Design and Performance Analysis of Rotor Blades of

Turbo Molecular Pump under Dynamic Loading

Conditions

By
Sahab Hafeez (170501049)
Azhar Hussain (170501043)

Supervisor
Dr. Muhammad Umer

Department of Mechanical Engineering

Institute of Space Technology
2021
 ACKNOWLEDGEMENT
Firstly we are thankful to our supervisors, Dr.M. Umer (internal) and Dr. Abdul Wasay (external)
for their indispensable advice, nonstop support, and poise during our thesis study. Their colossal
knowledge and abundant experience have encouraged us during our academic research and daily
routine. We would also like to thank Dr. Talha Irfan and Dr. M.Anwar for their practical support
of our study. We would like to thank all the lab instructors in the CNC Lab. It is their kind help
and support that have made our model possible. We would also like to thank the whole mechanical
department especially Dr. Asif Israr, HOD mechanical department for his valuable suggestions,
ever encouraging and motivating guidance. Finally, we would like to express our gratitude to our
parents, colleagues, and friends. Without their tremendous understanding and encouragement in
the past few years, it would be impossible for us to complete our project.

i
 ABSTRACT
Turbo Molecular Pump (TMP) performance is firmly linked to the rotational velocity of the rotor
blades. With the increase of rpm, rotor blades deflection starts to appear. To get the steady
performance of TMP at elevated rpm, its material quality and blade design need to be optimized.
So to improve the performance and reduce the deformation, different materials are used and the
blade design has been modified. The improved design of the rotor blade is modeled in CAD
software. The structural analysis is performed in ANSYS workbench to analyze and check for
structural integrity by using different materials. It was observed that aluminum has less
deformation as compared to titanium and structural steel. High cycle fatigue of turbo molecular
pump rotor blade is a significant design problem because fatigue failures can occur as TMP rotates
and hence resulting in the formation and propagation of cracks in the high-stress region. Fatigue
failure occurs when the part is subjected to fluctuating stresses and material Imperfections. Stress-
range diagrams are studied to estimate the allowable changing stress volumes to keep away fatigue
damage. Fatigue analysis of rotor blades is performed and optimized by modifying blade design
and using different materials. Different materials like Titanium alloy, aluminum alloy, and
structural steel are used for our analysis. The analysis is performed at 30,000 rpm using ANSYS
workbench.

ii
 TABLE OF FIGURES
Figure 1: TMP cross-sectional view ............................................................................................................. 3
Figure 2: compound turbo molecular pump .................................................................................................. 4
Figure 3: the principle of operation............................................................................................................... 5
Figure 4: (a) literature collisions (b) Reality ................................................................................................. 5
Figure 5: molecules leaving the surface of blade .......................................................................................... 6
Figure 6: molecules leaving rotors/stators .................................................................................................... 6
Figure 7: open blade...................................................................................................................................... 7
Figure 8: close blade ..................................................................................................................................... 7
Figure 9: The general concept of absolute and relative velocity .................................................................. 8
Figure 10: Geometry of Blade .................................................................................................................... 10
Figure 11: TMP rotor .................................................................................................................................. 13
Figure 12: Table_1 ...................................................................................................................................... 14
Figure 13: experimental setup ..................................................................................................................... 14
Figure 14: FBD ........................................................................................................................................... 15
Figure 15: CAD modal................................................................................................................................ 17
Figure 16: total deformation of structural steel ........................................................................................... 19
Figure 17: Total deformation for aluminum ............................................................................................... 19
Figure 18: Total deformation for titanium .................................................................................................. 19
Figure 19: Von misses stress for structural steel......................................................................................... 19
Figure 20: Eq.(von misses) stress aluminum .............................................................................................. 20
Figure 21: Eq. (von misses) stress titanium ............................................................................................... 20
Figure 22: Table_2 ...................................................................................................................................... 20
Figure 23: Total deformation along blade’s edge length ............................................................................ 21
Figure 24: Eq. (Von misses) stress along blade’s edge length. ................................................................... 21
Figure 25: Deformation and stress along blade’s edge length. ................................................................... 22
Figure 26: Blade profile (a) original (b) modified ...................................................................................... 23
Figure 27: Aluminum with modified blade (a) total deformation (b) Eq. (Von misses stress)................... 23
Figure 28: Table_3 ...................................................................................................................................... 24
Figure 29: comparison between the original and modified blade ............................................................... 24
Figure 30: Graph showing Goodman and Gerber line ................................................................................ 25
Figure 31: Fatigue life of Aluminum .......................................................................................................... 26
Figure 32: Fatigue damage of Aluminum ................................................................................................... 26
Figure 33: FOS of Aluminum ..................................................................................................................... 26
Figure 34: Graph for aluminum: load vs. life ............................................................................................. 27
Figure 35: Fatigue life for aluminum (modified blade) .............................................................................. 28
Figure 36: Fatigue life for titanium ............................................................................................................. 28
Figure 37: FOS for aluminum (modified blade) ......................................................................................... 28
Figure 38: FOS Titanium ............................................................................................................................ 28
Figure 39: Load vs. life (a) aluminum (original and modified). (b) Titanium ........................................... 29
Figure 40: Experimental setup .................................................................................................................... 30
Figure 41: Graph showing the comparison between FEA and experimental results .................................. 31

iii
 LIST OF TABLES
Table 1: rotor stages with pressure ............................................................................................................. 18
Table 2: comparison of materials ................................................................................................................ 20
Table 3: deflection and stresses in the original and modified aluminum rotor ........................................... 24
Table 4: comparison of procedures with maximum deformation ............................................................... 30

iv
 TABLE OF CONTENTS

ACKNOWLEDGEMENT .................................................................................................................................... i
ABSTRACT ..................................................................................................................................................... ii
LIST OF FIGURES…………………………………………………………………………………………………………………………………….. iii
LIST OF TABLES ........................................................................................................................................... ..iv
TABLE OF CONTENTS .................................................................................................................................... v
LIST OF ACRONYMS ………………………………………………………………………………………………………………………………viii
CHAPTER 1 ………………………………………………………………………………………………………………………………………………1

