Simulation Based Design for Large Module Gear Machining with Indexable Inserts

(1)

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Selection and peer-review under responsibility of the International Scientific Committee of “9th CIRP ICME Conference” doi: 10.1016/j.procir.2015.06.056

Procedia CIRP 33 ( 2015 ) 470 – 475

ScienceDirect

9th CIRP Conference on Intelligent Computation in Manufacturing Engineering

SIMULATION BASED DESIGN FOR LARGE MODULE GEAR

MACHINING WITH INDEXABLE INSERTS

Klocke, F.a, Brumm, M.a, Weber, G.a*

a_{Laboratory for machine tools and production engineering, Steinbachstraße 19, 52074 Aachen, Germany}

* Corresponding author. Tel.: +49-241-80-27366; fax: +49-241-80-22293 E-mail address: [email protected].

Abstract

Production processes for large module gears are characterized by small batch sizes and high workpiece costs. Therefore, the production of scrap parts cannot be accepted. Manufacturing processes are often not pushed to their limits so that the limits themselves are often not clearly known. To use the potentials of modern tool and machine tool concepts, a deeper theoretical understanding of the processes is necessary. This applies especially to modern indexable insert gear cutting tools, which gain importance in these processes.

In the past, gear cutting processes for large module gears have seldom been a subject of scientific research. Therefore, the design of these processes is largely based on experience. Especially, for the newer and more complex tool concepts using indexable inserts, a possibility to analyze processes theoretically is a key factor for deeper understanding of the processes and tool concepts.

In this paper the two most important processes for green machining of large module gears, gear hobbing and form milling, are analyzed using simulation and machining trials. For the process simulation, major changes to existing process simulation tools for gear machining processes were implemented. For suitable machining trials, model processes were developed for milling and hobbing. These allow an analysis of the wear behavior of the tools based on a reduced number of machined parts. In this paper the bridge between simulation and machining for large module gears will be drawn in terms of chip removal and deformation, wear behavior and tool design features. As final result a first method for a simulation based process design will be given.

Keywords: Gear; machining; wear; process design; process simulation

1. Introduction and challenge

For the production of large module gears tools based on indexable carbide inserts have been applied for several years. Due to high hardness and high thermal strength of the cutting material, these tools allow machining at higher cutting speeds than HSS tools and machining without coolant for large module gears.

These advantages have lead to a prevalence of these tools in the industry. The design of machining processes for large gears is mostly based on experience. The process design for new tool concepts is therefore often limited to an iterative procedure. This approach is not

usually applicable for small quantities of parts. Further on, the financial risk of a tool overload in these processes is high, as the prices of the tools themselves and the machined parts are high. Therefore, the limits of the applied processes are often unknown.

To exploit the potential of these processes more effectively, a base for the process design is necessary. This base is the objective of the project “Increase of productivity for green machining of large module gears with indexable carbide insert tools” (ProHM-WSP). Besides the theoretical analyses, the process is investigated using machining trials. In the comparison of tool performance with theoretical process values a possibility for a target-oriented process design will be

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

(2)

developed. In this report the focus lies on the form milling process with coolant. Theoretical results and results from machining trials are used to derive recommendations for a process design.

2. State of the art

For machining of large module cylindrical gears milling and hobbing are widely used due to their productivity. The applied tool concepts for these processes are shown in Fig. 1. For machining large module gears solid tools made of high-speed steel (HSS) were the dominating tool concept for many years. Due to its larger hardness and temperature strength, carbide has advantages over HSS in cutting speed and wear resistance. A production of solid carbide tools for these applications is not feasible due to large material and production cost, though. For some years, tools which are equipped with cutting inserts are increasingly used. These tools allow the use of carbide for large module gear production. These tool concepts are applied in gear hobbing and form milling processes. Due to the high temperature strength and hardness these tools are applied at higher cutting speeds than HSS tools. Additionally, dry processes are possible using these tool concepts.

