Turbomachinery Blade Vibrations
Turbomachinery Blade Vibrations
Turbomachinery Blade Vibrations
2012 Hawaii
12 B litres
2
1 M cycles
3
A Steam Turbine
The workhorse of power generation
Designed to operate for years
7 seconds
An Aircraft Engine
The workhorse of air travel
Powerful, light, reliable
failure
(Warwick, 2008)
Lessons Learnt
harmful
short time
Let us define a
turbomachinery
turbomachinery ideal
ideal?
?
A Turbomachinery Ideal
D
Durable
bl
Available
Reliable
Fuel-flexible
Lightweight
Silent
Eco-friendly
Low-maintenance
Powerful
Efficient
Affordable
Reaching or Not-Reaching
Not Reaching the Ideal
D
Durable
bl
Available
Reliable
Turbomachinery
blade vibrations are the
Fuel-flexible
show-stopper No 1 that
prevent us
Lightweight
Silent from reaching this ideal
Eco-friendly
Efficient
and
d
Low-maintenance
expensive
it is
i
Powerful
Affordable
10
30% of
11
12
Vibrations in Turbomachines
13
Vibrations in Turbomachines
Induced by unsteady loads
Structural
Aerodynamical (fluid-structure interaction)
Damped
p
p
preferred
14
Vibrations in Turbomachines
Induced by unsteady loads
Structural
Aerodynamical (fluid-structure interaction)
Unstable self-excited
Damped
p
p
preferred
Failure due to overload
15
Vibrations in Turbomachines
Induced by unsteady loads
Structural
Aerodynamical (fluid-structure interaction)
~100Mio
cycles
Vibrations in Turbomachines
Induced by unsteady loads
Structural
Aerodynamical (fluid-structure interaction)
~100Mio
cycles
unsafe
safe
Unstable self-excited
Limit Cycle
y
Oscillations (LCO)
(
)
Damped
p
p
preferred
Failure due to overload
HCF Haigh Diagram
17
Typical Flow
Flow-Induced
Induced Vibration
A windy day
An open landscape
Structures exposed to flow
18
19
flow
St
Structure
t
Structure (elastic)
20
mx kx F (t )
Structural part
Structure
St
t
frequency
mode
Excitation here
aerodynamic forces
x: deformation coordinate
modal coordinate
: natural frequency
q
y of
structure
Aero load
ffrequency
q
y
N t
Natural
l mode
d
k
m
(
(eigenmode)
d )
21
F ps n ds
Load
Structure
St
t
frequency
mode
Aero load
ffrequency
q
y
direction
Flow
ps
If ps=ps(t) then F=
F F(t)
22
Flow-Induced
Flow
Induced Vibration
Inertial forces
Aerodynamic forces
mx kx F (t )
Elastic forces
Resonance
phenomenon
Structure
St
t
frequency
mode
Aero load
ffrequency
q
y
direction
Same frequency
q
y
Force in direction
of mode
23
Flow-Induced
Flow
Induced Vibration
Inertial forces
Aerodynamic forces
mx kx F (t )
Elastic forces
Resonance
phenomenon
Structure
St
t
frequency
mode
Aero load
ffrequency
q
y
direction
Same frequency
q
y
Collars triangle
of forces (1946)
Force in direction
of mode
24
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
Re
Inertial forces
Elastic forces
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
Re
Inertial forces
Elastic forces
Aerodynamic
damping force
Re
Inertial forces
Aerodynamic
y
stiffness force
Elastic forces
Bringing it Together
mx cx kx F (t )
Fae (t ) Fdamping (t ) Fdisturbance (t )
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
mx (c cae ) x (k k ae ) x Fdisturbance (t )
Multiple degrees of freedom: scalars vectors
M X C Cae X K K ae X Fdisturbance (t )
29
Important Aspects
The ratio of structural to aero forces matters
(Structural) damping
forces
Im
Aerodynamic force
Aerodynamic
damping force
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
Re
Inertial forces
Aerodynamic
stiffness force
Elastic forces
30
Important Parameters
Mass
M
ratio
i
Ratio between airfoil mass
and mass of surrounding
fluid
4m
0 c 2
Reduced frequency
Relation between time-offlight of fluid particle
across airfoil during one
oscillation period
t 2fc
k
T
u
k aero damping reduced ( negative)
31
Application to Turbomachines
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
32
Turbomachine Environment
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
Blade rows
Stationary
Rotating
33
Vibration of Bladed
Bladed-Disk
Disk Structures
Blades
Disk
Bladed disk
34
Vibration of Bladed
Bladed-Disk
Disk Structures
2ND
3ND
+
-
Blades
1E
1F
1T
Disk
Bladed disk
35
TWM ND 0
ND 0
=0deg
ND nodal diameter
FT forward traveling
BT backwards traveling
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
36
TWM ND 6 FT
ND 6 FT
=90deg
ND nodal diameter
FT forward traveling
BT backwards traveling
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
37
TWM ND 12 FT
ND 12 FT
=180deg
ND nodal diameter
FT forward traveling
BT backwards traveling
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
38
f [Hz]]
4.2kHz
1T
2F
3.1kHz
1F
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
1 5kHz
1.5kHz
800Hz
N/2 ND
1F +-4 ND
1F +-1 ND
Frequencies vary with engine speed 1F 0 ND
F
Frequencies
i can vary with
i h nodal
d l diameter
di
6000
[rpm]
39
f [Hz]]
4.2kHz
1T
2F
3.1kHz
1F
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
1 5kHz
1.5kHz
800Hz
N/2 ND
1F +-4 ND
1F +-1 ND
Frequencies vary with engine speed 1F 0 ND
F
Frequencies
i can vary with
i h nodal
d l diameter
di
6000
[rpm]
40
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
41
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
span
n
u
42
t=60/[rpm]/N f=1/t
Aero load