INTRODUCTION ………………………………………………………………………………………………………………………………………1

1.1 .motivation ..……………………………………………………………………………………………………………………………………..1

1.2 .overview ……..…………………………………………………………………………………………………………………………………..1
1.3 .research objectives…………………………………………………………………………………………………………………………..1
1.4 .outline and methodology …………………………………………………………………………………………………………………2

CHAPTER 2 ……………………………….……………………………………………………………………………………………………………..2

LITERATURE REVIEW………………….…………………………………………………………………………………………………………….2

2.1. History of TMP ………………………………………………………………………………………………………………………………….2

2.2. Principle of TMP ……………………………………………………………………………………………………………………………….2

2.3. TMP operation …………………………..…………………………………………………………………………………………………….3

2.4. Analysis techniques for TMP structural design …..…………………………………………………………………………….7

2.4.1. Analysis technology TMP …………………………………..………………………………………………………………………….7

2.5. Vibration analysis of rotor ……………………………………..………………………….……………………………………………..7

2.5.1. Blade vibrations …………………………………………………..……………………….……………………………………………….7

2.5.2. Dynamic vibrations of rotor ………………………………….…………………….…………………………………………………8

2.5.3. Rotor vibrations of blade ……….………………………………………………….…………………………………………………..8

2.6. Inside TMP ………………………………………………………………..……………………………………………………………………..8

2.7. TMP velocity triangles ………………………………………………..…………………………………………………………………….8

2.8. Application for TMP ……………………………………………………..………………………………………………………………....9

2.9. Characteristics of TMP …………………………………………………..…………………………………………………………………9

2.9.1. Gas loads ……………………………………………………………………………………………………………………………………….9

v
2.9.2. Critical backing pressure ……………………………………………………………….……………………………………………..9

2.9.3. Compression ratio ……………………………………………………………………….……………………………………………….9

2.9.4. Pumping speed of TMP ………………………………………………………….….………………………………………………..11

2.9.5. Specific pumping speed ………………………………………………………………………….…………………………………..11

2.9.6. Emptying of the chamber in high vacuum region …………………………..…………..……………………………….11

2.9.7. Determination of suitable backing pump ..……………………………………..…………………………………………..11

2.10. Material fatigue ………………………………………………………………………………..………………………………………….12

CHAPTER 3 ……………………………………………………………………………………………………………………………………………13

METHODOLOGY ……………………………………………………………………………………………………………………………………13

3.1. Design and CAD model ………………………………………………………………………………………………………………….13

3.2. Ansys analysis ……………………………………………………………………………………….………………………………………14

3.3. Fabrication …………………………………………………………………………………………..………………..........................14

CHAPTER 4 ……………………………………………………………………………………………………………………………………………15

ANALYTICAL MODELLING ……………………………………………………………………………………………………………………..15

CHAPTER 5 ……………………………………………………………………………………………………………………………………………17

DIFFERENT TECHNIQUES ADAPTED ……………………………………………………………………………………………………….17

5.1. CAD modeling……………………………………………………………………………………………………………………………….17

5.2. ANSYS workbench analysis …………………………………………………………………………………………………………….17

5.3. Assumptions ………………………………………………………………………………………………………………………………….18

5.4. Boundary conditions ……………………………………………………………………………………………………………………..18

5.5. Structural analysis and results ……………………………………………………………………………………………………….19

5.6. Optimization by rotor’s blade modification …………………………………………………………………………………..23

5.7. Fatigue analysis ……………………………………………………………………………………………………………………………..25

CHAPTER 6 ………………………………………………………………………………………………………………………………………….. 30

EXPERIMENTATION ………………………………………………………………………………………………………………………………30

6.1. Experimental setup ………………………………………………………………………………………………………………………..30

CHAPTER 7 …………………………………………………………………………………………………………………………………………...32

RESULTS AND DISCUSSION ……………………………………………………………………………………………………………………32

CHAPTER 8 ……………………………………………………………………………………………………………………………………………33

vi
SOCIETAL IMPACT, ENVIRONMENT AND, SUSTAINABILITY …………….………………………………………………….….33

CHAPTER 9 ……………………………………………………………………………………………………………………………………………34

CONCLUSION ………………………………………………………………………………………………………………………………………..34

CHAPTER 10 ………………………………………………………………………………………………………………………………………….35

FUTURE RECOMMENDATIONS ………………………………………………………………………………………………………………35

REFERENCES ………………………………………………………………………………………………………………………………………….36

vii
 LIST OF ACRONYMS

M molar mass of the gas.

𝑉𝐵 The average tangential velocity of the blades.
𝐾𝐵 Boltzmann’s constant
𝑁𝐴 Avogadro’s number
T Operating Temperature
𝑓𝛼 Blade angle
n total number of blades
Df blade thickness factor

𝑐̅ Mean thermal speed of molecules

A A inlet area
K compression ratio
Q gas evolution

viii
 CHAPTER 1: INTRODUCTION
1.1 Motivation:
The work presented in this thesis was conducted at the institute of space technology,
Islamabad. The reason for conducting the research work on turbo molecular pump was its usage
in a broad range of high vacuum applications, research, and development organizations and in
semi-conductor industry where the TMPs have to support corrosive gases at hypercritical
conditions. TMP pumps are easy to operate and have a low maintenance cost. As compared to
many pumps they have low vibration, no hydrocarbon operation with no regeneration
phenomenon. They can provide pumping speed up to 4000 l/s in high, ultra-high, and extremely
high vacuum ranges. To provide the industry with improved TMP pumps is the main motivation
behind our research. At high rpm of TMP its blades start to deflect, so to improve the performance
and minimize the deflection at higher rpm is our first task. Secondly, Fatigue analysis of rotor
blades of TMP needs to be done under different material conditions. Different materials were
analyzed under different parameters. Our motivation behind this project is to improve the
performance of TMP at a higher speed and to avoid fatigue in pumps.

1.2 overview:
The project is related to the design and performance analysis of rotor blades of the turbo-
molecular pump under dynamic loading conditions. TMP is a pump that is used to create an ultra-
high vacuum. It is just like a turbine with rotor blades. At high RPM, its rotor blades are subjected
to deflection. To have the stable performance of TMP on high RPMs, its material quality and rotor
blade design needs to be improved. To improve the rotor blade’s design, CAD software is used to
model the rotor blade. Another software ANSYS workbench is used to done structural analysis to
verify the structural integrity by using different materials. Another design problem that can be
experienced by TMP is high cycle fatigue, as fatigue failure propagates when the rotation of rotor
blades results in cracks in the high-stress region. Different materials are analyzed for fatigue
analysis on parameters like fatigue sensitivity, the factor of safety, life and von misses stresses.
The analysis was performed on 20,000 RPM using ANSYS workbench.