Fig. 1 Tools for manufacturing large module gears

The process design for machining gears using gear hobbing is subject to different publications. Research on process and tool design for a single cut strategy in gear hobbing with solid carbide tools were investigated [1], [2], [3], [4]. For HSS tools the focus of research lies on the wear behavior and process design [5], [6], [7]. Investigations on surface defects and the chip flow are described by STUCKENBERG [8]. The focus in the

described investigations lies on gear hobbing of workpieces from automotive and smaller industrial applications with a module of mn < 6 mm. Gear hobbing in a second cut with PM-HSS tools for a part with a module mn = 11 mm was investigated [9]. Also work on specific measurement systems in the gear hobbing processes was carried out [10].

Machining with indexable inserts has been investigated in different publications. BOUZAKIS [11]

and KÖTTER [12] examined the influence of the cutting edge radius on the process. These publications mostly refer to turning and conventional milling processes.

A transfer of the previous results to machining of large module gears is not directly possible due to the differences in workpiece size, tool concepts and processes. The process design is therefore to large extends based on experience of the user, tool producer and machine tool supplier.

3. Research objective

The state of the art shows a need for research on the design process for machining processes for large module gears. This leads to experience based process design for these processes. Furthermore, the tool and workpiece cost in these processes is high, which leads to a high financial risk in experimental tests at high productivity.

In this paper a method for a process design for machining of large module gears is developed. This method uses theoretical investigations as well as cutting trials. For the theoretical investigations a penetration calculation is applied, which allows the calculation of the occurring undeformed chip geometries. From these geometries characteristic values are derived.

Due to high workpiece and tool cost, the cutting trials are carried out in model trials, which were developed in this project. These trials form identical chip geometries as the conventional process while the number of parts and cutting inserts is drastically reduced. In the cutting trials tool life, wear behavior and chip geometries are analyzed.

Fig. 2 Methodology for the investigation of gear machining processes

4. Numerical calculation of the chip geometries

The chip geometries in gear hobbing cannot be easily calculated due to the complex contact conditions. Therefore, a simulation model for these processes is necessary to determine local contact conditions and locally resolved characteristic values for the chip geometries. A similar method is necessary for an analysis of the form milling process.

Solid Tools

C arbide

Indexable Insert Tools

Material Advantages

Disadvan-tages Tool type

H igh profile accuracy

L ower invest per tool

E xperience in process design

H S S ; P M -H S S

L imited cutting speed

R econditioning external

H igh cutting speed

H igh wear resistance

D ry machining possible

C hange of inserts by user

H igh invest per tool

L ittle experience in process design

S ource: F ette S ource: S aacke

Objective

R ecommendations for the design of form milling processes for large module gears with cutting inserts

Process Simulation chip thickness (hcu) cutting length (lcu) C hip volume (V ) Model Trials VB L

(3)

4.1. Calculation method for machining of gears

The simulation of undeformed chip geometries gives the possibility to predict wear and design cutting processes [13]. For the calculation of the gear hobbing process the program SPARTApro was developed. This program is based on a penetration calculation. The workpiece is modelled using an equidistant set of transverse sections. The tool in these calculations is represented by a spline curve. The process kinematics are applied by a tool and workpiece movement, which represent complete relative movement of the blade and the workpiece in the process. The hob’s different teeth and their generating position are modeled by movements of one cutting edge.

In this calculation method chips are represented by chip thickness values, which are mapped over the tool rotational angle and the unrolled cutting edge of the tool profile. This set of data allows a representation of the chip geometry and the calculation of different characteristic values for the chip and tool load at different states of the process. The values maximum and average chip thickness, maximum, average and accumulated cutting length, specific chip volume and the number of cuts are determined. Additionally, the cutting force and the torques of the workpiece table and the spindle are calculated using the cutting force models of

GUTMANN [14] and BOUZAKIS [15]. A further

functionality is a variation calculation, which allows the process design based on different boundary conditions like the dimensions of the machine tool or maximum geometrical deviations of the workpiece flanks.

Fig. 3 Manufacturing analysis for gear hobbing SPARTApro

The simulation has been designed for solid gear hobs. This leads to limitations, which do not correspond to the degrees of freedom in tool design which can be used in a hob with indexable cutting inserts, Fig. 4.

In solid hobs the basic rack profile is represented by one single tooth in conventional tools or two teeth in roughing hobs. Unlike this, the basic rack profile can be divided into any number of different inserts in indexable insert hobs. This degree of freedom was not implemented in the simulation model. Therefore, the

restriction of the number of profiles in one tool was lifted.

