Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Geometri

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Mechanics & Industry 25, 1 (2024)

© E.B. Younes et al., Published by EDP Sciences 2024 Mechanics


https://doi.org/10.1051/meca/2023042 &Industry
Available online at:
www.mechanics-industry.org

REGULAR ARTICLE

Optimization of energy efficiency and NVH behaviour of a helical


gear unit
Emna Ben Younes1,2,3, Emmanuel Rigaud1,* , Joël Perret-Liaudet1, Jérôme Bruyère2, and Christophe Changenet3
1
Univ Lyon, Ecole Centrale de Lyon, CNRS, LTDS, UMR5513, 69134 Ecully, France
2
Univ Lyon, INSA Lyon, CNRS, LaMCoS, UMR5259, 69621 Villeurbanne, France
3
Univ Lyon, ECAM LaSalle, LabECAM, 69321 Lyon, France

Received: 1 February 2023 / Accepted: 22 November 2023

Abstract. The goal of this study is to perform a multi-objective optimization of a gear unit in order to improve
its performance in terms of mechanical power losses, gear dynamics and equivalent sound power radiated by the
housing. All these key performance indicators are closely related to the gear macro- and micro-geometry
parameters. Decision variables chosen are the helix and pressure angles as macro-geometry parameters, and the
amount and dimensionless roll length of tip relief as micro-geometry parameters corresponding to gear profile
modifications. The multi-objective optimization is carried out under geometric and load capacity constraints
using the evolutionary NSGA-II algorithm. Various results, observed in the form of 3D Pareto front confirm that
improvements in energy efficiency and vibroacoustic performance are antagonistic. Nevertheless, a significant
decrease of mechanical power losses is possible without degrading the vibroacoustic performance much.
Otherwise, the correlation between the gear dynamic response and the equivalent radiated sound power (ERP)
is partial. The minimization of the equivalent sound power radiated by the housing is not equivalent to the
minimization of the gear transmission error fluctuation. These results underline the interest of modelling the
whole gear unit to optimize its efficiency and NVH behaviour.
Keywords: Multi-objective optimization / NSGA2 heuristic /
helical gear macro-geometry and micro-geometry / mechanical power losses / transmission error fluctuation /
equivalent radiated sound power (ERP)

1 Introduction ments. As an example, the following studies can be


emphasised: (i) Sanghvi et al. optimized the load carrying
Mechanical engineering applications widely use gear capacity and volume [5]; (ii) Yao optimized the bearing
devices to transmit a force and a rotational motion, capacity coefficient, spur gear efficiency and centre distance
because they respond to power requirements and precision [6]; (iii) Padmanabhan et al. minimized the centre distance
imposed to modern architectures of kinematic chains [1]. and overall mass, while improving performance in terms of
Numerous items should be considered at design stage such transmitted power and gear efficiency [7]; (iv) Li et al.
as energy consumption, pollution or NVH performance and minimized, the transmission error, contact stress and gearbox
engineers have to combine sometimes antagonistic con- volume [8]; (v) Patil et al. minimized the power losses and
straints, such as minimize on-board mass, power losses, gearbox volume [9]. In these works, the gear design macro-
dynamic mesh loads, pressures and stresses, excitation geometry parameters were chosen as decision variables.
sources, vibroacoustic behaviour, etc. In order to do that, Some other studies have considered the gear micro-
designers increasingly resort to optimization methods [2]. geometry parameters. In fact, gear tooth profile deviation is
For example, some mono-objective optimizations have been often adopted to avoid premature contact between gear
performed in order to minimize the helical gear unit mass [3] or teeth and optimize contact conditions. In these works, the
the multi-stage gear driveline volume [4]. Multi-objective amount and dimensionless roll length of tip relief
optimizations (MOO) of cylindrical gears have also been modification were chosen as decision variables in order
performed, sometimes associated with contradictory require- to optimize the gear transmission error fluctuation [10–13]
mechanical power losses and efficiency, or noise and
durability [14–16]. Most studies in the literature resort to
* e-mail: emmanuel.rigaud@ec-lyon.fr optimizing only gear pair and shafts, rather than the whole