1.3 Research Objectives:

➢ Design and performance analysis of rotor blades of TMP under dynamic loading
conditions.

➢ Structural analysis of the rotor of TMP.

➢ To vary the blading configuration and rotor physical parameters.

➢ Fatigue analysis of TMP

1
 1.4 Outline and methodology:
Turbopump is a type of vacuum pump which is used to create a high vacuum. TMP is just
like a turbine with blades. The gas molecules are given momentum with the help of rotating blades.
When working on high RPM its rotor blades experience deflection. The stability of TMP is affected
due to the deflection of rotors at high RPM. A detailed literature review is done to understand the
working of TMP. To have the stable performance of TMP its blade design should be improved and
its material quality should be replaced. For modeling of rotor blades, SolidWorks (CAD) software
is used. Static structural analysis is done on ANSYS workbench to prove its structural integrity.
Different materials like Titanium, aluminum, and steel are incorporated. Fatigue analysis is done
in ANSYS workbench of different materials and results are compared. Different materials are
analyzed for fatigue analysis on parameters like fatigue sensitivity, the factor of safety, life and
von misses stresses. The analysis was performed on 30,000 RPM using ANSYS workbench. The
design and analysis of TMP’s rotor blades are done while maintaining sustainability and
developments goals in mind. Good jobs, economic growth, innovation, and infrastructure are the
main SDGs of our project.

CHAPTER 2: LITERATURE REVIEW

2.1 History of TMP:

The turbo-molecular pump was invented by Becker in 1857 which was just like the turbine
in terms of construction. The older molecular pumps were developed by Wolfgang Gaede in 1913,
Fernand Holweck in 1923, and Manne Siegbahn in 1944.

2.2 principle of TMP:

A turbo-molecular pump is used to create and maintain ultra-high vacuum in different
systems. The gas molecules are given momentum with the help of rotating blades. The gas
molecules are given directional motion. The gas molecules are directed from the inlet port to the
exhaust port. At the exhaust port, a backing pump is installed which is used to clear the chamber
which is earlier in vacuum range. Below 10-3 mbar which comes under molecular flow range. In
gas molecules flow in TMP, the mean free path of gas fragments is larger than the distance between
rotor and stator blades, gas fragments collide with rotor blades and their thermal velocity is
increased due to momentum from rotor blades. When the momentum of gas molecules is increased,
molecules move from one rotor blade to another until they reached the exhaust port. As molecular
flow must be continuous in TMP, so molecular collisions should also be continuous. The mean
free path should be always larger than the rotor-stator blade volume. Molecules of TMP do not
collide with themselves but with the rotor.

2
 Figure 1: TMP cross sectional view

2.3 TMP Operation:

TMPs are used for high speed, free oil pollution, and TMPs are widely used for vacuum
systems and room cleaning, mainly in the semiconductor industry, this guarantees the ultra-clean
environment for high precision work. The cost of TMP is US$25,000. During manufacturing, if
TMP got damaged, this results in economic loss and the corresponding repair cost. TMP is usually
damaged by vibrations during high RPM of up to 30000 RPM with inevitable disturbances due to
opening and closing of chamber doors to feed/eject chamber. At pressure below 10-3 mbar in the
range of molecular flow, the extent of the mean free path is always broader than the distance
between stator-rotor assemblies. That’s why molecules of gas collide with rotor blades and not
with each other. For this reason, the pump works best in the above-mentioned range. When
pressure surpasses the above-mentioned limit i.e. in the laminar region at a pressure above 10-1
mbar intermolecular collisions will be greater. That’s why TMP cannot evacuate the system when
it is under atmospheric pressure. Another fore vacuum pump is attached with TMP which shifts
the atmospheric to cross-over pressure. On cross overpressure, we can further operate TMP on the
pump down process. Its RPM can vary from 30,000 to 90,000 RPM. 1 Torr is the cross
overpressure of TMP. The above factor is ten times greater pressure than the maximum designed

3
cross-over pressure for mainly the oil vapor diffusion pump (100 Torr). When the pressure will be
above 1 Torr, TMP blades will get slow by impact with gas molecules. Due to this motor will get
overload and the speed of TMP blades will get slow which will be ineffective for pumping gas.
Just like diffusion pumps, TMP has some moving parts that can cause vibration which badly affects
precision instruments like electron microscopes and surface science probes. A mechanical backing
pump at the exhaust cause vibrations of about 60 or 120 Hz. Due to imbalances in TMP high-
frequency vibrations occurs. For many vacuum applications, this minute amount of vibrations is
ignorable. When vibrations remain minimum and pumping characteristics are required, a
magnetically rise rotor design can give the solution. Such types of TMPs can use oil-free bearings
on starting and exiting of turbo. On normal routine, the rotor is usually on top of bearings with the
help of strong magnets, this will eliminate mechanical vibrations. Magnetically rise TMPs are used
to operate for long time durations with few problems. The oil-free mechanical bearings need to be
replaced due to wear and tear when magnetically rise TMP is started or stopped. After minimizing
vibration, the magnetically rise rotor design gives the option of organizing in any direction, as no
oil sumps are found in most standard turbopumps.

Figure 2: compound turbo molecular pump

4
The internal view of the compound turbomolecular pump is shown in figure 2. They could reach
UHV and exhaust about 0.1mbar. The exhaust pressures and throughputs can be increased by
(>20mbar).

Figure 3: principle of operation

The molecules travel in the direction of the arrow. Each length of the arrow shows the chances
that molecules will travel in that direction.

Figure 4: (a) literature collisions (b) Actuality

When elevated rotational velocity (>1000 Hz) tip speed = molecular thermal speed.

5
 Figure 5: molecules leaving surface of blade

For proper motion, blade speed and molecular velocity need to be the same. In figure 5, it shows
there is a chance that molecules will leave the surface of the blade in a falling direction.

Figure 6: molecules leaving rotors/stators

6
 Figure 7: open blade
• When blades attached to the upper end of the pump are called open blades.
• The open blade structure will give a high pumping speed and low compression ratio.