Additionally, the input of cutter profiles had to be changed drastically. While one tooth in a solid hob represents either the complete profile of the basic rack or the tip section of it up to a certain profile height, the partitioning of the profile is not limited in indexable insert tools. Therefore, a calculation routine for the input of different blade profiles along one basic rack profile was implemented.

As a second limitation the number of the teeth over the tool circumference (number of gashes ni) in solid hobs is always whole-numbered. This is due to the fact, that the rake faces of one row of teeth are produced in the same grinding stoke. This limitation does not exist in tools with indexable inserts and had to be lifted in the simulation method.

Fig. 4 Adaptions to the calculation model

For the analysis of the form milling process, no calculation tool is available. Therefore, a penetration calculation method for these processes was designed. In this method first the penetration of a single collective of cutting inserts is calculated, leading to the input tooth gap design. Further on, the penetration of the cutter profiles of a second collective of cutting inserts is calculated, leading to the undeformed chip geometries in the process. These chip profiles can be evaluated in respect to the undeformed cutting length, chip thickness and chip volume.

5. Verification of the applied calculation models

In a further step cutting trials were carried out, which allow the verification of the adapted calculation methods. A workpiece from a wind turbine application with a module of mn = 16 mm, z2 = 35 teeth and an helix angle of β2 = 7.6° was machined using an indexable insert hob and a milling cutter. Both tools have an outer diameter of da0 = 300 mm.

The applied hob has Zeff = 17 effective tip cutters and 6 inserts per collective. The tool was applied in a climb cutting process with an axial feed on fa = 300 mm and vc = 110 m/min cutting speed. In Fig. 5 a photo of the

Input

Tool Geometry

Output

Workpiece Geometry Process

Chip Geometry Force Calculation Economic Feasability Study

hcu Fc/b vc fa

SPARTApro

vc R otation A ngle Fo rc e Adaptions Simulation Limitations S imulation of chip geometries for every blade of the tool in a laboratory version of the calculation core based on control files

G eneration and allocation of characteristic values on a blade possible in a M atlab routine

C alculation of the full cut of a process within a few minutes Number of Profiles M aximum three tool profiles in a blade group No limitation of the number of tool profiles in a blade group Profile Design Input of standard profiles or point clouds

Internal routine for the genertion of partial profiles Number of Groups L imitation to full blade groups on the tool circumference No limitation of the number of blade groups on the tool circumference

(4)

rake face of a left cutting insert in generating position -10 is compared to the calculated undeformed chip thickness. Additionally, the chip which was cut by the same cutting insert is compared to the calculated undeformed chip geometry.

Fig. 5 Comparison of simulated and cut chips in gear hobbing

In the middle of Fig. 5 the calculated undeformed chip thickness is represented by a black line along the cutting edge. The distance to the cutting edge represents the local maximum undeformed chip thickness. The resulting area corresponds well to a dark area on the rake face, which results from the contact with chip material in the hobbing process.

On the right side of the figure a photo of the machined chip and the calculated chip geometry are displayed. Both images show similar geometries with a flank area and a tip area, which are divided by a thinner section at the protuberance of the profile. The similarity of the results shows, that the simulation method for gear hobbing was successfully adapted to indexable insert hobs.

In a second set of cutting trials form milling processes were conducted and the chip geometries were compared to simulation results. For these trials a milling cutter with Zeff = 12 effective tip cutters was applied. Each collective consists of two inserts, which cut the one complete flank of the tooth profile and two inserts which only cut the tooth root. These tip inserts protrude over the flank inserts. The processes were conducted in a climb cutting with feed rates of fz = 0.6 mm, 0.8 mm and 0.95 mm.

The left side of Fig. 6, shows the order in which the different inserts engage in the process. For each of the chips the calculated undeformed geometry and a photo of the chip from the process at fz = 0.6 mm is shown. The triangular shape of the flank chips can be seen in simulation and practical result. The tip chips in the simulation and in the cutting trials have a constant width over the process. The calculated chip geometries correspond well with the chips from the machining trials.

