Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 33 (2015) 484 – 489

9th CIRP Conference on Intelligent Computation in Manufacturing Engineering - CIRP ICME '14

Design of Gear Hobbing Processes Using Simulations and Empirical Data

C. Brechera, M. Brumma, M. Krömera*
a
Laboratory for Machine Tools and Production Engineering, Steinbachstraße 19, 52074 Aachen, Germany
* Corresponding author. Tel.: +49-241-80-28295; fax: +49-241-80-22293. E-mail address: m.kroemer@wzl.rwth-aachen.de

Abstract

Gear Hobbing is one of the most productive manufacturing processes for pre-machining cylindrical gears. The process design as well as the
tool selection is often based on experience or is limited to an iterative procedure. Existing methods for process and tool design are limited
regarding the size of the gears and the process parameters.
The objective presented in this paper is to support the process design by suggesting process parameters. To achieve this goal, a simulation
for continuous gear hobbing was developed. By calculating planar intersections of transverse sections of both gear and tool, the generated chip
geometries are determined. Due to the general approach of positioning tool and workpiece in the program, all continuous processes with
defined cutting edges can be simulated. The generated chip geometries are analyzed and characteristic values for each position of the tool are
defined. Beside the chip geometries, other parameters such as working areas and the length of axis movements are calculated too.
Since the simulation process is time-consuming, it is not possible to calculate each hobbing process for different designs. Also, the
simulation program cannot be implemented into the machine control due to the needed computing capacity. Thus, a large amount of hobbing
processes with varying gear geometries as well as different tools with the corresponding profiles are calculated in advance. By the use of
regression analysis, the results of these variations are transferred to approximation formulas afterwards, which are easy to calculate and to
implement into other software products.
To support the tool and process design, existing hobbing processes will be simulated with the help of the developed manufacturing
simulation and the results will be stored in a database. By comparing the results of the approximation formulas with the values in the database,
it is possible to evaluate a given process.
© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
© 2014 The Authors. Published by Elsevier B.V.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review
Selection andunder responsibility
peer-review of the International
under responsibility Scientific Committee
of the International of “9th CIRP
Scientific Committee ICME
of “9th Conference".
CIRP ICME Conference”
Keywords: Machining; Gear; Hobbing; Process; Design; Expert System; Simulation

1. Introduction dependable method for calculating the maximally occurring

For designing gear hobbing processes, certain values are
necessary that can be compared and determined
unambiguously. An established value for hobbing is the
maximum chip thickness. Determining the chip thickness can
be conducted according to different methods and formulas. A
common and industrially as well as scientifically established
way is the approximation formula for the maximum chip
thickness according to HOFFMEISTER [18]. The approximation
formula represents a simplified calculation that is based on
empirical studies up to the module of mn = 4 mm.
Investigations have shown that for gears of larger modules,
the results according to HOFFMEISTER deviate from the actual
chip thickness and underrate them [10].
To define a reliable process design of large module gears, a
chip thickness in a process is necessary. For this purpose, a
new formula for calculating the maximum chip thickness up
to modules of mn = 30 mm will be created. In order to achieve
this objective, a variation model will be set up. This allows
determining approximation functions for different
characteristic values by means of results of a penetration
calculation and a regression analysis. With the help of this
method, subsequently a function for calculating the maximum
chip thickness for gear hobbing can be developed.
2. State of the Art
For turning or milling processes with defined cutting
edges, the process design and the cutting parameters are most
likely specified based on empirical knowledge, structured in

2212-8271 © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(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.2015.06.059
 C. Brecher et al. / Procedia CIRP 33 (2015) 484 – 489 485