Figure 8: close blade

• Those blades which are below the pump contain a closed blade structure.
• The closed blade structure will give low pumping speed and a high compression ratio.

2.4 Analysis techniques for TMP structural design:

2.4.1. Analysis technology TMP:
Due to the advanced capabilities of computer systems and the processing capability of mechanical
analysis tools, designing complicated systems like TMP using FEA techniques is widespread.
These techniques are used to analyze the TMP 100,000 up to 250,000 rpm. The following areas
are:
• Magnetic bearing’s static analysis
• Rotor blade’s modal analysis
• Full rotational system’s modal and harmonic analysis
• Blade rotor’s full modal analysis
• Blade rotor’s stress analysis
• TMP’s thermal analysis

2.5 Vibration analysis of rotor:

The life of TMP and its system performance largely depends upon the dynamic stability of the
rotor at high rotational speeds. Dynamic stability can be analyzed in three ways:

2.5.1. Blade vibrations:

The vibration issue is a common problem in turbo-machinery. Here single blade protruded
in its natural frequency. It causes rubbing of the tip and shows high cycle fatigue.

7
2.5.2. Dynamic vibrations of rotor:
Turbo-machinery rotating systems, including bearings and rotors, can produce vibrations
of different modes. In such dynamic behavior, stiffness of bearings and hosing becomes critical.

2.5.3. Rotor vibrations of blade:

In some circumstances, a combination of rotor and blades structure develops higher modes
in the rotor. It can produce high cycle fatigue.

2.6 Inside TMP:

TMPs can be used in a wide variety of high vacuum technologies between 10-4 to 10-10
Torr. TMP has rotor blades ranging from 25 to 60 blades, piled with alteration with rows of stator
blades, and gas molecules would be pumped between pathways. The rotor is attached to the shaft,
with a dc brushless motor and the motor would be driven at blade-tip speed as near to the thermal
speeds of the gas molecules to be pumped. TMP experiences problems related to rotor dynamics.
For achieving a 15%-20% critical speed margin, rotor dynamic design is of great concern. To
evaluate the dynamic characteristics of TMP, rotor bearing analysis would be investigated.

2.7 TMP velocity triangles:

Velocity triangles function is to relate the dynamics of the flow in a turbo-machinery. The
variation in the head is because of variation in the tangential velocity section times the tangential
speed at the specified area. For this motive, it is important to understand the speeds in the absolute
and relative frames of reference.
➢ An absolute frame of reference: zero rotation, secure concerning surface.
➢ The relative frame of reference: rotation exists, i.e. the frame of reference if observing from
the rotor.
In a turbo-machinery, it is the circumferential velocity section that associates absolute and relative
speeds. Absolute velocities are represented by “c” whereas relative velocities are represented by
“w”. The general concept is given in above figure 9.

Figure 9: general concept of absolute and relative velocity

8
2.8 Applications for TMP:
TMP has vast applications in manufacturing like semiconductor tools manufacturing, deposition
of thin-film, and noticing of bleed in fabrication. With the aid of throttled turbo-molecular pumps,
as it has features of high pumping speed, sputter deposition processes are done in industries of
coating based on the flow of process gas, usually at pressures of 3 to 50 millitorr are normally
done. Turbo pumps are efficient in pumping argon gas. Turbo pumps are also efficient in detecting
leaks. In the vacuum field, turbopumps are preferred for detecting leaks due to high vacuum
pumps. They have another application of detecting leaks in gases (oxygen, nitrogen, and carbon
dioxide) to compare it with helium as TMP has a high pumping speed for atmospheric gases.
Helium is used in some cases for leak detection as it has a small molecular size, is rare in the
atmosphere, and has low toxicity. Almost 90 percent of applications related to high vacuum use
turbo-molecular pumps which usually replace oil-based diffusion pumps.

2.9 characteristics of TMP:

2.9.1. Gas Loads:
The pump transports the gas flow which is given by:
𝑑𝑉
𝑞𝑝𝑣 = 𝑠 ∗ 𝑝 = ∗𝑝
𝑑𝑡
The equation of the gas load of the pump is shown. As gas load relates directly with pressure.
When the gas pressure is in a critical position, the turbo-molecular pumping velocity decreases
and finally reaches zero. TMP’s gas load depends upon the nature of gas and the type of cooling.
Noble gases which are heavy and have low specific heat, only a few heats is released in the
surrounding. Due to this phenomenon stator and rotor blades gets heat-up and unless effective
cooling is done. To avoid this, one should regularly check the temperature of the rotor and RPM
should be controlled to make sure the safety of the pump.

2.9.2: critical backing pressure:

It is certainly the elevated backing pressure given by the backing pump at the exhaust of TMP, and
the compression of the pump reduces. Nitrogen’s elevated critical backing pressure is always fixed.

2.9.3: compression ratio:

𝑝𝑜𝑢𝑡𝑙𝑒𝑡
The compression ratios for TMPs is noted as 𝑘 =
𝑃𝑖𝑛𝑡𝑎𝑘𝑒
𝑛
𝑉𝐵√𝑀
𝑘𝑚𝑎𝑥 = {𝑒𝑥𝑝 [ 2
] 𝑓𝛼 }
( √2𝐾𝐵 𝑁𝐴 𝑇)

9
Where,
M = molar mass of the gas.
𝑉𝐵 = is the standard tangential speed of the blades.
𝐾𝐵 = Boltzmann’s constant
𝑁𝐴 = Avogadro’s number
T= Controlling Temperature
𝑓𝛼 = blade angle
n = total number of blades
Maximum compression has exponential relation with a core of molar mass (M) and blade velocity
VB.

Figure 10: Geometry of Blade

𝑣
𝐾𝑚𝑎𝑥 = 𝑒𝑥𝑝 [ ]
𝑐̅ ∗ 𝑔 ∗ 𝑡 ∗ 𝑠𝑖𝑛𝛼
The above relation shows that maximum compression has an exponential relation with a core of
molar mass (M) and blade velocity VB.

8𝑅𝑇
Mean thermal speed of molecules, 𝑐̅ = √
𝜋𝑀

It is known that the molecular mass of nitrogen is elevated than oxygen, similarly, the compression
ratio for nitrogen is elevated than oxygen.