A quantitative verification of the calculation results is given by a comparison of the chip mass of calculation

and practical results. For this comparison the volume of each simulated chip is multiplied by the density of the chip material. The results for each process and each cutting insert are given in the right side of Fig. 6. In addition, the weight of the chips of each insert was measured and plotted into the same graph. The values of the chip mass in simulation on cutting trial have a similar range and increase with the feed rate in the process. Further on, the mass of the measured flank chips of the left flank insert are larger than those of the right insert. The tip inserts show an opposite trend. This effect also results from the calculations. The simulation results for form milling correspond well to the practical results.

Fig. 6 Comparison of simulated and cut chips in form milling

The results of the presented calculation methods for machining of gears with indexable insert tools are according to the practical investigations and are therefore used for further analysis.

6. Results from cutting trials

In the conducted cutting trials in the profile milling process investigations concerning the tool wear at different process parameters were carried out. The number of cutting inserts in the tool was reduced to two collectives. This leads to a reduction of necessary workpieces for the trials. In these processes the cutting speed was varied as well as the maximum chip thickness. The resulting tool life and process parameters are shown in Fig. 7.

For all processes, which were carried out, a similar failure mode was observed. First, a coating failure at the rake face and later on in the areas of coating failures thermal cracks occur. Tool life end is reached by chipping of the cutting edge along the thermal cracks. The tool life at which each of these wear phenomena occur is shown in the column diagram. Trial four at vc = 85 m/min and hcu,max = 0.54 mm was not finished yet.

In general thermal cracks occurred earlier with increased cutting speed and increased chip thickness. A decrease of cutting speed with an increase in chip

Chip Geometries Rake Face Workpiece: module mn: 16 mm no. of teeth z2: 35 pressure angle α0: 20° helix angle β0: 7.62° diameter da2: 615.3 mm Tool: diameter da0: 300 mm no. of starts z0: 1

eff. teeth Zeff: 17

Process:

roughing

axial feed fa: 3.0 mm

cutting speed vc: 110 m/min

cutting depth T: 34.4 mm climb cutting generating position: -10

unrolled cutting edge

tool rotatoin 0 2 4 6 L F simuliert L F gemessen R F simuliert R F gemessen 0 0,6 1,2 1,8 2,4 0,6 0,7 0,8 0,9 1 K L F simuliert K L F gemessen K R F simuliert K R F gemessen

Comparision of chip geometries Coparison of Chip Mass

C h ip m a ss mcu [g ] feed fz[mm] C h ip m a ss mcu [g ] F I right (1. cut) T I left (2. cut) F I left (4. cut) T I right (3. cut) F lank Inserts Tip Inserts L F simulated L F measured R F simulated R F measured L F simulated L F measured R F simulated R F measured 2.4 1.8 1.2 0.6 0.6 0.7 0.8 0.9

(5)

thickness had a positive influence on wear progression. The influence of the cutting speed on the tool wear in the process is larger than that of the chip thickness.

Fig. 7 Tool life in the model trials for form milling

Besides the differences is tool life three observations were made, which will be further investigated. The first of these observations is the characteristic wear progression in these trials. A second observation is an increased wear of the tip inserts. The third observation is an uneven wear distribution between left and right cutting inserts.

First of the described observations the wear mechanism is considered in more detail. The progression of the thermal cracks during the process until chipping of the cutting edge is of special interest for the processes. In Fig. 8 the wear characteristics at tool life end are shown for the trial with an increased cutting speed. In the middle of the figure the rake face of a left tip insert is shown. A darker area next to the cutting edge can be seen, which corresponds to the contact zone with the chip. In some areas of this contact zone the coating has failed and substrate material is visible. In these areas thermal cracks starting from the cutting edge and reaching to the end of the contact zone are marked with letters from A to D. While the cracks have a similar length at the rake face, the view on the clearance face reveals differences. While crack “D” has a length of about l = 0.2 mm and is almost orthogonal to the cutting edge crack “A” with a length of l = 1.2 mm shows two changes in the direction of the crack progression. After l = 0.2 mm the crack changes to an angle of 45° and after a further progression of l = 0.3 mm the crack changes the direction of progression parallel to the cutting edge. A similar progression of thermal cracks was found at cutting inserts in different processes. This progression is based on the stress distribution in different distances from the cutting edge. While the area close to the rake face is dominated by a thermally induced alternating stress with a direction of principal stress parallel to the cutting edge [16], the temperature changes are reduced at further distance from the rake face. The thermally induced stress in machining are always superposed by mechanically induced stress. This