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
2 E.B. Younes et al.: Mechanics & Industry 25, 1 (2024)

power transmission. Regarding NVH performance, the been described in detail in the companion paper [17].
objective function retained is often the minimization of Stationary operating conditions are assumed: nominal
the gear transmission error fluctuation rather than the input torque is 500 N and nominal input rotational speed is
whining noise radiated by the gearbox. Regarding 3000 rpm, corresponding to an input power equal to 25 kW.
efficiency, the power losses considered are often limited The mesh frequency is equal to fm = 1450 Hz. Gears are
to tooth friction losses and the interaction between power splash lubricated by oil. Shafts are supported by ball
losses and temperatures is to the best of our knowledge not bearings and enclosed in a (380  280  150 mm3) rectan-
considered. gular parallelepiped housing. The thickness of the steel
In a companion paper [17], a single-stage helical gear unit walls is 10 mm.
efficiency was evaluated using the thermal network method The gear design macro-geometry parameters chosen as
in order to evaluate the bulk temperature of each gear unit decision variables for the optimization are the helix angle b
component and take account of the interconnection between (b ∈ [0°–30°]) and the normal pressure angle a0 (a0 ∈ [15°–
the numerous heat sources, i.e. the tooth friction losses, 25°]. The centre distance is imposed (aw = 174 mm).
rolling element bearing losses, shaft seal losses and churning Consequently, the normal module mn depending from the
losses. MOO is carried out using the evolutionary NSGA-II helix angle b is evaluated as follows:
algorithm, in order to minimize power losses and transmis-
sion error fluctuation. The set of decision variables
2 aw cos ðbÞ
corresponded to macro- and/or micro-geometry parameters. mn ¼ : ð1Þ
The results were in the form of a Pareto front. They showed Z1 þ Z2
that the values of decision variables associated with optimal
solutions when considering the total power losses differ from Two micro-geometry parameters are chosen for each
those obtained when considering only the tooth friction gear wheel as decision variables [18]. Ei is the amount of
losses. In addition, they showed the interest of optimizing linear tip relief. G i is the dimensionless tip relief roll length
both the macro- and micro-geometry parameters to optimize (related to the path of contact length). The ranges of
performance in terms of efficiency and transmission error. variation of micro-geometry parameters are presented in
Therefore, MOO performed with macro- and micro-geome- Table 1.
try parameters considered simultaneously leads to a wider
Pareto front than successive MOO performed firstly with
macro-geometry parameters and secondly with micro- 2.2 Thermal network and calculation of power losses
geometry parameters.
Based on these conclusions, the goal of this work is to The total mechanical power losses (Qtotal) in the gear
extend the previous study which integrates a global vision unit are evaluated using the thermal network method
of power losses to a global vision of NVH behaviour, via the described in detail in the companion paper [17]. To this
analysis of the vibroacoustic response of the housing of the end, the gear unit is discretised into 14 isothermal
studied gear unit. Thus, MOO of the gear unit is carried out elements (see Fig. 2). The first one is associated with
under constraints. Both gear design macro- (helix and surrounding ambient air. The housing is decomposed
pressure angles) and micro-geometry parameters (amount into 3 parts, i.e. an upper part, a lateral part supporting
and dimensionless roll length of tip relief) are chosen as bearings and a lower part containing the oil sump
decision variables. The objective functions are not only the (modelled by a sole isothermal element). Other isother-
mechanical power losses of the gear unit and transmission mal elements are introduced to model each of the 4
error fluctuation (associated with the gear pair behaviour), bearings, 2 shafts and 2 gear wheels, as well as the
but also the equivalent radiated sound power (ERP) meshing zone where friction between gear teeth occurs.
associated with the vibroacoustic response of the housing. Elements are connected by thermal resistances depend-
The first part of the document presents the system under ing on the type of heat transfer. Heat exchange between
study, as well as the thermal network and the different the housing and surrounding ambient air are modelled by
mechanical power losses considered (tooth friction losses, convection and radiation resistances. Heat exchange
bearing losses, shaft seal losses and churning losses). Then, between housing internal walls, gear wheels and bearings
the dynamic models of the driveline and the housing are on the one hand and oil in the other hand are modelled by
described, allowing estimation of the transmission error convection resistances. Heat exchange between gear
fluctuation, the dynamic mesh force and the equivalent sound wheels, bearings and shafts are modelled by conduction
power radiated by the housing. The objective functions, thermal resistances. Analytical models used to quantify
constraints and decision variables are presented, as well as resistances are presented in detail in [19,20].
NSGA-II algorithm parameters. Finally, MOO results are The different mechanical power losses considered (tooth
presented and discussed. friction losses, bearing losses, shaft seal losses and churning
losses) are evaluated by taking account of node temperatures
to insure the thermo-mechanical coupling through oil
2 Modelling of the gear unit properties variation (see Tab. 2). The steady-state tempera-
2.1 Description of the gear unit ture distribution in the gear unit is simulated from an
iterative procedure. As a result, total mechanical power
Figure 1 displays the gear unit under study. The main losses are calculated from the sum of the above-mentioned
characteristics of the helical gear pair (Z1:Z2 = 29:80) have contributions.
E.B. Younes et al.: Mechanics & Industry 25, 1 (2024) 3

