Optimization in engineering design

PROJECT PROPOSAL
Optimization design of helical cylinder gear reducer

Nariman fouladinejad

Nima fouladinejad

April 17, 2009

1.0 INTRODUCTION
Innovation, defined in Oxford American Dictionary as ‘the act of introducing a new process or the
way of doing new things’, has always fascinated humankind. In the context of engineering design of a
system, a product or a process, researchers and engineers usually look for innovative solutions. But
there is another aspect that innovation without optimization is not as useful as we accept it to be.
Optimization techniques can permit significant improvement in the process itself and that’s the
reason, why beside every innovation and design we should have a simultaneous optimization. By
extending this criterion we can handle so many aspects like quality, cost etc of a good and useful
design.
For example, in designing the gears it is well recognized and crystals clear that we should take so
many criteria into account to have a well based and workable design. Such as:

 Bearing capability
 Volume
 Transmission precision
 Noise

Surprisingly, we can see that Ancient engineers successfully used Direct Gear Design. They knew the
desired gear ratio and centre distance, and available power source (e. g. water current, wind, horse
power). They used them to define the gear parameters: diameters, number and shape of the teeth for
each gear. Then they manufactured gears and carved their teeth using available materials (mostly
wood), technology and tools. And without computer and programming, indeed they did a great work.

Figure1 .Ancient Gear Design

Modern gear generating process was developed during the technological revolution of the 19 th
century. This process uses a gear rack profile as a cutting edge of the hob that is in mesh with the
gear blank.

After that, by the appearance of computers, programming and advancing the technology, designing
the gears and gearbox become much more cumbersome in terms of attributes that should be
mentioned in optimizing the design .For example the current emphasis on a quieter environment has
prompted many manufacturers to undertake the study of noise and implement programs for its
reduction. Although gear noise is lower in amplitude than the overall noise in most cases, its
relatively pure tone makes it one of the most annoying sounds to an equipment operator, shop worker,
or even the passer-by.

2. PROBLEM EXPLAINATION

2.1 uses of gears

Gear usage can be classified under four main heading for parallel shaft gears as use for
synchronization, speed step-up or down, variable ratios, or for power addition.

Synchronous drives, whether by gears, toothed belts, or chains are necessary when exact position
relationship must be maintained as with overhead cam drives in engines, color register in printing
presses, and needle drives in knitting machines. Powers are often low in relation to the size of the
gears and the main requirement is high accuracy, i.e., low T.E. (Transmission Error) Noise problems
with this type of drive are often associated with low torque or reversal of torque giving movement
across the backslash and then the amount of backslash must be kept low.

Variable ratios are required for machine tools and car gearboxes to give matching between a power
unit whose speed is fixed or cannot vary greatly and a load which may vary speed over a 100:1
range. In a wide ratio gearbox it is customary to give speeds on a geometric progression with a ratio
of about 1.4 so that speed doubles every other gear as this allows selection within 20% of speed over
the whole range.

Power addition in its simplest form is used when more than one engine or crankshaft drive a common
load; a typical example is the use of two turbines driving a single propeller. Two (or more) pinions
mesh with the same wheel to give addition of torque. In this type of power addition, relative speeds
remain unaffected and torques adds. If the requirement is for torques to remain constant but speeds to
add then a differential or epicyclic system must be used.

Speed reduction use parallel shaft gears are straightforward; problems arise in the economic balance
between increasing number of stages and increasing ratio per stage. In practice 5 to 1 is the usual
limit for a single stage though 10 to 1 is possible and an overall ratio of 69 to 1 would require a
detailed costing to choose whether 2 or 3 stages would be better. High ratios of the order of 60 to 1
are easily achieved using worms and wheels in one stage but this type of gearbox is not suitable for
high speeds. Our project contents are mainly about gear reducer and can be placed in this category.

2.2 gear reducer

Cylindrical gear reducer is an important mechanical part which is widely used in aerospace,
automobile, lathe and so on. How to scientifically choose the design parameters is paid general
attention by the designers since the traditional design method could not meet the demands. Lots of
works has been done on optimization design of gear reducer e.g. Thompson used negotiation to make
optimized design aimed at the minimum volume, and Gao adopted evolutionary program. These
researches have achieved some promising results. However none of them do anything about reducing
the noise in gear reducer. Hence, our main purpose in this project is to minimize the noise in gear
reducer, to manage this we should firs take a look at the causes of noise because, there is no exact
function that can show the variation of noise in a gear. So let’s have a brief review on the causes of
gear noise.
2.3 possible causes of gear noise
To generate noise from gears the primary cause must be a force variation which generates a
vibration (in the components), which is then transmitted to the surrounding structure. It is only when
the vibration excites external panels that airborne noise is produced. Inside a normal sealed gearbox
there are high noise levels but this does not usually matter since the air pressure fluctuations are not
powerful enough to excite the gear case significantly. Occasionally in equipment such as knitting
machinery there are gears which are not sealed in oil tight cases and direct generated noise can then
be a major problem.

There are slight problems in terminology because a given oscillation at, for example, 600 Hz is called
a vibration while it is still inside the steel but is called noise as soon as it reaches the air. Vibrations
can be thought of as either variation of force, movement or of angles, though, in reality, all of them
must occur together. Also, unfortunately, mechanical and electrical engineers often talk about
“noise” when they mean the background random vibrations or voltages which are not the signal of
interest. Thus we can sometimes encounter something being described as the signal-to-noise ratio of
the (audible) noise. An additional complication can arise with very large structures especially at high
frequencies because force and displacement variations no longer behave as conventional vibrations
but act more as shock or pressure waves radiating through the system but this type of problem is rare.