mechanically induced stress leads to transverse cracks, which are parallel to the cutting edge. At certain distances from the rake face the thermal load is no longer dominating the state of stress and the crack progression is dominated by mechanically induced stress. This type of crack progression leads to areas in which the cutting edge is destabilized and in which the occurrence of chipping is likely.

Fig. 8 Wear characteristics at tool life end in form milling

7. Comparison of experiment and simulation

The uneven distribution of wear between tip and flank inserts and between left and right inserts is approached by an analysis of the undeformed chip geometries. Fig. 9 shows the order of engagement of the inserts in the left side and the simulated maximum undeformed chip thickness and the simulated chip volume for different feed rates and types of inserts.

Fig. 9 Comparison of chip thickness and chip volume

The maximum undeformed chip thickness and the chip volume increase with the feed rate. The maximum chip thickness of the tip inserts and the flank inserts differ by a value of Δhcu = 60 μm. This is an indicator for increases load and an explanation for the increased wear of the tip cutters. The difference in chip thickness is due to the protruding tip inserts in the tool design.

The values of the chip volumes also differ largely between tip and flank inserts. The larger volume of the flank chips is due to the larger chip width. Additionally, differences between the left and right inserts in terms of

185 110 85 0 2 4 6 8 10 12 14 16 18 20 0,42 0,54 0,64 1.2 3.4 2,3 6.6

max. chip thickness hcu,max [mm]

tool life L/Z ef f [m] reference process T1; Zeff= 4 T2; Zeff= 4 trial running T4; Zeff= 4

coating failure thermal crack tool life end

6.7 2.3 16.8 3.4 14.5 T3; Zeff= 4 15.5 14.5 Workpiece: 18C rNiM o7-6 module mn= 16 mm no. of teeth z2= 35 pressure angle αn= 20° helix angle β2= 7.6° diameter da= 615.3 mm Tool: material H M K 30 coating (Ti,A l)N diameter da0= 300 mm

eff. teeth Zeff= 4

Process:

climb cutting, wet (oil)

cutting depth T = 34.4 mm _1.2 0.42 0.54 0.64 500 μm 200 μm A B C D A B C D 200 μm 200 μm Workpiece: 18C rNiM o7-6 module mn= 16 mm no. of teeth z2= 35 pressure angle αn= 20° helix angle β2= 7.6° diameter da= 615.3 mm Tool: material H M K 30 coating (Ti,A l)N diameter da0= 300 mm

eff. teeth Zeff= 4

Process:

climb cutting, wet (oil) cutting speed vc= 185 m/min

feed speed vf= 470 mm/min

feed rate fz= 0.60 mm chip thickness hcu = 0.42 mm cutting depth T = 34.4 mm 0,3 0,4 0,5 0,6 0,7 V P L inks K P R echts K P L inks V P R echts 0 200 400 600 800 0,6 0,7 0,8 0,9 1

Simulated Chip Geometries Chip Thickness and Chip Volume

feed rate fz[mm] ch ip v o lu m e Vcu [m m 3] m a x. c h ip t h ic kn e ss hcu,m a x [m m ] F I right (1. cut) hcu,max

TI left (2. cut) hcu,max

F I left (4. cut) hcu,max

TI right (3. cut) hcu,max 58% 60 μm F P L inks F P R echts F I left TI right TI left F I right 0.3 0.6 0.7 0.8 0.9 0.4 0.5 0.6 0.7

(6)

chip volume can be seen. The volume of the chips of the

left flank inserts is increased by ΔVcu Ĭ 120 mm3 in

contrast to the right flank inserts. A similar effect, can be observed for the tip inserts. These differences are an explanation for the uneven distribution of wear over the cutting inserts.

8. Further approach

With help of the introduced simulation models and the results from the practical and theoretical analysis the design of tools and processes can be based on deeper knowledge of the chip geometries and local loads.