form of charts. These charts contain axial feeds and cutting wear. MUNDT [MUND92] also developed a formula for
speeds for different workpiece and tool material combinations calculating the average chip thickness. His approximation
and are provided by the tool manufactures. The parameters in formula is based on the results of a geometrical penetration
these charts are based on the chip thicknesses. Fig. 1 shows calculation. In recent past, HIPKE [17] developed a method for
the maximum chip thicknesses hcu,max occurring in the turning, calculating process parameters for HSS hobs. The method is
milling and gear hobbing process. Because of the continuous based on the average chip thickness and cutting length.
chip formation on one single cutting edge, the calculation of Altogether, the module range in these mentioned studies is
the maximum chip thickness in the turning process is trivial. mn < 10 mm. While HOFFMEISTER and HIPKE used gears up to
The chip thickness depends on axial feed fa, cutting speed vc mn = 4 mm, the formulas of MUNDT include gears with
and a tool angle. For milling processes, the chip thickness hcu
mn = 10 mm.
varies along the cutting length. Although the milling process
Because none of theses studies provide a comprehensive
has a discontinuous chip formation, it is possible to calculate
method for the full module range of currently used gears, the
the maximum chip thickness with the help of analytic
formulas. In gear hobbing this is possible. As it is shown in alternative is to design a process by the maximum chip
the right column in Fig. 1, the chip formation varies with the thickness as well as experience. On this basis, an objective
cutting length. Because of the rolling process of gear hobbing, process design is nearly impossible.
the chip formation takes place at different cutting edge
sections. Thus, hcu varies with the profile and for each single 2.2. Calculation Methods for the Chip Load in Gear Hobbing
blade. Also, the chip geometries are highly depending on the
superposition of the single cuts. Due to these unique and To calculate the chip thickness in gear hobbing, different
complex kinematic conditions, a calculation based on analytic methods can be used. A simple and effective option is the
formulas of the maximum chip thickness is not possible. maximum chip thickness according to HOFFMEISTER, (1).
3 ( 0.015˜ x p )
˜E 0 0.542)
Turning Milling Gear Hobbing h1,max 4.9 ˜ mn ˜ z 2 (9.25˜10 ˜ e ( 0.015˜E0 ) ˜ e
3
( 8.25˜10 ˜E 0 0.225) 0.877 0.511 0.319 (1)
§ d · §n · § f · § T ·
˜ ¨¨ a 0 ¸¸ ˜ ¨¨ i 0 ¸¸ ˜ ¨¨ a ¸¸ ˜ ¨¨ ¸¸
© 2 ˜ mn ¹ © z0 ¹ © mn ¹ © mn ¹
hcu,max

hcu,max HOFFMEISTER approximates the real maximum chip

thickness by exponential and potential functions with
C hip formation continuous discontinuous discontinuous influence factors of tool, workpiece and process. These
C hip geometrie:
ƒ C utting edge constant constant variable factors are the module mn, the number of workpiece teeth z2,
ƒ B lades
ƒ C utting path
-
constant
constant
variable
variable
variable
the helix angle β, addendum modification factor xp, the axial
C hip thickness hcu,max = f(fa, v c ) hcu,max = f(fa, nz, v c ) hcu,max = f(…) feed fa, the cutting depth T as well as the number of gaps ni0
© WZL and the number of threads z0. The tool profile with pressure
angle and the tool tip radius had been neglected in the
Fig. 1. Comparison of chip formation in various cutting processes investigations.
Another method for calculating the chip thicknesses is
using a process simulation software such as SPARTAPRO.
2.1. Characteristic Values in Gear Hobbing
The software was developed at the WZL of the RWTH
Aachen University, Fig. 2 [8].
Gear hobbing is the main process for manufacturing
cylindrical gears [3]. However, designing a productive and
stable process poses still a challenge. In the past, there were
many studies analyzing the influence of different
characteristic values according the tool life and wear
behavior. Research was done with carbide tools [22, 18, 23,
13, 14, 11,24] as well as high speed steel hobs (PM-HSS) [19,
1, 7, 16, 15, 17, 21]. HOFFMEISTER [18] developed an
approximation formula for the maximum cutting length lmax
and total cutting length l of the process. He also developed
formulas for calculating the maximum chip thickness at the
profile tip hcu,max and an average chip thickness hav. Especially
the approximation formula for the maximum chip load is still © WZL
widely established in industrial and scientific environment.
BOUZAKIS [1] conducted extensive research regarding the Fig. 2. Features of SPARTApro
chip formation and the effects to tool wear. For this,
BOUZAKIS classified the three flank chips and developed a
Input parameters are the geometrical information of the
mathematical model for calculating the corresponding tool workpiece like module, number of teeth and outside diameter
486 C. Brecher et al. / Procedia CIRP 33 (2015) 484 – 489