Fig. 1. Single stage gear unit.

Table 1. Gear micro-geometry characteristics.

Pinion Wheel
Helix crowning 10 mm 10 mm
Amount of tip relief E (mm) E1 ∈ [0 – 100 mm] E2 ∈ [0 – 100 mm]
Dimensionless tip relief roll length G G 1 ∈ [0 – 0.5] G 2 ∈ [0 – 0.5]

2.3 Model of the drive line and calculation of dynamic The equation of motion of the drive line is [21]:
mesh force and transmission error fluctuation

The drive line of the gear unit is modelled using the finite x þ ½Cv x_ þ ½Kðt; xÞx ¼ F0 þ F1 ðt; x; de ðM ÞÞ
½M€
 
element method. Each gear wheel is model by a rigid _ 1;2 ;
þ F2 t; V ð2Þ
cylinder with 6 degrees of freedom (dof). The shafts are
modelled by Timoshenko’s beam elements with two nodes [M] and [Cv] are the mass and damping matrices. [K(t,x)]
and six dof per node. Each bearing is modelled by 1 axial is the non-linear time-varying stiffness matrix. F0 is the
stiffness’s element and 2 radial and stiffness’s elements. external load vector. F1 (t, x, de (M)) is the time-varying
Each mechanical coupling is modelled by a torsional force vector including the contribution of tooth profiles
stiffness and connected to an inertial element modelling the deviations.
  de(M) is the relative deviation and
motor or the receiver inertia. The corresponding model of _ 1;2 takes account of inertial effects due to
F 2 t; V
the drive line includes 36 degrees of freedom. The contact unsteady rotational speed. The total dynamic mesh
lines between the pinion and the driven wheel are force and the root mean square value of the transmission
discretized into several elements associated with elemen- error fluctuation (TERMS) are evaluated from equation
tary stiffness elements taking account of gear tooth (2) and from the unloaded static transmission error
elasticity and with initial gap vectors taking account of (corresponding to very low applied torque and input
the micro-geometry tooth profile deviations. rotational speed). Equation (2) is solved using a
4 E.B. Younes et al.: Mechanics & Industry 25, 1 (2024)

Fig. 2. Thermal network for the studied single stage gear unit [17]. Numbering of isothermal elements: (1) = ambient air; (2) = upper
part of the casing; (3) = lateral part of the casing; (4) = lower part of the casing;(5) = oil sump; (6) and (7) = bearings on pinion’s shaft;
(8) and (9) = bearings on wheel’s shaft; (10) = pinion’s shaft; (11) = wheel’s shaft; (12) = pinion; (13) = wheel; (14) = meshing zone of
gear teeth.