A trial without tip inserts will be conducted to analyze possible influences of the chip geometries on wear behavior and chip flow. For a more detailed evaluation of influences of the process parameters and obstacles in the chip flow on the cutting force and temperature, simulation in a 2D-FE Model are planned. Further need for research lies in the design of the chip space.

Acknowledgements

The investigations described in the present paper were conducted as a part of a project (AiF-No. 17385 N) sponsored by German Machine Tool Builder’s Association (VDW) and the German Federation of Industrial Research Associations (AiF).

α0 Workpiece pressure angle

β0 Workpiece helix angle

da0 Tool outside diameter

da2 Workpiece outside diameter

fa Axial feed

fz Feed rate (per effective tooth)

hcu,max Maximum chip thickness

mcu Mass of chip mn Module T Cutting depth vc Cutting speed vf Feed velocity Vcu Chip volume

z0 Tool number of starts

z2 Workpiece number of teeth

Zeff Tool effective number of teeth

References

[1] Peiffer, K.: Untersuchung des Einsatzes von Hartmetall-werkzeugen beim Wälzfräsen. Diss. RWTH Aachen, 1986 [2] Knöppel, D.: Trockenbearbeitung beim

Hochgeschwindigkeits-wälzfräsen mit beschichteten Hartmetall-Werkzeugen. Dissertation RWTH Aachen, 1996

[3] Kobialka, K.: Prozessanalyse für das Trockenwälzfräsen mit Hartmetallwerkzeugen. Dissertation RWTH Aachen, 2002 [4] Winkel, O.: Steigerung der Leistungsfähigkeit von

Hartmetall-wälzfräsern durch eine optimierte Werkzeuggestaltung. Dissertation RWTH Aachen, 2005

[5] Lux, S.: Einfluß von Oberflächenstrukturen auf den Verschleiß von Verzahnwerk-zeugen aus Schnellarbeitsstahl. Dissertation RWTH Aachen, 1997

[6] Herzhoff, P.: Trockenbearbeitung und Minimalmengen-schmierung beim Wälzfräsen mit HSS-Werkzeugen. Diss. Universität Magdeburg, 2003

[7] Hipke, M.: Wälzfräsen mit pulvermetallurgisch hergestelltem Schnellarbeitsstahl. Dissertation, Magdeburg, 2011 [8] Stuckenberg, A.: Oberflächendefekte beim Trockenwälzfräsen.

Diss. RWTH Aachen, 2014

[9] Klocke, F.; Gorgels, C.; Schalaster, R.: Wälzfräsen im zweiten Schnitt. Abschlussbericht zum Vorhaben Nr. 576 der FVA. FVA-Heft Nr. 955, Frankfurt, 2010

[10] Frohmüller, R.; Beutner, M.; Köchig, M.; Karpuschewski, B.: High speed temperature measurement in gear hobbing: Part I – design, concept and physical operation mode of the infrared-camera. Prod. Eng. Res. Devel. (2014) 8:73–79

[11] Bouzakis, K.-D.; et al.: Effect of the Cutting Edge Radius and its Manufacturing Procedure on the Milling Performance of PVD Coated Cemented Carbide Inserts. In: Annals of the CIRP 51, 2002, p. 61-64

[12] Kötter, D.: Herstellung von Schneidkantenverrundungen und deren Einfluss auf das Einsatzverhalten von Zerspanwerkzeugen. Dissertation, Universität Dortmund, 2006

[13] Brecher, C.; Klocke, F.; Brumm, M.; et al.: Simulation based model for tool life prediction in bevel gear cutting. Prod. Eng. Res. Devel. (2013) 7:223–231.

[14] Gutmann, P.: Zerspankraftberechnung beim Wälzfräsen. Dissertation, RWTH Aachen, 1988.

[15] Bouzakis,K.-D.: Konzept und technologische Grundlagen zur automatisierten Erstellung optimaler Bearbeitungsdaten beim Wälzfräsen. Habilitation, RWTH Aachen, 1980

[16] Vieregge, G.: Stahleisen-Bücher, Band 16: Zerspanung der Eisenwerkstoffe. Düsseldorf: Stahleisen, 1970