10
2.9.4. TMP’s pumping speed:
TMP’s pumping speed S0 contains direct relation to inlet area (A) and the blade velocity v.
S0 = A*v*sin𝛼*cos𝛼, as 𝛼 is blade angle. The above pumping speed is less than above due to entry
𝑐̅
conductivity. The entry conductivity is represented by 𝐿𝐵𝑚 = ∗ 𝐴 and the blade angle is 450,
4
then the TMP’s whose molecular weight>20 has effective pumping speed Seff calculated as:
𝑆0 +𝐿𝐵𝑚 𝐴.𝑣
Seff = = 𝑣
𝑆0 ∗𝐿𝐵𝑚 4∗[𝑐̅+1]

2.9.5. Fixed pumping speed:

Fixed pumping speed shows the ratio between pumping speed and area of the topmost ring is called
fixed pumping speed. The thickness of the blade causes the actual area to be reduced. Relation
becomes:
𝑆𝑒𝑓𝑓 𝑑𝑓 .𝑣𝜋
SA = = 𝑣
𝐴 4∗[𝑐̅+1]

Df = blade thickness factor≈ 0.9. Its unit is 1/S.cm2

2.9.6. Emptying of the chamber in high vacuum region:

To achieve the wanted high vacuum pressure, deaerating from inside surface occurs and also
preheating as it has much importance in high and ultra-high vacuum pumps. A large amount of
outgassing occurs from regions having holes. That’s why the surfaces need to be cleaned and
smooth. The choice of material and surface conditions depends on the gas evolution rate. With the
help of the pressure elevation phenomenon: first, the set-up is emptied, then valves are
unconnected from the chamber. Time is noted down as the pressure rise with a certain amount in
the room of volume V. The gas amount that got upper unit instant calculated from:
∆𝑝 .𝑣
Q= , ∆𝑝 is calculated pressure rise. When gas expansion Q and wanted pressure Pend are
𝑡
familiar, the pumping speed can be found by the above relation:
𝑄
Seff =
𝑃𝑒𝑛𝑑

2.9.7. Determination of suitable backing pump:

The backing pump function is to handle the gas or vapor quantity. During high vacuum pump
operation, the maximum permissible backing pressure must never exceed. The high vacuum pump,
pumps the effective quantity of vapor Q, with a good pumping speed Seff at the vent pressure Ph,

11
the backing pump transport the gas extent with pressure PV and pumping speed SV. For the
effective quantity of Q, according to the continuity equation:
Q = ph*Seff = Pb*Sb
So, the pumping speed of the backing pump can be calculated from
𝑃
Sb = 𝑃ℎ ∗ 𝑆𝑒𝑓𝑓
𝑏

2.10. Material fatigue:

Due to cyclic load structure becomes fail and material fatigue phenomena occur. This form of
structural deformity happens when given a stress line that is much lower than static material
strength. The mechanical structures fail due to fatigue. The failure of the component by repeating
load consists of three types:
➢ The maximum amount of cycle produces damage on the micro amount and after that, it
becomes a macroscopic crack.
➢ The macroscopic cycle becomes large for every turn and it reaches its critical amount.
➢ The maximum load can never be beard, so the deformed component breaks.
In some cases, 2nd case cannot be experienced. Due to microscopic crack, sudden material failure
occurs. Fatigue is studied in the first stage. When there is no constant external load, the condition
in the material also changes with time. The material contains different parameters such as stress
and strain. A load cycle is a time from one peak in the observed variable to the next peak. The
stress fluctuates under a maximum and minimum stress while load cycle. In the phenomenon of
fatigue, stress change can be express by amplitude and mean stress. The stress range and the R-
value are mostly used to define a stress cycle. The main parameter for the fatigue damage is the
stress amplitude and mean stress. The tensile mean stress is responsible for sensitivity to fatigue
and compressive mean stress is used for higher stress amplitudes. The material behavior to
continuous load cycles is dependent on the external load that could be periodic, uneven, and
contains repeated blocks. To develop a crack, it is usual to know the difference between low and
high cycle fatigue. In HCF, the stresses are much low so that stress-strain becomes elastic and it
tells the local state. In the case of LCF strain range is used. Work on fatigue was started in the 19th
century and many methods are invented for fatigue prediction. One of the famous methods is of
S-N curve. It is the relation between stress amplitude and lifetime. For long lifetime is obtained
stress amplitude should be below. According to this relation, a reduction of the stress amplitude
by 10% will increase the lifetime up to 50%. In a few materials exhibit, there exists a threshold in
fatigue testing. When stresses will be less than the above threshold, regarded as the endurance
limit, no fatigue damage is experienced and components can run for an infinite lifetime. Few
materials have an endurance limit. They usually fail when fatigue on low levels of stress. In our
analysis, we will perform high cycle fatigue analysis and pump rotor blades and then modified
their shape to optimize the results. First, we will see for aluminum but we will also use titanium
as it is a stiffer material and is also one of the few trials having an endurance limit while aluminum
doesn’t have this feature.

12
 CHAPTER 3: METHODOLOGY
3.1. Design (CAD modeling):
TMP is modeled in software named Solid works, which is an important and inbuilt
3D CAD software to deal with the difficulties organizations face as they analyze, design, and share
knowledge with respective partners. Its function is to deal with 3D mechanical systems for
modeling. This software has many applications in the mechanical, aerospace, and defense industry.
Design and modeling of the TMP rotor are done under the following suggestions derived from the
experimental results. As the length of the blade, angle of the blade, and rotational speed are directly
related to the pumping speed. Similarly, the length of the blade and angle of the blade is inversely
related to the compression ratio. The angle of the blade is remained under 200 to 450 [1]. Frictional
effects are experienced that may burden the motor beyond this limit that can cause failure.