Table 2. Lubricant properties. r0 and c0 are respectively the air density (1.2 kg m3) and
the sound velocity in air (340 ms1) at ambient tempera-
Kinematic viscosity at T = 40 °C 120 Cst ture, S is the radiating surface area of the gear unit housing,
v (M,t) is the normal velocity at a point M of the housing
Kinematic viscosity at T = 100 °C 15.9 Cst
surface and < v2 >S is the time (upper bar ) and space
Density at T = 15 °C 860 (kg/m3) averaged (brackets <>s) mean square vibrational velocity.
The evaluation of sound power from the housing surface
vibration is an approximate method because the radiation
Newmark’s implicit scheme coupled to a normal contact efficiency s rad is difficult to assess. As a first assumption,
algorithm. The procedure is described in detail in [21] the acoustic couplings between the different faces of the
and in the companion paper [17]. parallelepiped housing are assumed to be negligible.
Consequently, the acoustic power is evaluated as the
2.4 Model of the housing and calculation of the sum of the acoustic powers radiated by each radiating face
equivalent radiated power (ERP) assumed to be baffled. As a second assumption, the critical
frequency fc is introduced, which for uniform infinite flat
The vibroacoustic response of the gear unit excited by the plates is [23,24]:
dynamic mesh force is computed assuming that the 10 mm
thick- steel housing is elastic and fitted on its base. It is sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
discretized by 500 3-D structural solid elements with 8 c0 2 12rð1  n2 Þ
fc ¼ ; ð4Þ
nodes and 3 dof per node and by 6000 structural shell 2p Eh2
elements with 4 nodes and 6 dof per node (see Fig. 3).
The model of the whole gear unit, including the drive r is the material density, E and v its Young’s modulus and
line, encompasses 7400 nodes and around 40000 degrees of Poisson’s ratio and h its thickness. The radiation efficiency
freedom. From numerical modal analysis, the whole gear s rad is usually greater than 1 at frequencies around fc, close
unit shows 280 modes between 0 and 20 kHz. An equivalent to 1 at frequencies above fc and lower than 1 at frequencies
viscous damping equal to 5% is introduced for each mode. below fc. Likewise, for a baffled finite plate, it is usual to
The sound power radiated by the gear unit is estimated assume that s rad tends towards 1 beyond fc. The critical
from the mean square vibration velocity averaged over the frequency calculated for the studied steel housing is
outer surface [22]. It is defined as follows: fc = 1175 Hz and is lower than all exciting frequencies,
Z corresponding to harmonics of mesh frequency
(fm = 1450 Hz). Therefore, a value s rad ≈ 1 is assumed.
P ac ¼ r0 c0 s rad ∫ vðM; tÞ2 dS ¼ r0 c0 s rad S < v2 >S ; ð3Þ
S
The radiated sound power corresponds to the Equivalent
E.B. Younes et al.: Mechanics & Industry 25, 1 (2024) 5

where N is the total number of elementary surfaces of area


sj located at each point Mj.
11
< v2 ðvÞ>S ¼ jF ðvÞj2 J ðvÞ: ð9Þ
2S
The first five harmonics of the periodic dynamic mesh
force at frequency vm are retained, so that:

11X 5
< v2 >S ¼ jF ðkvm Þj2 J ðkvm Þ: ð10Þ
2 S k¼1

and finally:

1 X5
ERP ¼ r0 c0 jF ðkvm Þj2 J ðkvm Þ: ð11Þ
2 k¼1

The ERP level in decibels is:


 
ERP
Fig. 3. Housing meshing. Lw ¼ 10 log10 ð12Þ
P ref

with reference acoustic power IIref = 10-12 W.


Radiated Power [25], noted ERP, which is retained as the The respective weight of axial and radial components
vibroacoustic indicator of the whole gear unit. associated with the dynamic mesh force transmitted to the
housing is affected by a change in gear design macro-
P ac ≈ ERP ¼ r0 c0 S < v2 >S : ð5Þ geometry parameters a0 and b. In practice, the surface
integral J(w,a0, b) is evaluated for a discrete number of
For a harmonic excitation at frequency v, the mobility normal pressure angle and helix angle values which extend
which relates the output velocity V(v) at point M to the over the entire range of variation of these decision variables.
dynamic mesh force F(v) is described by the complex For each individual of the multi-objective optimization
frequency response function HV,F(M,v). The time and algorithm with design parameters (a0, b), a bilinear
space averaged mean square vibrational velocity is: interpolation of the surface integral J(w,a0, b) is performed
Z from these discrete values, for each of the first five harmonics
11 V ;F V ;F   of the periodic dynamic mesh force. Then, the ERP level in
< v ðvÞ>S ¼
2 ∫ H ðM; vÞH ðM; vÞF ðvÞF ðvÞdS; decibels is estimated from equations (11) and (12).
2S S
ð6Þ
3 Functions, decision variables and
Z constraints of the MOO algorithm
11  2
< v2 ðvÞ>S ¼ jF  ðvÞj2 ∫ H V ;F ðM; vÞ dS: ð7Þ
2S S The goal of this work is to perform a MOO of the gear unit
under constraints. The 3 objective functions to minimize are:

For R each 2
harmonic, the surface integral – The mechanical power losses (Qtotal) of the gear unit
J ðvÞ ¼ ∫S H V ;F ðM; vÞ dS is the transfer function evaluated from interaction between temperature and the
between the quadratic value of the dynamic mesh force different power losses considered (tooth friction losses,
and the mean square vibration velocity averaged over the bearing losses, shaft seal losses and churning losses),
housing surface. It is estimated beforehand via the finite – The transmission error fluctuation (root mean square
element analysis. For this, a unit dynamic mesh force F(v) value TERMS associated with the gear pair behaviour),
is applied on the gear unit finite element model without – The ERP level in decibels (Lw) associated with the
elastic coupling between gear wheels, and normal velocities vibroacoustic response of the housing.
are computed on each element of the housing affected by its
elementary external area (collocation technique). The The decision variables are:
estimated integral is then given by: – 2 design macro-geometry parameters corresponding to
the helix angle b and normal pressure angle a0,
X
N   2 – 4 micro-geometry parameters corresponding to the amounts
J ðvÞ ≈ sj H V ;F M j ; v  ð8Þ of tip relief for pinion and driven wheel teeth (E1,E2) and the
j¼1 dimensionless tip relief roll lengths (G 1,G 2).
6 E.B. Younes et al.: Mechanics & Industry 25, 1 (2024)

Fig. 4. Pareto front of the multi-objective optimization. (*) Reference point. P1, P2, P3, P4: specific points.

Therefore, the decision variables vector is: (method B) [27] (h2(X)  500 MPa). For each angular
position, the maximum contact pressure h3(X) is also
X ¼ fa0 ; b; E1 ; E2 ; G 1 ; G 2 g: ð13Þ checked (h3(X)  1.2 GPa). Finally, the MOO problem is
written as follows:
The boundaries of the decision variables have been 8 8
presented in x 2.1. < min Qtotal < h1 ðXÞ ≥ 1:2
The design space must be constrained in order to ensure min T ERMS subjected to h2 ðXÞ  500 MP a X ∈ S⋅
that the solutions are feasible. First, geometric constraints : :
min Lw h3 ðXÞ  1:2 GP a
are introduced in order to ensure that the gear is functional
[26]. Operating centre distance and gear design macro- ð14Þ
geometry parameters have been chosen in order to avoid
interference and too thin thicknesses at the tooth tip. The The problem is solved using the NSGA-II algorithm. In
total contact ratio h1(X) must also ensure a continuous fact, this one is usually adopted for solving MOO problems
motion transmission (h1(X) ≥ 1.2). Then, constraints due to its fast convergence, efficiency and ability to test and
associated with load capacity are introduced in order to preserve diversity (i.e. a distribution of solutions from the
ensure that the gear can withstand operational conditions. Pareto front) [28,29]. Suitable values for parameters of the
The tooth root bending stress h2(X) is evaluated for pinion NSGA-II algorithm (selection strategy, mutation type and
and driven wheel, from standard formula ISO-6336 rate, crossover type, population size, maximum number
E.B. Younes et al.: Mechanics & Industry 25, 1 (2024) 7

Fig. 5. Decision variables associated to the Pareto front. P1, P2, P3, P4: specific points.

Table 3. Features of reference point and specific points of the Pareto front.

Reference point P1 P2 P3 P4
a0 (°) 20 23.0 22.0 17.7 18.1
b (°) 20 15.3 15.0 16.5 13.8
E1 (mm) 0 80 46 17 16
G1 0 0.44 0.38 0.19 0.28
E2 (mm) 0 79 33 12 15
G2 0 0.33 0.28 0.19 0.21
Qtotal (mm) 1900 1240 1400 1780 1710
TERMS (mm) 0.35 1.80 0.63 0.23 0.34
Lw (dB) 92.0 101.3 91.3 85.3 82.1