as well as the tool data and its profile geometry. With the 4. Designing a Variant Calculation
given axial feed, a penetration calculation is executed and the
undeformed chip geometries occurring in the hobbing process The chip formation in gear hobbing processes depends on a
are determined. Afterwards, these geometries are analyzed large amount of influencing factors. Fig. 4 shows these
and characteristic values such as the maximum and average influences categorized into workpiece, the tool and the
chip thickness hcu,max and hcu,av, the specific chip volume Vʹ process. The highlighted geometrical influences can be
and the maximum and average cutting length lmax and lav are considered in the described penetration calculation
calculated. The values are then displayed along the unrolled SPARTAPRO. The bold factors in Fig. 4 influence the chip
cutting edge as well as maximum and average values for the thicknesses in the hobbing process directly and are
whole process. Besides these values, SPARTAPRO is capable
subsequently discussed in detail.
of calculating the cutting forces of the hobbing process and
the profitability. Workpiece Tool Process
Because HOFFMEISTER only used gears with a module
Part radius ra2 Tool radius ra0 Feed fa
mn < 4 mm, it is questionable to use the approximation Module m2 Number of gaps ni0 Plunging depth T
Helix angle β2 Number of threads z0
formula and transferring it to large gears up to module

Geometrical
Pressure angle α2 Lead angle J0 Cutting process

Influence
Addendum Protuberance P rP 0  C limb / conventional
mn = 30 mm. Thus, a comparison of the maximum chip modification
Width
x
b
Tip radius
Pressure angle
σa0
αP 0
cutting
 Identical / opposite lead
thickness according to HOFFMEISTER and SPARTAPRO was G ap angle
C learance angle
JN
α
orientation of part and
tool
conducted [8]. As it is shown in Fig. 3, the results of these R ake angle J S hift strategy
C utting velocity
sH
vc
two methods differ widely with increasing modules. The M aterial M anufacturing tolerances M achine / setting

simulation results are constantly higher than the results by Type C ooling agent

HOFFMEISTER and at some parts the difference of the M aterial E nvironment

calculated and simulated chip thickness is 55%. This shows

Chip formation
that the approximation formula according to HOFFMEISTER
underestimates the tool load for large module gears and © WZL

therefore is not the best option to use in a process design.

Fig. 4. Factors influencing the chip formation
Workpiece:
M odule mn: 1 - 30 mm Influence of the module on the chip thickness
Number of teeth z 2: 35 1.0
P ressure angle α2 : 20° For analyzing the maxima for different characteristic
max. Chip thickness hcu,max [mm]

H elix angle β2 : 0°
D iameter da2: (z 2 + 2)∙mn values, the gear width is often of little interest. This is because
Tool:
B asic rack acc. D IN 3972 IV
in the run-in and the overrun of the tool chip thicknesses,
D iameter da0: (1.2∙mn + 3.5)∙ni
Number of threads z 0: 1, right 0.5 cutting lengths and other values increase respectively
Number of gaps ni: 10 decrease. In the full-cut section of the workpiece, the
Process:
A xial feed fa: 2.0 mm characteristic values are nearly constant and have the highest
P lunging depth T: haP 0 + mn
C limb process
S PA R TA pro
H offmeister values.
0
5 10 15 20 25 30
To examine the influence of the pressure angle,
Module mn [mm]
preliminary investigations on the profile addendum and the
© WZL
generating addendum modification were preformed in
advance. While the cutting depth was kept constant, in each
Fig. 3. Comparison of approximate formula and simulation according to simulation the profile angle was varied between αn = 10°,
different modules [8] αn = 20° and αn = 30°. It shows that this modification has no
effect on the resulting maximum chip thickness. Furthermore,
the addendum of the tool profile was also changed in three
3. Research objective steps. This change leads to a varying tooth width of the gear,
but also has no effect on the maximum chip thickness.
The state of the art shows the need for an approximation Because the generating addendum modification factor also
formula for the maximum chip thickness in gear hobbing changes by varying the addendum height, this factor has no
which can be used for large module gears. So far, the process effect on the maximum chip thickness as well.
parameters for industrial application of the gear hobbing
process are chosen based on experience. By developing 5. Developing of Functions describing Characteristic
formulas for calculating the chip thicknesses, it is possible to Values
create a database for tool and workpiece materials with the
corresponding process parameters to support the process and After gear width, pressure angle and the addendum
tool design. modification could be excluded, as they have no effect on the
The objective of this paper is to illustrate a method chip thickness, the parameters listed in Table 1. are varied
determining approximation formulas for characteristic values. throughout the specified range. For each parameter up to
This method will be afterwards used to generate an seven incremental steps are used. Furthermore, the process is
approximation formula for the maximum chip thickness. simulated for climb as well as conventional cutting. A design
of experiments (DOE) method is used to reduce the number of
 C. Brecher et al. / Procedia CIRP 33 (2015) 484 – 489 487

necessary simulations. Because of the multifactorial

influences, a D-optimal design is used. This reduces the
number of simulations to 77 for each cutting strategy (climb
and conventional cutting).
Table 1. Variation Range of the influence factors
Parameter Symbol Range Unit