Figure 11: TMP rotor

3.2. Ansys Analysis:
Ansys structural analysis software is used to solve intricate structural engineering problems and
make better design decisions. It is used in different industries to aid engineers to optimize their
product development and cut the expenses of experiments. Structural static analysis of pump rotor
has been analyzed in (ANSYS) workbench for both materials namely Aluminum and Titanium
blades.
The pressure varies from 1×10-12 𝑀𝑃𝑎 𝑡𝑜 1×10-2 𝑀𝑃𝑎
in seven stages as shown in Table.1 [1]

Figure 12: Table_1 [1]

13
3.3. Fabrication (model):
In the fabrication, rotor blades were designed in a wire cut machine. Original and modified blades
were designed to check deflection. The modified blade is designed by chamfering its one end while
the original is kept constant in length. Both blades were compared on the experimental setup and
the modified blade performed better in terms of deflection due to applying different weights. In
the case of the modified blade, deflection in blades reduces as compared to the original shape.
Several readings were taken and after that average value is taken out.

Figure 13: experimental setup

CHAPTER 4: ANALYTICAL MODELLING
In general, for TMP, the blade angle of the initial step should have a high pumping speed. For the
last step, the blade designs should achieve a high compression ratio with a small angle. To make
sure the pump performance and its manufacturing, values of blade angle from the initial step to
the end step are gradually taken as 45◦ to 20◦. In ANSYS Workbench, analysis is first done for
TMP rotor and then also for an individual blade for both cases (original and modified) is analyzed
separately which acts just like a beam with one end fixed and the FEA results are compared with
experimental results. Analytically also the maximum deflection is calculated for the blade by the
following method:

d12 y1
EI ( dx1 2 ) = M1(x)

d12 y1
M1(x) = 0; or EI ( dx1 2 ) = 0
Figure 14: FBD

14
 dy1
EI = C1 … (a)
dx1

EI y1 = C1x1 + C2 … (b)
M2(x) = Px2 - P (L - a)

d22 y2
EI ( dx2 2 ) = P (L - a) x2 + C3

dy2 𝑃
EI = 2 𝑥22 - P (L - a) x2 + C3 … (c)
dx2

𝑃 𝑃(𝐿−𝑎)
EI y2 = 6 𝑥23 - 𝑥22 + C3x2 + C4 … (e)
2

Boundary Conditions:
dy1
At x2 = 0, =0
dx1

From Eq. (c) C3 = 0

At x2 = 0, y2 = 0
C4 = 0

Continuity Condition:
dy1 dy2
At x1 = a, x2 = L - a; =
dx1 dx2

From Eqs. (a) and (c),

𝑃(L − a)2
C1 = - [ 𝑥22 - P (L - a) 2]
2

𝑃(L − a)2
C1 = 2

At x1 = a and x2 = L - a, we have y1 = y2
From Eqs. (b) and (d),
𝑃(L − a)2 𝑃(L − a)3 𝑃(L − a)3
( ) a + C2 = -
2 6 2

𝑃𝑎(L − a)2 𝑃(L − a)3

C2 = -
2 3

From Eq. (b),

P
y1 = [3(L - a) 2 x1 - 3a (L - a) 2 - 2(L - a) 3] … (1)
6EI

15
From Eq. (4),
P
y2 = [𝑥23 - 3(L - a) 𝑥22 ] … (2)
6EI

E = 90 GPa for aluminum

1
I= bh3 = 2.8125 E-12 as b = 10mm, h = 1.5 mm
12

From Eq. (2),

y2 = ymax when x2 = L and L = 30 mm and a = 10 mm
ymax = 0.228 mm
These results obtained are comparable with FEA and experimental results with FEA and
experimental results being equal to 0.225 mm and 0.21 mm respectively.

CHAPTER 5: DIFFERENT TECHNIQUES ADAPTED

5.1. CAD modeling
CAD modeling is done for optimized dimensions and the values of blade angle from the first stage
to the last are gradually taken between 45◦ and 20◦. The CAD modeling is performed by using
different commands such as sketching, extrude, pattern, etc. The CAD is done using solid works
which are used to model 3D mechanical systems for sketching and assembling the systems. This
software has much application in the aerospace and defense industry. Design and modeling of
TMP rotor are made under the following observations:
• The length of the blade, angle of the blade, and rotational speed are related directly to the
pumping speed.
• The angle of the blade remains in the design limit from 200 to 450 . Frictional effects occur
that can overload the motor if reaching above design limit, due to this failure may occur.
[2]

Figure 15: CAD modal

16
5.2. ANSYS Workbench analysis:
ANSYS software with the help of structural analysis aids us to solve complex engineering
problems and give better decisions. Ansys Workbench software has applications in industries to
aid engineers to get better design results and reduce the costs of experiments. After designing out
TMP modal in Solid works, it is imported in ANSYS workbench software to perform structural
analysis of pump rotor blades workbench in the following way.

5.3. Assumptions:
• Linear analysis is performed.
• Structure analysis is performed.
• For fatigue analysis, Gerber theory is used.

5.4. Boundary conditions:

• The maximum RPM applied is 30,000 rpm.

• Cylindrical supports are applied.
• The pressure range from 1×10-12 𝑀𝑃𝑎 𝑡𝑜 1×10-2 𝑀𝑃𝑎
in seven stages as given in Table.1 [2]

Table 1: Applied pressure at different rotor stages [2].

Rotor stages Pressure (MPa)

Stage 1 1.0E-12
Stage 2 1.0E-11
Stage 3 1.0E-10
Stage 4 1.0E-8
Stage 5 1.0E-6
Stage 6 1.0E-4
Stage 7 1.0E-2

17
 5.5. Structural analysis and results:
Structural analysis of turbo molecular pump rotor is performed using ANSYS workbench and the
following results are obtained at a rotational velocity of 30,000 rpm.

Figure 16: Total deformation for structural steel Figure 17: total deformation of aluminum

Figure 18: Total deformation for titanium Figure 19: Eq.von misses stress for structural steel

18
Figure 20: Eq. (von misses) stress aluminum Figure 21: Eq. (von misses) stress titanium

Fig:(16,17 and 18 ) shows total deformations for structural steel, aluminum, and titanium
respectively with maximum values of deformation are 0.62126 mm, 00.61152 mm, and 0.7462
mm for structural steel, aluminum, and titanium respectively which indicates that aluminum has
less deformation comparatively. Fig:(19, 20, and 21) shows Equivalent (Von misses) stress for
structural steel, aluminum, and titanium respectively with maximum values are 519.98 MPa,
182.56 MPa, and 302.71 MPa for structural steel, aluminum, and titanium respectively which
indicates that aluminum has less deformation comparatively so it means that aluminum material
should be preferred as it is stiffer and hence has less deformation and stresses comparatively.
Table: 1 shows the comparison for these materials in terms of total deformation, directional
deformation, Eq. (von misses) stresses, normal stresses, shear stresses. The analysis is performed
at 30,000 rpm.