of generations) correspond to those presented in detail in Figure 4d displays the 3D Pareto front resulting from
the companion paper [17]. MOO, that is to say the set of non-dominated solutions
corresponding to the best compromises between objectives.
Figures 4a, 4b and 4c display the projections of the Pareto
4 Results and discussion front on respectively TERMS–Qtotal plane, Lw–Qtotal plane
and TERMS–Lw plane. Four specific points are introduced:
Before the MOO is performed, objective functions have
– P1 corresponds to the minimum mechanical power losses
been evaluated for a reference point (a0 = 20° and b = 20°)
(Qtotal = 1240 W). Among all points of the Pareto front, it
without micro-geometry tooth profile modifications and
is also associated with both the maximum of transmission
which respects geometric and load capacity constraints.
error fluctuation (TERMS =1.80 mm) and ERP level
The corresponding decision variables and objective
(Lw = 101.3 dB).
functions values are listed in Table 3. The mechanical
– P2 is the selected optimal solution among all points of the
power losses (Qtotal = 1900 W) are equal to 7.5% of the
Pareto front.
input power. They are generated in order of importance by
– P3 corresponds to the minimum transmission error
tooth friction losses (60%), churning losses (22%), bearing
fluctuation (TERMS = 0.23 mm).
losses (15%) and seal losses (3%). The oil sump tempera-
– P4 corresponds to the minimum ERP level (Lw = 82.1 dB).
ture for the reference point stays below 100 °C (in this case
96.3 °C). In contrast, the pinion bulk reaches 110 °C. The For clarity, the Pareto front is broken down into 2 sets.
transmission error fluctuation is TERMS = 0.35 mm and the The first side of the Pareto front starts from P1 and ends at
ERP level is Lw = 92 dB. P2. It corresponds to the line displayed using red empty
8 E.B. Younes et al.: Mechanics & Industry 25, 1 (2024)

circles. The second side starts from P2 and ends at P3 and time, a reasonable and slight increase of Qtotal is observed
P4. It corresponds to the surface displayed using blue solid (from 1240 W at P1 to 1400 W at P2). When the Pareto
circles. Figure 5 displays the evolution of decision variables front is browsed from P3 and P4 towards P2, a steep slope of
associated to the Pareto front and Table 3 lists the the Pareto front is observed. The significant decrease of
objective functions values and decision variables for power losses can be explained by the large increase of the
specific points P1 to P4. normal pressure angle.
The Pareto front clearly shows that the improvement of Otherwise, the first side of the Pareto front (from P1 to
one performance leads to the degradation of another and P2) shows a strong correlation between the transmission
objectively highlights the antagonism between the objec- error fluctuation TERMS and the ERP level Lw while they
tives related to the improvement of energy efficiency on the become independent on the second side (from P2 to P3 and
one hand and dynamic and vibroacoustic performances on P4). Thus, P3 corresponding to the minimum transmission
the other hand. error fluctuation shows a much higher ERP level than P4
On one side of the Pareto front, the 2 points P3 corresponding to the minimum ERP level (Lw = 85.3 dB
(minimum transmission error fluctuation) and P4 (mini- instead of 82.1 dB). Furthermore, a difference of up to
mum ERP level) correspond to similar small removals of +6 dB can be observed between 2 points of the Pareto front
material from the gear tooth flanks. Normal pressure angles showing a similar transmission error fluctuation. This can
are similar (a0≈18°) but the helix angle values differ be explained by the slight decrease of helix angle b which
(b = 16.5° at P3 and b = 13.8° at P4). As expected, leads to a decrease of the surface integral function J(w, a0, b)
minimum transmission error fluctuation is obtained for an corresponding to a different balance between the axial and
overlap ratio eb ≈ 1 which minimizes the fluctuation of gear the radial forces transmitted to the bearings and the housing.
contact length. But the mechanical power losses still reach So, the minimization of the ERP level Lw, associated with the