Module mn 1 – 30 mm
Number of teeth Z 10 – 120 -
Helix angle Β -40 – 40 °
Outside diameter of tool da 50 - 380 mm © WZL

Number of gaps ni0 8 – 23 - Fig. 5. Transfer functions of the variation parameters

Number of threads z0 1–5
Profile edge radius factor σaP0 0.1 – 0.5
Cutting depth factor T* 0.1 – 1
Axial feed fa 0.5 – 6
Climb / conventional cutting - A or B
For evaluating the resulting characteristic values, it is
possible to use several regression functions. The evaluating
software used in this paper provides a quadratic, cubic and
biquadratic regression model. To compare the different
approaches, the coefficient of determination R² is used [6].
The coefficient R² can vary between 0 < R² <100%. It
indicates the fitting of the statistical model to the original data
points. In general a R² > 90% is deemed to be good [12]. In
this study a quadratic model is used because it fits the transfer
functions best. The coefficient of determination for this
approximation function is R² = 94%.
In Fig. 5, the function for each parameter is shown. It is
striking that the function curve for the module is decreasing
with higher modules. Because this is not plausible, the
function is manually adjusted and reintegrated into the
approximation model. Furthermore, an analysis of the curves
also shows that the profile edge radius σa0 and the helix
angle β2 of the gear do not have a significant effect on the
resulting maximum chip thicknesses and therefore both can be
neglected in the model. In addition, the influence of the
number of gaps ni0 and threads i0 can be condensed to one
parameter, the effective number of gaps i0. Because the
regression functions of the tool diameter da0, effective number
of gaps i0 and cutting depth T are nearly linear, these
functions are linearized. Due to these changes in the
regression model, the approximation formula is now less
complex but the coefficient of determination only deteriorates
by ΔR² = 0.4%.
-
-
The resulting approximation formula is shown in (2). To
- improve the coefficient of determination, a logarithmic
mm transformation of the target value, the maximum chip
thickness, was done. This transformation is a common method
- to improve the R². Because of the described modifications to
the original approximation model, the resulting equation is not
more complex than the equation to HOFFMEISTER.
hcu ,max ex
with
x 1.446
 2.78 ˜10 3 ˜ d a 0
 0.065 ˜ ni 0 / z 0 (2)
 0.0186 ˜ T
 0.122 ˜ mn  2.24 ˜10 3 ˜ mn 2
 0.02621 ˜ z 2  0.123 ˜10 3 ˜ z 2 2
 0.6447 ˜ f a  0.07486 ˜ f a 2
The validation of the approximation formula for the
characteristic value can be done with the help of the process
simulation SPARTAPRO. During the experimental design,
three points of stability were defined by the DOE.
Furthermore, the data of two very different, real cylindrical
gears are available for validation. Fig. 6 shows the gear, tool
and process parameters of these five gears. At the bottom of
Fig. 6, the resulting maximum chip thicknesses according to
the simulation SPARTAPRO, HOFFMEISTER and the new
approximation formula are given. This shows that the chip
thickness calculated with SPARTAPRO and the developed
formula are in four of these five cases nearly the same.
Especially the large gear with a module of mn = 16 mm shows
the advantage of the new formula compared to HOFFMEISTER.
While the calculated chip thickness by HOFFMEISTER is 25%
lower than the simulated one, the value calculated by the new
approximation formula is only 3% higher than the simulation
and therefore nearly the same.
488 C. Brecher et al. / Procedia CIRP 33 (2015) 484 – 489

will not lead to an optimal process but it will result in most

cases in a stable process which can be optimized.

