or material are needed. Alpha prototypes are the first generation of prototypes, manufactured after the first intended design. They are used to see if the product fulfills the requirements. The alpha prototypes doesn’t need to be manufactured with mass production in mind, that is for the beta prototypes. Beta prototypes are the next generation of prototype, they are the tested and refined alpha prototype. They are also manufactured according to the planed production method. They are also tested to see if it fulfills the requirements, but also for security and reliability. Like for example strength tests, tensile tests and lifecycle tests. The product can be refined further, but it is at its final stages of development.

For this thesis the product only reaches the alpha test, and is tested for performance. But no other tests are made and no reﬁnement of the product is made.

3.2 Experimental procedure

The test carried out featured three milling tools, the designed tool for cooling called the copper milling tool, the previously designed tool for increasing the temperature in the inserts called idun milling tool, and Coro Mill 245 which is used as a reference, so therefore called the reference milling tool.

The idun milling tool is made in the material idun, which has a low thermal conductivity of 20 W/(m^◦C. Less heat is therefore conducted from the insert and the temperature increases in the insert. Both the reference milling tool and the idun tool uses a shim, but the idun milling tool uses a ceramic shim instead of a cemented carbide shim. The ceramic has a low thermal conductivity of 25 W/(m^◦C), to further increase the temperature in the inserts.

The tools are tested with two diﬀerent inserts. Insert 4330 which uses a chemical vapor deposition (CVD) coating and insert 1130 which uses physical vapor deposition (PVD) coating.

For the testing the cutting data is shown in Table 1. The cutting parameters used is meant to create as much heat as possible. The cutting parameters cuts at the edges of the work piece and moves towards the center of the work piece, with constant engagement with the work piece. When the cutting depth of 4 mm is cut from a layer of the work piece, a pass is made. For one insert tool life test for one milling tool, the total passes is 51, and the wear is recorded every 17 passes with a high-resolution camera. The radial and axial position for each insert (run-out) for the milling tool is measured before every machining. This test is repeated for every tool three times.

Table 1: The cutting data for testing a_p[mm] f_z[mm] v_c[m/min] a_e[mm] z_n[-] D_c[mm] κ [^◦]

4 0.2 250 51 6 80 45

3.2.1 Familiarize test

The machining time at Sandvik Coromant test facility is limited, so the testing must be proceeded eﬃciently. That means a familiarize test is carried out to resolve some questions for the testing.

The milling tool’s machining cutting parameters is programmed so that the milling tools are constantly engaged with the workpiece. The milling tool cuts the shoulder of the workpiece continuously until one layer of the workpiece is cut. The cutting parameters is tested for the familiarize test to see if it worked.

(16)

The insert is tested to see when it reaches a wear of 0.3 mm. To prepare for the insert tool life test. To test that, CoroMill 245 is machined in small increments, and measured for each increment.

The amount of passes needed to reach the 0.3 mm mark is tested. This test is repeated, but with longer machining time before measurement, so that more heat is generated, since steady state is achieved after longer machining times. These tests are repeated until the insert tool life for the inserts are known. So, when comparing the milling tools, a reasonable amount of wear is achieved and therefore measurable.

3.2.2 insert tool life test

The insert tool life test compares the inserts’ wear. The run-out for the tool is measured before machining, to ensure that the tools have similar run-out for both axial and radial measurements.

In the case of a too high run-out, interference with the wear results would occurre due to larger cutting forces on the inserts positioned further out.

Each milling tool is tested for three test repetitions with insert 4330, and since there is three tools, there is nine tests in total for insert 4330. For each test, the milling tool does a total of 51 passes, in intervals of 17 passes, so that the wear process can be analyzed if needed.

Insert 1130 is tested for one repetition with the idun milling tool and reference milling tool. Insert 1130 is measured after 20 passes.

High-resolution pictures is taken for all inserts after the 17 passes. Every insert got pictures taken from two diﬀerent angles, one of the ﬂank, and one of the rake face. To semi-automate the process of taking pictures, a robot arm is used. Placing the mill at a stand, the programmed robot arm takes pictures at the desired angles of the insert, always from the same distance, since the focal length of the camera didn’t change. The pictures are used to calculate the wear of the inserts. By taking a picture of a ruler, the distance for every pixel can be calculated, and therefore the wear is measurable.