Table 2: comparison of materials

Material Total Directional Eq. (von Normal shear

deformation deformation misses) stress stress stresses
(mm) (mm) (MPa) (MPa) (MPa)

structural steel 0.62126 0.034799 519.98 531.39 257.51

Aluminum alloy 0.61152 0.032134 182.56 187.72 90.565
Titanium alloy 0.7462 0.046562 302.71 313.27 150.48
FIGURE 22: Table 2

19
 Figure 23: Total deformation along blade’s edge length

Figure 24: Eq. (Von misses) stress along blade’s edge length.

20
 Figure 25: Deformation and stress along blade’s edge length.

It shows the behavior of stress and deformation along the length of the blade and it can be observed
from the figure (25) that deformation is minimum near the fixed end of the blade i.e. where the
blade is attached to the shaft and increases towards the free end of the blade. Figure (24) shows
the stress variation and it is clear that stress is maximum near the fixed end of the blade as here
more resistance is experienced and hence more stresses are generated and the values of stresses
decreases towards the free end of the blade opposite to deformation and this is shown in figure
(23). Here deformations and stresses are shown in the same figure along the length of the blade
and their behavior is almost opposite to each other as was expected.

21
5.6. Optimization by rotor’s blade modification:

(a) (b)

Figure 26: blade profile (a) original (b) modified

The blade profile is modified based on the stress concentrated region identified from structural
analysis. Stresses are more near the fixed ends of blades and decrease as we move towards the
free ends of blades so based on these observations, blade thickness is increased near the stress
concentrated regions for stresses to distribute over a wide range of areas.

(a) (b)
Figure 27: Aluminum with modified blade (a) Total deformation (b). Eq. (Von misses) stress

22
Table: 3
Deflection and stresses in the original and modified aluminum rotor
parameter Original blade Modified blade

Minimum Maximum Minimum Maximum

Total deformation 0.00 mm 0.62126 mm 0.00 mm 0.35764 mm

Eq. (Von misses) stress 0.02286 MPa 182.56 MPa 0.02675 MPa 121.7 MPa
Normal stress -95.475 MPa 187.72 MPa -53.938 MPa 116.17 MPa
shear stresses -90.684 90.565 MPa -59.609 MPa 59.102 MPa
Max Principal Stress -10.787 MPa 190.84 MPa -6.3133 MPa 136.07 MPa

Figure 28: table_3

Figure 29: comparison between original and modified blade

Rotor’s blade is modified in such a way as to decrease the different effects such as total
deformation, stresses, etc. and for this first stress concentrated region is identified and in our case,
it is the region near to the fixed end of the blade and so the blade is modified and hence is optimized
and results are shown in figure (26, 27 and 29) and Table 2. Table 3 shows the comparison for the
original as well as a modified blade in terms of different parameters and it shows that total
deformation is reduced to 58 percent of the original blade. Equivalent stresses to 66 percent of the
original blade, Normal stresses, and shear stresses to 61 and 66 percent respectively of original
blade and hence it is clear that results are improved and design is optimized with the modified
blade.

23
5.7. Fatigue analysis:
Now fatigue analysis is performed for our design part i.e. TMP and since aluminum material is
found to be a better choice from our structural analysis so fatigue analysis is performed and
different criteria can be used for fatigue analysis and they include Goodman criteria and Gerber
criteria. Goodman criterion is a straight line drawn from endurance limit value down to the ultimate
strength value and here most of the data lie above this line so, in other words, the Goodman line
does not do the best job of describing the mean of representing the experimental data although
simple as it is linear. Gerber criterion is more like parabolic in shape and it does a better job of
describing the mean of the phenomenon and so this criterion is used for our analysis with a fully
reversed loading type.

Figure 30: graph showing Goodman and Gerber line

24
Figure 31: Fatigue life of Aluminum Figure 32: Fatigue damage of Aluminum

Figure 33: FOS of Aluminum

25
According to Fatigue Life analysis when the loading remains constant, then the available life
shows the number of cycles that the structure can bear without failure and the structure can fail if
the required limit is exceeded. In figure 32, the results got for TMP structure comprised of
aluminum and it depicts that it is safe for a life of 88000 cycles only much lower than our designed
life of 1 million cycles. Fatigue Damage in the form of a contour plot is shown in the given design
life. It is the ratio of design life and available life. In the case of Fatigue Damage, when values are
greater than 1 that means the system will get fail until the design life is reached. In figure 31, it is
observed that the damage values are more than 1 which shows that system is not in safe criteria
for the designed life and will lead to fatigue damage. In the case of the Fatigue, Safety Factor is a
contour plot of the factor of safety concerning a fatigue failure at an observed design life. In the
case of the Fatigue Safety Factor, values less than one shows failure before the design life is
observed. For aluminum as is shown in figure 33, the region below 1 exists that is 0.7 safety factor
which indicates it will fail before our designed life as a factor of safety must be greater than 1 to
avoid fatigue failure and hence safe design.

Figure 34: graph for aluminum: load vs. life

It is a graph showing the number of cycles a material can withstand against different loading
conditions and as can be seen at 1 i.e. under given loading conditions of 30,000 rpm, life is about
88000 cycles and it decreases as loading conditions are increased and vice versa. For example, as
shown if loading conditions are decreased to 0.7 or 70 percent of given loading conditions then it
will increase our part life from 88000 cycles to about 1 million cycles. So to obtain the desired
results of million cycles for our loading conditions of 30,000 rpm, fatigue analysis for the
aluminum case with the modified blade is performed. So with this, the part will last for million
cycles and the analysis is also performed for titanium because of the interesting feature that it
exhibits endurance limit while aluminum doesn’t have this luxury. Titanium and certain other
materials have this characteristic that if the stress value is low enough then there will be a point at
which it will essentially run the number of cycles as has desired and it will not have any additional
effect on when the material is going to fail. As in mechanical design, the normal way is to design
for infinite life, and one million cycles are mostly regarded as the “infinite threshold”.