7. Summary and Outlook

In practice, designing the maximum axial feed for hobbing

Fig. 6. Comparison of calculation methods using different gear geometries
A similar pattern emerges by comparing the results of the
second and third points of stability. Also, in these cases the
deviations of the developed formula and SPARTAPRO are
minimal (1.5% and 4%) while the chip thickness according to
HOFFMEISTER differs to the simulation by 20% and 28%. This
improvement of the results calculating the maximum chip
thickness is mainly attributable to the fact that the module
range of the new formula is wider than HOFFMEISTER. Only
the first point of stability does not match to the good results of
the other cases. While the calculated chip thickness by
HOFFMEISTER and the new formula are nearly the same, the
thickness by simulation is 20% higher.
Beside these five gear sets, also the known module
variation of Fig. 3 was done. The curves of the maximum chip
thickness for the three calculation methods are shown in
Fig. 7. Here again, the new formula matches the simulation
results of SPARTAPRO well.
Workpiece:
M odule mn: 1 - 30 mm
Number of teeth
P ressure A ngle
z 2:
α2 :
35
20°
1.0
max. chip thickness hcu,max [mm]
processes is often done by means of the maximum chip load
according to HOFFMEISTER. For developing the approximation
formula, however, only gears up to a module of mn = 4 mm
have been regarded and extrapolating the results on large
module is only possible by inducing deviations.
© WZL For this reason, a new formula for determining the
maximum chip load for hobbing processes was developed.
Based on a regression analysis, an approximation function for
gears up to a module of mn = 30 mm has been set up.
This newly developed regression formula according to
SPARTAPRO regards influencing factors from part, tool and
process, as does the approximation formula according to
HOFFMEISTER. Values regarding the workpiece are module mn
and the number of teeth z2. The outside diameter da0 and the
effective number of gaps i0 as a result from the quotient of
number of gaps ni0 and the number of threads z0 are influences
of the tool. Only the plunging depth T and the axial feed fa
have to be extracted from the kinematics for calculating the
necessary chip thickness. Contrary to the approximation
formula according to HOFFMEISTER, the tool helix angle as
well as the generating addendum modification are disregarded
for the developed regression formula, as considering these
factors did not cause an improvement of the results. Also,
considering the edge radius of the tool could not contribute in
a decisive improvement of calculating the maximum chip
thickness, contrary to initial assumptions. Thus, the newly
developed approximation formula does not need more
influencing factors than calculating the chip thickness
Comparison of the calculating methods according to HOFFMEISTER, which makes the complexity of
the developed approximation formula and HOFFMEISTER
comparable.

H elix angle β2 : 0°
D iameter da2: (z 2 + 2)∙mn
Tool: By means of five example gears, the new approximation
B asic rack acc. D IN 3972 IV
D iameter da0: (1.2∙mn + 3.5)∙ni formula according to SPARTAPRO has been validated. For
0.5
Number of threads z 0: 1, right
Number of gaps ni: 10 this, the maximum chip thicknesses have been calculated by
Process: the simulation, by the approximation formula according to
A xial feed fa: 2.0 mm S PA R TA pro
P lunging depth T: haP 0 + mn H offmeister HOFFMEISTER as well as by the new approximation formula. It
C limb cutting New formula
0 was shown, that in four of five cases the maximally occurring
5 10 15 20
Module mn [mm]
25 30
chip thickness in the process corresponds almost exactly to
the ones of the simulation. In one case, the new formula
© WZL provides a comparable result to HOFFMEISTER which,
however, underrates the maximum chip thickness by nearly
Fig. 7. Comparison of calculation methods based on a module variation 20%. Additional to these five examples, it has been shown by
varying a gear in the module range of mn = 1 mm to
mn = 30 mm that the newly developed approximation formula
6. Process Design provides good results within the entire range.
For future investigations, it is possible to develop
This developed approximation formula for the maximum approximation formulas for further characteristic values like
chip thickness can now be used to calculate existing hobbing the average chip thickness with the help of the presented
processes. By setting up a database with the calculated chip method. Also, calculating the axial feed with the help of the
thickness, the workpiece and tool material as well as the used presented approximation at a given geometry and maximal
axial feed and cutting speed the existing processes can be tolerable chip thicknesses is possible. For this, however, the
evaluated and new processes can be designed according to knowledge of the maximally possible chip thickness of a
same material and chip thickness. It is clear that this approach hobbing process is necessary.
 C. Brecher et al. / Procedia CIRP 33 (2015) 484 – 489 489