3.3 Simulation Method

The cutting parameters used for the simulations is ANSYS since it’s used at Sandvik Coromant.

FEM is used for phase 1 of the product development. To validate if a concept can withstand the cutting forces. Temperature is also analyzed with FEM to see the temperature of the insert.

3.3.1 Strength Simulation

The cutting force, FC , is applied on the insert on the cutting zone, see Figure 6. The cutting force is divided into three components, where 70 % is in the tangential direction, FT , 20 % is in the radial direction, FR, and 10 % is in the axial direction, FA. With the angle directions θ and ϕ from the vertical respectively horizontal plane. The specific cutting force, k_cfor SS 2541 is defined as 2000 N/mm², which means that the cutting force will be the specific cutting force times the area of the cutting zone. The cutting force will then be distributed in tangential, radial and axial direction with the correct θ and ϕ.

The milling tool model for simulation is cut off at 60^◦at its symmetry, to decrease computing time. CAD models of the milling tool and insert are also simplified to further reduce computing time. The insert has a fine mesh, so is the mesh for the contact area where the insert is held in place. There is a coarser mesh for the rest of the milling tool.

(17)

Figure 6: The cutting force in red, and the cutting force components in orange

A fixed support is added to the areas where the tool is cut off for symmetry. The cutting force is applied on the cutting area. Cutting depth a_pmultiplied by the feed per tooth f_zgives the cutting area. The depth of cut radially on the insert is estimated to be two times the feed (D. Lundström, 2020).

By adding the cutting force to the cutting area results in a simulation of the stress and deformation on the tool when milling. If the stress results is within the yield strength of the material and has a reasonable deformation, the tool can be produced in that material.

3.3.2 Temperature

For the temperature simulation, it is critical to create a model that is structured so that the computing time is reasonable. This means that the mesh needs to be refined. The model used for the milling tool, insert, shim and screw is simplified for more robust surfaces. The CVD coating is also added to the simulation model with a thickness of 6 µm, (D. Lundström, 2020). Similarly to Section 3.3.1, the mesh for the milling tool, screw and shim is coarse. The mesh for the insert is fine, especially around the cutting zone because of its temperature gradient. The tool’s symmetry is also implemented so that only 60^◦of the model is used. The mesh of the milling tool with its components can be seen in Figure 7

Figure 7: The mesh of an idealised model of Coro Mill 245, the insert is black because of its ﬁne mesh.

Figure 8 shows the insert’s mesh at 0.4 mm, and the mesh around the cutting area at 0.05 mm.

To further reduce the computing time, the project schematic in ANSYS is set-up as following, see Figure 9. The time for the temperature to reach a temperature equilibrium during dry machining with a milling tool takes several minutes (V. Kalhori, 2001). It’s unrealistic to use transient calculations for a simulated time of several minutes. So steady-state is ﬁrstly used so

(18)

Figure 8: The mesh of an idealised model of Coro Mill 245 where the insert is shown.

that the system reaches the temperature equilibrium, and the results from the steady-state is then transferred to the transient simulation.

Figure 9: The results from the steady-state simulation is transferred to the transient simulation.

For the steady-state simulation, the average heat into the cutting zone over time is used. To derive the average heat ﬂow, the time in which the insert is engaged for one total rotation is needed. The mill can be represented by a unit circle to show when the insert is in engagement, see Figure 10.

The diameter, D_C, gives the radius, which is 40 mm, and α is the angle in which the milling tool is in engagement.

Figure 10: Unit circle that represent the milling tool in engagement

Given that the center of the milling tool is placed at the origo of the unit circle, x1 is the x-value of the point where the milling tool end its engagement. The variable ae is the radial depth of cut of the milling tool. The value of x1 is given through Equation 4.

(19)

x₁= DC /2 − ae (4)

With the unit circle, α can be calculated through the trigonometric function in Equation 5.