7% of the input power which is barely better than the vibroacoustic response of the housing is not equivalent to the
reference point. minimization of the transmission error fluctuation TERMS,
On the other side of the Pareto front, point P1 associated with the gear pair behaviour.
corresponds to power losses equal to 4% of the input
power. This minimum value is associated with a high-
pressure angle (a0 = 23°) as predicted by mono-objective 5 Conclusion
optimizations. For example, Höhn et al. [30,31] showed
that tooth friction power losses decrease with high pressure A multi-objective optimization based on NSGA-II algo-
angle. They also generally increase with helix angle because rithm has been performed in order to improve perfor-
the tooth wetted surface involved in churning power losses mances of a single-stage helical gear unit, under geometric
increases, as well as the axial forces transmitted to the and load capacity constraints. Its efficiency is improved via
housing involved in bearing losses. Point P1 also corre- minimization of mechanical power losses (taking account of
sponds to large removal of material from the gear tooth tooth friction losses, bearing losses, shaft seal losses and
flanks. This significant profile modification is required in churning losses), and its dynamic and vibroacoustic
order to reduce contact forces in the area of high sliding behaviour is improved via minimization of the transmission
velocities but it leads also to a decrease of the effective error fluctuation (associated with the gear pair response)
contact length, which affects the transmission error and the ERP level (associated with the vibrating housing
fluctuation [32]. response). The set of decision variables matches macro-
So, the Pareto front allows us to conclude on the (helix and pressure angles) and micro-geometry parame-
strength of the relationship between power losses and ters (amounts and roll lengths of tip relief).
NVH behavior: A gain of 25% on power losses leads to a The 3D Pareto front resulting from the multi-objective
too large deterioration of radiated noise (+20 dB). The optimization confirms that the objectives related to the
vibroacoustic response amplitude becomes unreasonable improvement of energy efficiency and vibroacoustic
with regard to the size of the transmission and the power performance are antagonistic. Nevertheless, a significant
transmitted. improvement of the efficiency is possible without degrad-
The point P2 results from the implementation of a ing the vibroacoustic performance much. Otherwise, the
decision maker in order to select the optimal solution correlation between the gear dynamic response and the
among all points of the Pareto front (see [6]). Compared to ERP level is observed only on the side of the Pareto front
the reference point, the specific point P2 shows a large corresponding to power losses minimization. On the other
decrease of power losses (from 1900 to 1400 W) without side, the analyse of the frequency response function which
degradation of the vibroacoustic performance (Lw = 91.3 relates the dynamic mesh force and the housing vibration
dB instead of 92 dB for the reference point). Compared to shows that the optimization of ERP level is not equivalent
the point P1 the normal pressure and helix angles are to the minimization of the gear transmission error
almost constant. The significant decrease of both TERMS fluctuation. These results underline the interest of
(from 1.80 to 0.63 mm) and Lw (from 101.3 to 91.3 dB) can modelling the whole gear unit to optimize its efficiency
be explained by reduction by more than half of removed and NVH behaviour.
material: a large decrease of the amount of tip relief E and a
small decrease of the dimensionless tip relief roll length G Acknowledgments. The authors are indebted to the French
lead to variations of products (E1G 1; E2 G 2) from (35.4 mm; National Research Agency (ANR) and to the institute Carnot
26.3 mm) at P1 to (16.6 mm; 9.3 mm) at P2. At the same Ingénierie@Lyon for their support and funding.
E.B. Younes et al.: Mechanics & Industry 25, 1 (2024) 9