26
Figure 35: fatigue life for Aluminum (modified blade) Figure 36: fatigue life for Titanium

Figure 37: FOS for aluminum (modified blades) Figure 38: FOS for Titanium

27
 Figures 35 and 36 show the fatigue life for aluminum with modified blade configuration
and titanium and both show that it will sustain for one million cycles but since titanium is
among those which exhibit endurance limit and hence is safer and is clear from the figure.
For aluminum also the lower stress value is achieved to get one million cycles but still
there are more chances that it will fail as having no endurance limit. Figures 37 and 38
show the factor of safety and it is clear that the region below 1 does not exist and the design
is safe.

(a) (b)
Figure 39: load vs. life (a) aluminum (original and modified). (b) Titanium
Figure 39(a) shows that the design life for the part has improved from 88000 cycles to one 1 million
cycles when a modified blade is used. Loading condition of 1 indicates the life against the applied
boundary conditions i.e. angular velocity of 30,000 rpm indicated by dots. Figure 39b shows that
for our boundary conditions, titanium has achieved its endurance limit and is considered to have
infinite life.

28
 CHAPTER 6: EXPERIMENTATION

6.1. Experimental setup

Blade with original as well as the modified profile is manufactured and tested for deformation
through an experimental setup shown in figure and results are shown in the table.

Figure 40: experimental setup

Table 1: comparison of procedures with maximum deformation

Procedure Maximum deformation

Figure 42:
Analytical 0.22 mm
FEA 0.22 mm
experimental 0.21 mm

For modified blade profile, max deflection = 0.14 mm

29
 Figure 41: Graph showing comparison b/w FEA and experimental results

From the experimental setup, it is found that deflection is maximum at the free ends of the blade
and minimum at the fixed end as observed through FEA analysis using ANSYS and it also shows
that the modified blade has less deformation compared to the original blade and hence has better
performance thereby verifying our FEA results. Both analytical and FEA results are the same but
when compared with experimental results for different loading conditions then some error is
observed which is found to be approx. 8% as can also be observed from the graph.

30
 CHAPTER 7: RESULTS AND DISCUSSION

Aluminum with a modified blade configuration is preferred as it has less deformation and stresses
comparatively. The total deformation and Von-Mises Stresses are reduced to 58% and 66%
respectively of the original blade as is cleared from figure 20. Figure 39a is the graph showing that
the design life for the modal has improved from 88000 cycles to one 1 million cycles when a
modified blade is used. Loading condition of 1 indicates the life for part against the applied
boundary conditions i.e. angular velocity of 30,000 rpm. When 21000 rpm is applied then TMP
design with an original blade is having a life of 1 million cycles but with modified blade design,
an addition 9000 rpm can be applied for the same boundary conditions and is now good for 30,000
rpm, so performance is optimized by small modification in blade’s shape. When titanium is used,
then the same 30,000 rpm can be applied without modifying the blade design and the reason for
using the titanium is because it makes the design safer comparatively as it is one of those few
materials that exhibit endurance limit while aluminum doesn’t. Figure 36 shows that stress value
for endurance limit is achieved and the design has a life of 1 million cycles which is considered as
“infinite threshold” and designing is generally done for this endurance limit. This is also clear from
the graph shown in figure 39b. Although titanium is expensive comparatively but also durable and
is safer so from an economic point of view aluminum is a better choice and from a durability and
strength perspective, titanium is preferred but some compromised is to be done in both cases as
per requirement.

31
 CHAPTER 8: SOCIETAL IMPACT, ENVIRONMENT, AND
SUSTAINABILITY

In this project, SGDs goal numbers 8 and 9 are followed. According to goal 8, it targets good jobs
and economic growth. Goal 9, targets innovation and infrastructure. TMPs are used in research
and development areas. They have very harsh applications in the semiconductor industry where
they handle corrosive gases and go through with critical processes. So basically TMPs are the
backbone of industries and playing their part in economic growth. Moreover, the more industries
will play part in economic growth, the more opportunities for jobs will open. The turbo-molecular
pump’s vacuum technology is one such innovation that has originated as a result of its use in
manufacturing activities across the electrical and electronics, power, consumer durables, and
scientific instrumentation industries and has a strong link with the development of
infrastructure.

32
 CHAPTER 9: CONCLUSION

In this project, structural analysis of TMP is performed at 30,000 rpm using ANSYS workbench
and it is concluded from our analysis that aluminum performs better as it has less deformation and
stresses compared to titanium and structural steel and these results are shown in Table 2.
Aluminum is also economical and light in weight. Further fatigue analysis is performed and results
show that TMP design made up of aluminum fails after 88000 cycles much less than our design
life of million cycles and factor of safety is also less than 1 indicating that design is not safe for
the given boundary conditions. However, it runs for 1 million cycles when a rotational velocity of
21000 rpm is applied. The results are optimized by modifying the blade’s shape and can now
sustain 1 million cycles at 30,000 rpm with modified blade configuration as is clear from the graph
shown in figure 39. Fatigue analysis is also carried out for titanium as it is safer and durable
because it is one of those few materials which exhibit endurance limits while Al does not have an
endurance limit like other ferrous metals. Thus, it is very difficult to predict its failure under fatigue
loading. It may be thousands or even millions of cycles before it fails (all depending on the loads/
cycles). For durability and a more safe design, stiffer material like titanium having high fatigue
strength can be used. However, Ti is comparatively expensive than Al. This solution is only useful
for highly precise equipment. Moreover, Ti exhibits the "endurance limit". Therefore, its fatigue
limit can be predicted numerically for a given number of cycles. It is observed that when aluminum
with a modified blade or titanium is used then an extra velocity of 9000 rpm can be observed. So
aluminum with modified blade configuration is preferred as it has less deformation and stresses
comparatively and as observed in figure 20 that total deformation and von missed stresses are
reduced to 58 percent and 66 percent respectively of the original blade. Aluminum is also cheaper
but if durability is needed then stiffer material like titanium having high fatigue strength can be
used.

33
 CHAPTER 10: FUTURE RECOMMENDATIONS

Optimization of the turbomolecular pumps blade profile is a vast and continuing research area in
which many options exist and work is required to further optimize and improve the performance
characteristics of these vacuum pumps. Furthermore, future work will focus on the improvement
of their performance concerning compression ratio, pumping speed, etc. for different gases.

34