2 ∗ x1

α = arccos , (5)

D_C

The rotational speed of the milling tool is given through Equation 6 where n is the rotational speed in rpm

1000

n = VC ∗ . (6)

πDC_ap

Where V_Cis the cutting speed of the mill and DC_apis the maximum radial cutting depth of the milling tool given the cutting data. DC_apis given in Equation 7. It’s dependent on the entering angle, κ, and the cutting depth, ap.

2a_p

DC_ap= D_C+ . (7)

tan κ

Given the rotational speed, the time for one rotation t_rotationin seconds is shown in Equation 8 where

t_rotation= 1 , (8)

60 ∗ n

The time in witch the milling tool is in engagement for one rotation, tengagement, is then given from Equation 9

tengagement = trotation ∗ α . (9)

2π

The ratio between the tengagement and trotation is then used in Equation 10 to get the average heat ﬂow into the insert, Qsteadystate, which is the heat ﬂow used in the steady state simulations.

tengagement

Qsteadystate = Q ∗ . (10)

t_rotation

Where Q is the heat ﬂow into the insert.

Materials’ properties change with temperature. To simulate this, the CVD coating’s material data is dependent on temperature. The other materials don’t have temperature dependent material data in the simulation. The speciﬁc heat increases with the temperature, and the thermal conductivity is also varying.

The mesh is gradually refined until the results converges. The body, screw and shim has a redefined size of 2 mm, the insert has a mesh size of 0.4 mm and the CVD coating has a mesh size of 0.05 mm. Since the size difference between the insert and coating is too great, a contact sizing is applied between the contact region of insert and CVD coating. Contact sizing means that the mesh density

(20)

is ﬁner at the area of contact of the components, to provide a better distribution of contact pressure (S. Imaoka, 2008). The contact sizing is set at 0.08 mm. Figure 11 displays the meshing for the model. The coating only covers the insert partly, this is also for reducing computing time. If the coating covered the entire insert, the mesh size of 6 µm would be needed for entire component.

Figure 11: The mesh of a the insert, the mesh is set to be much ﬁner around the cutting area The convection for the circumference area is derived from (B. Ozerdem, 2000), which gives that ¨ the average Nusselt number N u for a rotating cylinder is according to Equation 11

N u = 0.318Re^0.571_r . (11)

The rotational Reynolds number Re_ris given through Equation 12. Where the angular velocity Ω in rad/s, D_cis the diameter of the milling tool, µ_airis the dynamic viscosity of air, and ρ_airis the density of air.

ΩD_c²ρair

Re_r= . (12)

2µ_air

By deriving h0 from Equation 13, the average outer convection will be given. The variable kair is the thermal conductance of air.

h₀D_m

N u = h0. (13)

k_air

Which gives that the outer convection of the milling tool h0 which is 32 W/m².

In the interface of the tool and the machine, a convection coefficient value is set to simulate the thermal conductance of the tool interface with the machine. The temperature is varying between the milling tool interface and the machine, to represent the heat transfer, it’s set as a convection coefficient similarly to (A. Malmelöv, A. Lundbäck, L. Lindgren, 2020). The value of the convection coefficient is derived through calibration with the testing temperature results.

The radial depth of the cutting area is given from the test results. An illustration of how the radial depth of cut is measured is seen in Figure 12. The deepest radially wear is measured, either at a comb crack or at the abrasive wear of the insert.

(21)

Figure 12: The radial measurement is illustrated in red.

The length of crater wear is also measured. It is measured on the rake face of the insert, which is shown in Figure 13.

Figure 13: The crater wear measurement is illustrated in red.

The program used to measure the radial depth of cut is ImageJ. Which counts the amount of pixels that makes up the distance of the measurement. The amount of pixels is then multiplied by the distance/pixel. The distance/pixel is calibrated by taking a picture of a ruler with the same camera settings as the settings used for the insert.

The radial depth of cut is measured for 1 repetition of test for all inserts of the mill. The radial depth diﬀers for the inserts because of the radial run-out of the milling tool. So the average radial depth is used for the simulation. The average radial depth of cut is 0.267 mm.