ffrequency
q
y
direction
phase
span
n
Structure
St
t
frequency
mode
damping
Fdisturbance
(t) u
43
An Excitation Diagram
f [Hz]]
EO 60
EO : Engine Order
f=[rpm]/60*60
EO 20
f=[rpm]/60*20
f2
2kHz
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
EO 11
f1
f=[rpm]/60*N1 (=11)
1 1kHz
1.1kHz
6000
[rpm]
44
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
Aerodynamic coupling
45
=180deg
ND 12 FT
ND nodal diameter
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
FT forward traveling
BT backwards traveling
=-90deg
ND 6 BT
46
Bringing it Together
EO 60
f [Hz]]
Forced response
EO 20
1T
2F
1F
EO 11
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
Fl tt
Flutter
6000
OP range
[rpm]
47
Turbomachinery Aeroelasticity
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
(Giles, 1991)
48
49
Flow instability
e.g. vortex
shedding
EO 60
Non-Synchronous Vibrations
EO 20
1T
2F
1F
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
EO 11
Fl tt
Flutter
6000 Forced response
[rpm]
OP range
50
The Reality
Experimental Campbell Diagram
Kielb et al. ASME Turbo Expo,
2003
Acceleration to 95% Speed
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
51
52
Remedy problems
Ensure that certain operating points are avoided
Ensure that the problems are made harmless
High
g p
positive damping
p g
HCF tolerant materials
53
Structure
St
t
frequency
mode
damping
Aero load
ffrequency
q
y
direction
phase
Aero load
f
frequency
direction
phase
h
Structural analyses
Static loads, mode shapes,
frequencies, damping
Mutual interaction
Aerodynamical analyses
Unsteady aerodynamic forcing
Aerodynamic damping
Mayorca, 2011
55
Mayorca, 2011
56
An
AnExample
ExampleAeromech
AeromechDesign
DesignProcess
Process
Back to Aero design
Silent
Efficiency is king
Fuel-flexible
Lightweight
Eco-friendly
Low-maintenance
safety
y
Powerful
Efficient
Affordable
58
59
An Example Aeromech
DesignStructural
Process
Computational
Dynamics (CSD)
60
An Example Aeromech
DesignFluid
Process
Computational
Structural
Dynamics (CFD)
(CSD)
Cyclic
symmetric
models
Forced
response
(0)
100k
per sector RANS
Full-size
3DDOF
time-marching
p clearance,, inter-row
Details ((tip
Model cavities)
gaps,
size not extremely
modeled (but
critical
not
for modal analysis (other than
always)
y )nodes p
stress
analysis)
(0)
(
) 100k-1M
per p
passage
g
Usually single or few passages
Updated system matrices (e.g.
stiffening
effects)
Aerodynamic
damping
3D time-marching or linearized
Modeling
of material damping,
viscous approaches
friction
damping,
Mode shapes
fromdamping
FEM (loose
coatings
coupling) or time-marching
CFD/CSD (strong coupling)
61
Example:
p
Aero Damping
p g CFD
62
63
Proficiency
Accuracy
Reliability
in use
Clarity
about objectives
64
L t us do
Let
d a test
t t
65
Task
To predict the steady aerodynamic loading
Students performed
Meshing
Simulation setup
Solving
Extraction of loading
Centralized post-processing
66
67
FUTURE
Input
Design intent: low ( negative) aero
Geometry
damping as stall is approached
Task
To predict the minimum aerodynamic damping vs pressure ratio
Industries performed
CSD analyses ( modes)
Steady CFD ( speedline)
y CFD (
( damping
p g at various OPs))
Unsteady
FUTURE
69
0.8%
0.2%
-0.3%
70
71
Managers
Manager
s vs Engineers
Engineer s Views
What is the
73
Key Challenges
Aerodynamic forcing
Correct prediction of forcing levels
Taking into account details (tip clearances, cavities, etc)
Aerodynamic damping
Correct prediction of damping levels
Strongly dependent on steady flow phenomena
Transition usually not modeled at all
Non-synchronous vibrations
Extremely difficult to delineate where to search for
Post-diction possible, pre-diction extremely challenging
U
Usually
ll iinvolving
l i
360d
360deg
models,
d l multi
lti row
Damping
Correctt prediction
C
di ti
off friction
f i ti
d
dampers and
d novell damping
d
i
concepts (coatings, air film, piezo, eddy current)
74
Key Challenges
Aerodynamic forcing
Correct prediction of forcing levels
Taking into account details (tip clearances, cavities, etc)
Aerodynamic damping
Correct prediction of damping levels
Strongly dependent on steady flow phenomena
Having
that are trained in
Transition usually not modeled at all
engineers
i
interdisciplinary analyses and problem solving
Non-synchronous vibrations
Damping
THRUST
Turbomachinery
y Training
g
Aeromechanical University
Correctt prediction
C
di ti
off friction
f i ti
d
dampers and
d
novell damping
d
i
www.explorethrust.eu
concepts (coatings, air film, piezo, eddy current)
75
76
Realistic Components
115m
115
+64%
64%
70m
the story
77
79
Summary
An overview over turbomachinery blade vibrations,
analyses techniques and challenges has been given
Despite the fact that we nowadays have very
sophisticated analysis tools, we are not in a position
to predict turbomachinery blade vibrations down to
single digit accuracies
Still, turbomachines have and will be designed with
these
h
methods
h d while
h l taking
k
into account
conservative safety margins
The future calls for top-of-the-line analyses taking
into account variability of engines and yielding
failure probabilities
80
mahalo
h l
81