Acknowledgements [14] Kobialka, K.: Prozessanalyse für das Trockenwälzfräsen mit

Hartmetallwerkzeugen. Diss. RWTH Aachen, 2002
[15] Lux, S.: Einfluß von Oberflächenstrukturen auf den Verschleiß von
αn Pressure angle Verzahnwerkzeugen aus Schnellarbeitsstahl. Diss. RWTH Aachen, 1997
[16] Mundt, A.: Modell zur rechnerischen Standzeitbestimmung beim
α0 Profile angle of the tool Wälzfräsen. Diss. RWTH Aachen, 1992
[17] Hipke, M.: Wälzfräsen mit pulvermetallurgisch hergestelltem
β0 Helix angle of the workpiece Schnellarbeitsstahl. Dissertation Otto-von-Guericke-Universität
Magdeburg, 2012
da0 Outside diameter of the tool [18] Hoffmeister, B.: Über den Verschleiß am Wälzfräser. Diss. RWTH
Aachen, 1970
σa0 Profile edge radius [19] Joppa, K.: Leitungssteigerung beim Wälzfräsen mit Schnellarbeitsstahl
fa Axial feed durch Analyse, Beurteilung und Beeinflussung des Zerspanprozesses.
Diss. RWTH Aachen, 1977
hav. Average chip thickness [20] Scheffler, E.: Statistische Versuchsplanung und –auswertung. Deutscher
Verlag für Grundstoffindustrie Stuttgart, 1997
hcu,max Maximum chip thickness [21] Schalaster, R.: Optimierung des Fertigwälzfräsens von Verzahnungen.
Diss. RWTH Aachen 2012
lav. Average cutting length [22] Sulzer, G.: Leistungssteigerung bei der Zylinderradherstellung durch
genaue Erfassung der Zerspankinematik. Diss. RWTH Aachen, 1973
lmax Maximum cutting length [23] Venohr, G.: Beitrag zum Einsatz von Hartmetallwerkzeugen zum
Wälzfräsen. Diss. RWTH Aachen, 1985
mn Module [24] Winkel, O.: Steigerung der Leistungsfähigkeit von Hartmetall-
wälzfräsern durch eine optimierte Werkzeuggestaltung. Diss. RWTH
ni0 Number of threads (tool) Aachen, 2005
[25] Ziegler, K.: Untersuchung der Hauptschnittkraft beim Wälzfräsen von
R² Coefficient of determination Stirnrädern. Diss. RWTH Aachen, 1967
T Cutting depth
xp Generating addendum modification
z0 Number of threads (tool)
z2 Number of teeth

References
[1] 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
[2] Bouzakis, K.-D.; Lili, E.; Michailidis, N.; Friderikos, O.: Manufacturing
of Cylindrical Gears by Generating Cutting Processes: A Critical
Synthesis of Analysis Methods. CIRP Annals - Manufacturing
Technology 57 (2008) No. 2, pp. 676-696
[3] Bouzakis, K.-D.: Konzept und technologische Grundlagen zur
automatisierten Erstellung optimaler Bearbeitungsdaten beim
Wälzfräsen. Habilitation, RWTH Aachen, 1980
[4] CRGraph: Versuchsmethoden. DoE und Datenanalyse. www.crgraph.de
[Stand: 18.03.2014]
[5] DIN 3972: Bezugsprofile von Verzahnwerkzeugen für Evelventen-
Verzahnungen nach DIN 867 Berlin, 1952
[6] Fahrmeir, L.; Kneib, T.; Lang, S.: Regression: ;Modelle, Methoden und
Anwendungen. Springer Verlag, Berlin, 2009
[7] Kauven, R.: Wälzfräsen mit Titannitrid-beschichteten HSS-Werkzeugen.
Diss. RWTH Aachen, 1987
[8] Weck M; Winter W; Klocke F; Winkel O: Analysis of Gear Hobbing
Processes by Manufacturing Simulation. In: Wissenschaftliche
Gesellschaft für Produktionstechnik (WGP). Production Engineering:
Research and Development. Volume X, Issue 1. WGP Annals, WGP
Eigendruck, Berlin, 2003.
[9] Klocke, F.; König, W.: Fertigungsverfahren – Drehen, Fräsen, Bohren,
Band 1. 8. Aufl. Springer. Berlin, 2008
[10] Klocke, F.; Brumm, M.; Weber, G.: Prozesssimulation für
Wendeschneidplatten-Wälzfräser. In: Seminar Aktuelle Entwicklungen
beim Vorverzahnen, WZL, RWTH Aachen, 2012
[11] Kleinjans, M.: Einfluss der Randzoneneigenschaften auf den Verschleiß
von beschichteten Hartmetalwälzfräsern. Diss. RWTH Aachen, 2003
[12] Kleppmann, W.: Taschenbuch Veruschplanung – Versuche und Prozesse
optimieren. Carl Hanser Verlag, 5. Auflage, 2008
[13] Knöppel, D.: Trockenbearbeitung beim Hochgeschwindigkeits-
wälzfräsen mit beschichteten Hartmetall-Werkzeugen. Diss. RWTH
Aachen, 1996