The unknown variables in the simulation of the temperature is the heat flow into the insert and the convection coefficient on the interface of the milling tool. The heat flow into the insert is between 50 watt and 120 watt (D. Lundström, 2020). To calibrate the values of the unknown variables, Coro Mill 245 is used because of its well defined material data. The heat flow is set in intervals of 10, and the convection coefficient is set accordingly to match the temperature points values given from the thermal camera for points [8] and [9] seen in Figure 14. Points [8] and [9] is chosen because it’s the greatest distance from the cutting zone, which means the temperature change over time is minimal for the chosen points.

Figure 14: The points where the temperature is measured

The same heat ﬂow into the insert, and convection coeﬃcients for the interface of the milling tool, is used for the copper milling tool and idun milling tool. The models for the idun tool and copper tool had the same parameters otherwise. The simulation results that correlated the best with the temperature measurement from the thermal camera is used.

(22)

With the tools calibrated, points [8] and [9] had both similar results to the temperature measurement of the testing, which can be seen in Table 2. But the temperature of the insert in the cutting zone fluctuates with several degrees dependent on the heat flow. The reliability of this method of temperature reading on the insert is inconsistent. Which should be taken into account when analyzing the data. But the temperature measurement will give an understanding of the temperature fluctuation over time for one rotation, and an overall estimation of the temperature.

Table 2: The simulated temperature compared to the measured temperature from the thermal camera.

Milling Tool Copper [^◦C] Reference [^◦C] Idun [^◦C]

Simulated Point [7] 101 146 152

Simulated Point [8] 102 156 161

Measured Point [7] 102 146 153

Measured Point [8] 101 154 165

(23)

4 Results

4.1 Product development

For the concept development there is several concepts that is considered and analyzed. All use the angular velocity of the milling tool, or try to increase the thermal conductance to remove heat from the insert.

Figure 15 displays the first concept, similar to (E. Uhlmann, H. Riemer, D. Schröter, F. Sammler, S. Richarz, 2017). This uses heat sink fins that are curved, so that the air inlet is parallel to the tangent of the circumference, to maximize the air flow into the heat sink. Since the heat sink is dependent on the mass flow of the medium that passes through the heat sink. The questions that remain with this concept is how much of the airflow is led into the inlets, The heat sink is placed far from the inserts, which makes it problematic to transfer the heat generated in the insert to be conducted to the heat sink efficiently. The tool will also become taller and heavier. The production is more difficult and costly compared to the other concepts. The plates needs to be curved, which means either a much longer production times, or a third party for manufacturing.

Both production are more costly and timely. This can be prevented by implementing a diﬀerent shape for the heat sink, which uses a more simple geometry in a production method perspective and still have the same, or more, surface area.

Figure 15: Concept 1, with curved heat sink fins angled parallel to the turbulante air flow The next concept is based on the previous concept. However, this concept incorporates columns to increase the turbulence of air flow with more inlets of air. The concept is shown in Figure 16.

However this concept still faces similar problems to concept 1.

Figure 16: Concept 2, heat sink with ﬁns but with several columns to increase airﬂow Another concept is to use a clamp to lock the insert in place, instead of a screw. This will result in more contact area compared to a screw, and therefore more thermal conductance. This concept

(24)

can be seen in Figure 17. There are turning tools that uses clamps to lock the inserts in place. This would increase the amount of heat conducted from the insert, it would conceptually act closer to the cutting zone, which would result in more heat removal. This concept leaves many uncertainties.

If the detail construction of the clamp would result in a clamp that can whithstand the cutting forces. If it can hold the insert with the ﬁne tolerances that is needed, so that the run-out is not eﬀected. The concept have too many risks in order to be developed within the timeframe that is given for this thesis.

Figure 17: Concept 3, using a clamp to increase thermal conductance

One of the concepts is to create a small wind tunnel to increase the forced convection on the insert.

This concept can be seen in Figure 18. The advantage of this concept is that it is applied close to the source of the heat generation. But the wind tunnel has some problems. It might not be possible to create an inlet that is suﬃcient. The angular velocity of the milling tool is low because the tests are carried out on steel, which means a lower cutting speed. The forced convection will be very limited. The windtunnel can also remove the possibility to implement a heat sink, which is more eﬃcient given the cutting data that is used.