In general it is possible to reduce gear noise by:

(a) Reducing the contact stress difference between all step gears, usually numerical method
can used to do so.
(b) Increasing the transmission precision and reducing rotation angle error of gear axels,
like the previous procedure this one is also possible through numerical calculation.
(c) Reducing the excitation at the gear teeth. Normally for any system, less amplitude of
input gives less output (noise) though this is not necessarily true for some non-linear
systems.
(d) Reducing the dynamic transmission of vibration from the gear teeth to the sound
radiating panels and out of panels often by inserting vibration isolators in the path or by
altering the sound radiation properties of the external panels.
(e) Absorbing the noise after it has been generated or enclosing the whole system in
soundproof box.
(f) Using anti-noise to cancel the noise in a particular position or limited number of
positions, or using cancellation methods to increase the effectiveness of vibration
isolators.

Of these approaches, (e) and (f) are very expensive and tend to be temperamental and delicate or
impracticable; also for (c) and (d) we cannot find a good function to manage them by numerical
optimization, so our main intention is to use numerical optimization techniques to reduce the contact
stress difference between all step gears and also minimize rotation angle error of gear axels in a gear
reducer through an multi-objective optimization.
3. PROBLEM STATEMENT OBJECTIVES

3.1 optimization variables

In the gear reducer optimization design, we do not consider the deflection problem on gears, and
assume the four gears are standardized gears, as shown in figure2:

Figure 2.the structure sketch of a reducer

From our preliminary studies, the parameters in the table below are essential in the design:

Table 1: Design parameters of gear reduce

Step Molar dimension Helix angle Transmission ratio Tooth width Teeth
no. number
І mn1 β1 i1 b1 Z1
II mn2 β2 b2 Z3

So we have 9 design variables and we can show them like this:

X= [mn1, z1, b1, β1, i1 , mn2 , z3, b2 ,β2] = [x1, x2, x3, x4, x5, x6, x7, x8, x9]

Among these design variables, x 1, x6 are discrete variables; x 2, x7 are integer variables; x3, x4, x5, x8,
x9 are continuous variables.
3.2 Objective functions

3.2.1. Contact stress difference

In order to reducing the noise and vibration in gears and assure the reducer’s life, the minimum
contact stress difference between all step gears is an optimization object on equal strength between
stages.

where KI = KA K v K hb K Hβ, ZI= ZH ZE, T1 is the input torque, Nxmm, i is the total transmission ratio,
Ɛ α is the transverse contact ratio, K A is the usage coefficient, Kv is the dynamic load coefficient, KHα is
the load distribution factor of gear teeth, K Hβ is the longitudinal load distribution factor on gear
teeth, ZH is the area coefficient, ZE is elastic influence coefficient.

3.2.2. Rotation angle error of gear axels

For reducing noise and increasing the transmission precision between gears, the minimum rotation
angle error of output axles is the second objective function.

Where round ( ) is rounding integer operation, n1 is the rotate speed of a high-speed axle.
3.3 constraints

We put forward the following constraint conditions, considering the factors - contact fatigue, bending
fatigue, and strength condition on big and small gears, the non - interference condition between the
high-speed big gears and the low-speed axles, and the restriction conditions on the gear parameters.

(1)The contact strength condition of high-speed gears

(2) The contact strength condition of low-speed gears

(3) The bending strength condition of high-speed Gears

(4) The bending strength condition of low-speed Gears

Where K = KA Kv KF K Fα and Y= YFα YSα Yβ

(5)The non - interference condition between the high-speed big gear and the low-speed axle

(6) The upper and lower restrictions on design variables

3.4 mathematical model

The optimization mathematical model of a two-step helical cylinder gear reducer is

4.0 DELIVERABLES
The results of the project would be in the form of:
1. Computer programming for the optimization scheme in MATLAB
2. Graph for showing the effects of sensitive variables on the design

5.0 CONCLUSUION
The chosen topic includes method of optimizing the design of gear reducer in order to minimizing the
noise. Although the objective functions of this problem have 9 variables and it’s beyond our studies in
our optimization course, we hope to manage this and it will help us in our further studies, and
designs.

6.0 REFERENCES
1. Y. Gao, G. Wang and Y.Ding “Optimal Design of Helical Gear Reducers with Generic Algorithm”, Hoisting
and Conveying Machinery, 2003

2. Nicholas P.chironis “Gear design and application”,1967

3. Rui Li, Tian Chang, Jianwei Wang, Xiaopeng Wei “Multi-objective Optimization Design of Gear Reducer
Based on Adaptive Genetic Algorithm”, Advanced Design Technology Center, Dalian University, Dalian, P.R.
China, 2008

4. J. Derek smith “Gears and their vibration (A basic approach to understanding gear noise)”, university
engineering department, Cambridge university, Cambridge, England, 1983

5. J. Derek smith “Gear noise and vibration (second edition, revised and expanded)”, Cambridge university,
Cambridge, England, 2003

6. Leroy S. Pitts “Bevel and Hypoid gear noise reduction”, New York, U.S.A, 1977