References [15] C.I.L. Park, Multi-objective optimization of the tooth


surface in helical gears using design of experiment and the
[1] H. Xu, A. Kahraman, N.E. Anderson, D.G. Maddock, response surface method, J. Mech. Sci. Technol. 24, 823–829
Prediction of mechanical efficiency of parallel-axis gear pairs, (2010)
ASME J. Mech. Des. 129, 58–68 (2007) [16] N. Driot, E. Rigaud, J. Sabot, J. Perret-Liaudet, Allocation
[2] D. Miler, M. Hoic, Optimisation of cylindrical gear pair  a of gear tolerances to minimize gearbox noise variability, Acta
review, Mech. Mach. Theory 156, 104156 (2021) Acust. United Acust. 87, 67–76 (2001)
[3] O. Buiga, C.O. Popa, Optimal mass design of a single-stage [17] E. Ben Younes, C. Changenet, J. Bruyère, E. Rigaud, J.
helical gear unit with genetic algorithms, Proc. Rom. Acad., Perret-Liaudet, Multi-objective optimization of gear unit
Ser. A 13, 243–250 (2012) design to improve efficiency and transmission error, Mech.
[4] T.H. Chong, I. Bae, G.J. Park, A new and generalized Mach. Theory 167, 104499 (2022)
methodology to design multi-stage gear drives by integrating [18] ISO 21771, Gears  Cylindrical Involute Gears and Gear
the dimensional and the configuration design process, Mech. Pairs  Concept and Geometry, 2007
Mach. Theory 37, 295–310 (2002) [19] J. Durand De Gevigney, C. Changenet, F. Ville, P. Velex,
[5] R.C. Sanghvi, A.S. Vashi, H. Patolia, R.G. Jivani, Multi- Thermal modelling of a back-to-back gearbox test machine:
objective optimization of two-stage helical gear train using application to the FZG test rig, Proc. Inst. Mech. Eng., Part
NSGA-II, J. Optim. 2014, 670297 (2014) J: J. Eng. Tribol. 226, 501–515 (2012)
[6] Q. Yao, Multi-objective optimization design of spur gear [20] C. Changenet, X. Oviedo-Marlot, P. Velex, Power loss
based on NSGA-II and decision making, Adv. Mech. Eng. 11, predictions in geared transmissions using thermal networks-
1–8 (2019) applications to a six-speed manual gearbox, J. Mech. Des.
[7] S. Padmanabhan, S. Ganesan, M. Chandrasekaran, V. Trans. ASME 128, 618–625 (2006)
Srinivasa Raman, Gear pair design optimization by genetic [21] P. Velex, M. Maatar, A mathematical model for analyzing
algorithm and FEA, in: Proceedings of the International the influence of shape deviations and mounting errors on gear
Conference on Frontiers in Automobile and Mechanical dynamic behaviour, J. Sound Vib. 191, 629–660 (1996)
Engineering F IEEE, 2010, pp. 396–402 [22] M. Harrison, Vehicle Refinement: Controlling Noise and
[8] R. Li, T. Chang, J. Wang, X. Wei, J. Wang, Multi-objective Vibration in Road Vehicles, Elsevier, 2004
optimization design of gear reducer based on adaptive [23] M.P. Norton, D.G. Karczub, Fundamentals of Noise and
genetic algorithms, in: International Electronic Conference Vibration Analysis for Engineers, Cambridge University Press,
on Computer Science, 2008, pp. 273–277 2003
[9] M. Patil, P. Ramkumar, K. Shankar, Multi-objective optimi- [24] C. Lesueur, Rayonnement Acoustique des Structures,
zation of the two-stage helical gear box with tribological Eyrolles, Paris, France, 1988
constraints, Mech. Mach. Theory 138, 38–57 (2019) [25] D. Fritze, S. Marburg, H.J. Hardtke, Estimation of radiated
[10] A. Carbonelli, J. Perret-Liaudet, E. Rigaud, A. Le Bot, sound power: a case study on common approximation
Particle swarm optimization as an efficient computational methods, Acta Acust. United Acust. 95, 833–842 (2009)
method in order to minimize vibrations of multimesh gears [26] S.P. Radzevich, Dudley’s Handbook of Practical Gear
transmission, Adv. Acoust. Vib. 2011, 195642 (2011) Design and Manufacture, 2012, pp. 1–878
[11] M.S. Tavakoli, D.R. Houser, Optimum profile modifications [27] ISO 6336, Calculation of Load Capacity of Spur and Helical
for the minimization of static transmission errors of spur Gears, 1996
gears, ASME J. Mech. Des. 108, 86–94 (1986) [28] Y. Yusoff, M.S. Ngadiman, A.M. Zain, Overview of NSGA-II
[12] J.A. Korta, D. Mundo, Multi-objective micro-geometry for optimizing machining process parameters, Proc. Eng. 15,
optimization of gear tooth supported by response 3978–3983 (2011)
surface methodology, Mech. Mach. Theory 109, 278–295 [29] K. Deb, T. Goel, Controlled elitist non-dominated sorting
(2017) genetic algorithms for better convergence, in: International
[13] D. Ghribi, J. Bruyère, P. Velex, M. Octrue, M. Haddar, Conference of Evolutionary Multi-Criterion Optimization,
A contribution to the design of robust profile modifications 2001, pp. 67–81
in spur and helical gears by combining analytical results [30] B.R. Höhn, K. Michaelis, A. Wimmer, Low loss gears, Am.
and numerical simulations, ASME J. Mech. Des. 134, 1–9 Gear Manuf. Assoc. Tech. Paper 05FTM 11, 11 (2005)
(2012) [31] B.R. Höhn, Improvements on noise reduction and efficiency
[14] A. Artoni, M. Gabiccini, M. Guiggiani, A. Kahraman, Multi- of gears, Meccanica 45, 425–437 (2010)
objective ease-off optimization of hypoid gears for their [32] P. Velex, F. Ville, An analytical approach to tooth friction
efficiency, noise and durability performances, ASME J. losses in spur and helical gears—influence of profile
Mech. Des.133, 1–9 (2011) modifications, J. Mech. Des. 131, 101008 (2009)

Cite this article as: E.B. Younes, E. Rigaud, J. Perret-Liaudet, J. Bruyère, C. Changenet, Optimization of energy efficiency and
NVH behaviour of a helical gear unit, Mechanics & Industry 25, 1 (2024)

You might also like