Figure 18: Concept 4, implementing a wind tunnel

The pin-fin heat sink can be seen in Figure 19. This is an early version of the concept chosen for development. The advantage with this type of heat sink array is that it has an increased surface area since it uses cylindrical pins, which has a higher surface area per volume compared to regular fins. A pin heat sink is not as dependent on wind direction compared to fins. Regular fins needs the air flow direction to be at the tangents of the fins to be efficient. The pin-fin heat sink is easier to produce and it can be placed closer to the inserts. One of the negative aspects of using a bottom plate is that the heat sink is still placed a far distance from the inserts. This also results in a taller tool, facing the same problems as concept 1 and concept 2.

(25)

Figure 19: Concept 5, the concept used for detail construction

4.2 Detail Construction

The detail construction of the copper milling tool can be seen in Figure 20.

Figure 20: Detail design of the copper milling tool

The material used for the construction is a chromium copper alloy, SS 5716-20, which is a strong alloy with a thermal conductance of 322 W/mK. This high thermal conductance increases the heat transfer from the insert to the heat sink. This material is simulated for the cutting forces given. The stress results of the static structural analysis seen in Figure 21. As seen in the ﬁgure, the highest stress is 137.8 M P a and the yield strength of the copper alloy used is 400 M P a. The simulation also showed that the max value for deformation is at 0.021 mm, which won’t eﬀect the cutting performance of the milling tool.

(26)

Figure 21: The Von-Misses stress of the milling tool

Threaded M3 holes are placed on the milling tool, which are meant for the pins to be threaded on.

To decrease the distance between the heat sink and the inserts, instead of using a bottom plate as in concept 5, the pins are mounted directly upon the surface of the milling tool.

The distance of the holes/pins are influenced by (A. Bejan, A. Fowler, G. Stanesce, 1995), so that the heat sink is as efficient as possible. The conditions are different compared to the source, the source uses laminar flow when testing the distance between the pins. Heat sinks are rarely used with turbulent flow resulted from angular velocity. But the source is still used as a reference to distance the pins, so that the pins have a reasonable distance between them.

The rotational Reynolds number is relatively high, 16500, so the benefits of the splayed pin fin heat sink are not perfect. However, there are only benefits in using the splayed pin heat sink, since the air will pass through the heat sink more efficiently. So the design of the splayed heat sink is used. To create the splayed design, the threaded holes are angled. For production purposes, the pins closest to the axle are angled outwards, so that the holes can be threaded.

The pins are made of pure copper to have a higher thermal conductivity of 385 W/mK. At the top of the pins there is a slot for screwdriver so that the copper pin can be threaded onto the milling tool. There is also a diameter diﬀerence which can be seen in Figure 22. The diameter diﬀerence is to create a contact surface area between the milling tool and the pin, so that a clamping axial force can be created. This is to prevent the pins from unscrew themselves during the cutting process, due to vibrations and axial forces.

(27)

Figure 22: Detail construction of the copper milling tool assembly

The detail construction of the copper tool can be seen in Figure 23. The pins are mounted upon the milling tool with a screw driver. The copper pins has a diameter of 4 mm and has a smaller diameter of 3 mm where the pin is threaded. The copper pins are threaded onto the milling tool with thermal paste to create as high conductance between contact of the mill and pins. There is no shim used in the detail construction of the copper milling tool, so there is direct contact between the insert and the copper alloy. Thermal paste is also used at the contact between insert screw, insert and tool to further increase the conductance between the contacts.

Figure 23: Detail construction of the copper milling tool assembly

(28)

4.3 Temperature Simulations and Measurements

The computed temperature on the insert of the reference tool can be seen in Figure 24. The maximum temperature is marked with a red probe and a second probe is marked in blue. The blue probe is a distance of 1 mm from the maximum temperature. The blue probe is to see if there is diﬀerent temperature results at the given point.

Figure 24: The result of the FE-simulation of the reference milling tool, with a max temperature of 537 ^◦C, and only a distance of 1 mm from the max temp the temperature drops to 330 ^◦C The Figure 25 displays the temperature change over time for the red probe of the insert in the cutting zone, the red probe is shown in Figure 24. Note that the increase of the the temperature is when the insert is in engagement and the decline is when the insert gets out of engagement.

Figure 25: The Ansys calculations for the max temperature of the insert over time for one rotation of the tool, where orange is the reference tool, blue is idun and gray is copper.

Table 3 shows the maximum and minimum temperature for the red probe in Figure 24 for one rotation of the respective milling tool, and the temperature diﬀerence. The Average Temp in Table 3 is the average temperature of each milling tool for one rotation. The average temperature shows that: The idun milling tool is 3.1 % hotter than the reference tool and copper is 20 % colder than the reference tool.

(29)

Table 3: The maximum, minimum and average temperature for each milling tool for the red probe, and the diﬀerence between the max and min temperature is shown in the table.

Milling tool Copper [^◦C] Reference [^◦C] Idun [^◦C]

Max Temp 472 537 547

Min Temp 149 213 222

Temp Diﬀerence 323 324 325

Average Temp 257 321 331

Figure 26 shows the temperature change over time for the blue probe. A comparison of the maximum temperature at 18 ms for each probe, shows that there is a 200 ^◦C diﬀerence at the given time.

Figure 26: The Ansys calculations for the blue probe’s temperature of the insert over time for one rotation of the tool, where orange is the reference tool, blue is idun and gray is copper.

Table 3 shows the maximum temperature, minimum temperature, average temperature and temperature diﬀerence for the blue probe for one rotation of each milling tool. The average temperature for one rotation for the blue probe shows that the idun milling tool is 3.5 % hotter than the reference tool and the copper milling tool is 25 % colder than the reference tool.

Table 4: The maximum, minimum and average temperature for each milling tool for the blue probe, and the diﬀerence between the max and min temperature is shown in the table.

Milling tool Copper [^◦C] Reference [^◦C] Idun [^◦C]

Max Temp 265 329 339

Min Temp 148 212 221

Temp Diﬀerence 117 117 117

Average Temp 194 258 267

4.4 Wear

4.4.1 Wear test for insert 4330

Each tool had 3 repetitions of tests, where the wear is measured after 17 passes, 34 passes and 51 passes. All wear data can be seen in Appendix A-D. The wear is shown after 51 passes in the following ﬁgures and tables, because it’s easier to identify what type of wear that aﬀects the insert after 51 passes, and the wear is the closest to the insert tool life, which is a measured radial wear of 0.3 mm or above.

(30)

The wear types differ for the three milling tools. The reference tool shows the wear types crater wear, comb cracks and chipping. The flank side of the tool can be seen in Figure 27. The comb cracks are the cracks that are perpendicular to the edge line. All insert samples of the reference milling tool have been affected by comb cracks. This is predicted since it’s a thermo-mechanical wear and common in CVD coated inserts machining in steel. Chipping has also occurred in between the comb cracks for 16 out of 18 samples.

Figure 27: Reference tool wear sample: flank side, 1st repetition, insert 6, a wear of 0.176 mm The reference milling tools’ samples are also affected by crater wear, 7 out of 18 inserts have crater wear. One of those is shown in Figure 28. The crater wear can be seen at the corner of the insert, the same position as the maximum temperature is situated on the insert. A number of wear mechanisms can cause crater wear, but for this case it’s temperature related. It’s at the maximum temperature of the insert, and the amount of crater wear differs for the three milling tools, with more crater wear for higher temperatures during machining.

Figure 28: Reference tool wear sample: rake face, 1st repetition, insert 6, a wear of 0.176 mm Figure 29 shows a typical sample of the copper milling tool’s insert on the ﬂank side. There is comb cracks since there’s cracks that are perpendicular to the edge line. All tested inserts of the copper tool are aﬀected by comb cracks. Chipping in between the cracks has occurred for 7 out of a total of 18 samples. So cracks are less common for the copper tool compared to the reference tool.

Figure 29: Copper tool wear sample: ﬂank side, repetition 1, insert 6, a wear of 0.189 mm, both comb cracks and chipping can be seen

On the rake face of the copper tool’s insert, the same type of wear can be seen, Figure 30. No crater can be seen on this particular sample, only 2 of 18 inserts had crater wear. The samples that didn’t have crater wear didn’t show any signs of crater wear, it is only 2 exceptions, where the crater wear is minor. Which is far less compared to the reference milling tool.

(31)

Figure 30: Copper tool wear sample: rake face, repetition 1, insert 6, a wear of 0.189 mm, both comb cracks and chipping can be seen

Idun show comb crack wear for all inserts. However, compared to the reference milling tool and the copper milling tool, the idun milling tool has no chipping in between the cracks. One of the typical insert samples is shown in Figure 27.

Figure 31: Idun tool wear sample: ﬂank side, 2 repetitions, insert 5, a wear of 0.234 mm On the rake face of the idun tool’s insert, crater wear can be observed. 10 out of 18 inserts have crater wear, with the same severity as the reference milling tool, but it is more common that the insert have crater wear.

Figure 32: Idun tool wear sample: rake face, 51 passes, 2 repetitions, insert 5, a wear of 0.234 mm Table 5 shows the overall wear results of the testing with insert 4330. Average comb crack wear is the average comb crack wear after 51 passes for three repetitions for given tool. Average maximum crack wear is the wear measurement of the insert that had the most wear (since the milling tools were fully equipped), and the average is taken for the tree repetitions. Average crater wear is the average length of the crater wear, if no crater wear is found on a speciﬁc insert the crater wear is set to 0 mm for that insert.

(32)

Table 5: The wear for 4330

Milling tool Reference [mm] Copper [mm] Idun [mm]

Average comb crack wear 0.150 0.154 0.169

Average maximum comb crack wear 0.200 0.249 0.257

Average crater wear 0.37 0.023 0.455

The wear compared to the run-out can be seen in Figure 33. The radial run-out is compared to the wear of all inserts for testing of insert 4330. A trend line is set for the measurements, to see if the inserts with higher run-out had more wear. The graph shows a slight positive inclination, which means that the run-out had some eﬀect on the wear results.

Figure 33: The wear can be seen in the y-axis, and the run-out can be seen in the x-axis. The wear data points’ trend lines for each milling tool is also shown in the graph, note that the trend lines have a positive inclination for all tools

Table 6 shows the surface ﬁnish of the three milling tools after 51 passes of milling. The surface ﬁnish is similar for all three milling tools.

Table 6: The wear for 4330 Milling tool Reference [µm] Copper [µm] Idun [µm]

Surface ﬁnish 0.150 0.154 0.169

(33)

4.4.2 Wear test for insert 1130

The wear test for 1130 is done for one repetition for the reference milling tool and the idun milling tool. The insert samples are measured after 20 passes.

Figure 34 shows the typical wear for the reference milling tool on the ﬂank side. The wear type is clearly comb cracks with the cracks being perpendicular to the edge line. No chipping has occurred between the cracks of any sample.

Figure 34: Reference milling tool sample: insert 4, ﬂank side, a wear of 0.225 mm Figure 35 shows the same sample from the rake face of the insert, where the comb cracks are also apparent. No crater wear is shown for any sample of insert 1130, neither for the reference milling tool or idun milling tool.

Figure 35: Reference milling tool sample: insert 4, rake face, a wear of 0.225 mm

Figure 36 shows the typical wear for the idun milling tool on the ﬂank side. The wear type is the same as for the reference tool, only of the wear type comb cracks.

Figure 36: Idun milling tool sample: insert 5, ﬂank side, a wear of 0.188 mm Figure 37 shows the same sample from the rake face, which also shows the comb cracks.

(34)

Figure 37: Idun milling tool sample: insert 5, rake face, a wear of 0.188 mm

Table 7 shows the measured wear of insert 1130. Insert 1130 is tested with the idun milling tool and the reference milling tool for one repetition. The wear is measured after 20 passes. No correlation between temperature and insert wear is seen.

Table 7: The average wear of insert 1130 Milling Tool Idun [mm] Reference [mm]

Average comb crack 0.118 0.134

Max comb crack 0.188 0.238

Temperature effect on insert tool life in dry machining