This action might not be possible to undo. Are you sure you want to continue?
https://www.scribd.com/doc/154270747/FischerSEA
07/17/2013
text
original
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
1. Methods used for vibration problems:
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
1. Methods used for vibration problems:
Usually we are dealing with models like FEM, (BEM) and analytical models
which enable us to calculate for deterministic loads and defined model
parameters deterministic responses.
Typically the calculated value is given in detail with respect to frequency,
time and location.
However, the level of discretization of time/frequency and the geometric
data has to be defined at the basis of theoretical considerations regarding
wavelengths, eigenmodes etc.
The following introductory example shows, that at higher frequencies the
reliability of the result of calculation might be considerably reduced.
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
2. Introductory example:
 room (25 m3)
 limited by a steel plate
 one of the boundary surfaces is excited by a harmonic load
 18 points in the room are considered
by Michael Fischer JASS 2006 in St. Petersburg
2. Introductory example :
The figure shows for the 18 points
in the room all measured transfer
functions between the harmonic
load and the sound pressure.
lt can clearly be seen, that at
higher frequencies the transfer
functions differ considerably.
Hz
frequency Hz
L
e
v
e
l
d
i
f
f
e
r
e
n
c
e
s
o
u
n
d
p
r
e
s
s
u
r
e

h
a
r
m
o
n
i
c
f
o
r
c
e
Wheel of a bike
The Statistical Energy Analysis (SEA)
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
2. Introductory example:
 reason for the high differences:
different contributions of single modes
which are close together regarding
their eigenfrequency.
So e.g. in the centre of the room and a
tonal excitation at 250 Hz, a difference
of about 20 dB (factor 10) between the
individual functions is observed.
Hz
frequency Hz
L
e
v
e
l
d
i
f
f
e
r
e
n
c
e
s
o
u
n
d
p
r
e
s
s
u
r
e

h
a
r
m
o
n
i
c
f
o
r
c
e
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
2. Introductory example:
 Even slight temperatur differences in
the room, which practically cannot be
eliminated, influence the positions of the
Eigenfrequencies so that a detailed
prediction cannot be given
Hz
frequency Hz
L
e
v
e
l
d
i
f
f
e
r
e
n
c
e
s
o
u
n
d
p
r
e
s
s
u
r
e

h
a
r
m
o
n
i
c
f
o
r
c
e
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
2. Introductory example:
The air inside the room also shows
modes (starting at about 50 Hz)
mode
empty room
Frequency [Hz]
room with disturbing objects
Frequency [Hz]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
…
56,8
70,2
85,4
90,3
102,6
110,6
114,3
124,3
134,1
141,8
142,7
152,8
159,0
165,6
173,2
173,5
…
49,0
68,6
79,5
85,3
96,2
104,9
107,9
118,7
127,5
134,6
139,8
149,0
149,9
153,9
162,5
171,2
…
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
2. Introductory example:
Possible Uncertainties of…
• boundary conditions (e.g. clamped/free edge)
• dynamic material properties (e.g. concrete: E ~ 30kN/mm^2)
• masses of the materials (e.g. concrete: 25 kN/m^2)
• damping
• load distribution (e.g. position of the machine)
• frequency of excitation (e.g. velocity of train)
• ...
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
3. Historical example:
In the early 1960s:
prediction of the vibrational response to
rocket noise of satellite launch vecicles and
their payloads
problem: the frequency range of significant
response contained the natural frequencies of
a multitude of higher order modes:
the Saturn launch vehicle possessed about
500.000 natural frequencies
in the range 0 to 2000 Hz
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
4. Motivation for SEA:
The both examples above are leading to the insight that
at higher frequencies a method with less detailing has to be accepted.
A detailed analysis at the basis of FEM approach (input at a point of
excitation, output at a point of observation) would lead to results which
are very sensitive to slight changes in the input parameters
(factor 10!).
In order to obtain acceptable sensitivities of the results, but to describe
nevertheless the system response, we will give the results in an averaged
sense.
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
4. Motivation for SEA:
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
5. Deterministic approach: modal superposition
¿
+ · =
i
i i
¸ pressure velocity,
contribution of the i.th mode
mode shape (point of observation)
system response
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
5. Deterministic approach: modal superposition
q
t
t
¸
j
f
f
f j
f m
dV p
i
i i
V
i
i
+ ÷
·
·
+ ·
=
}
) 1 (
) 2 (
) 2 ( *
2
2 2
influence of the geometry
of excitation
amplification function
influence of the frequency of excitation
1
1 2
q
( ) q V
0 = D
0
1
> D
1 2
D D >
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.1 Shift to energy
2
i
i i
2
v


.

\

¸ + =
¿
In the first step a shift from velocities to energy is carried out.
the mean kinetic energy is proportional to the mean square velocity
contribution of the i.th mode
mode shape (point of observation)
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.2 Averaging in the SEA
The Statistical Energy Analysis (SEA)
 Now we increase the prediction accuracy by appropriate averaging
in several steps
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.3 Averaging over the points of observation („ Step 1“)
The Statistical Energy Analysis (SEA)
 by this step the phase information gets lost
2
i
i i
2
v


.

\

¸ + =
¿
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.3 Averaging over the points of observation („ Step 1“)
¿
}
¿
} }
¿
+ · o ·
t ·
+ ·
·
·
=
+ ¸ ·
·
=


.

\

¸ + ·
·
=
i
V
2
i
2
i
4
i
2
*
i
2
i
2
i
V
2
i
2
i
2
i
i i
2
dV
) f 2 ( m
F
V 2
1
dV
V 2
1
dV
V
V 2
1
v
(„Summing up the modal energy“)
The Statistical Energy Analysis (SEA)
Orthogonality of modeshapes
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.4 Averaging over the points of excitation („ Step 2“)
The Statistical Energy Analysis (SEA)
By this averaging, the information about the shape of the individual
eigenmodes is eliminated and has no longer to be considered
This means: the modes don‘t have to be calculated!
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.4 Averaging over the points of excitation („ Step 2“)
2
4
2
2
2 2
2
2
1
i
i
V
i
V
i
i
f dV
dV F
V
o
t µ
¸ ·
·


.

\

+ ·
+ ·
=
}
}
) (
modal
mass
mean modal force
The Statistical Energy Analysis (SEA)
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.4 Averaging over the points of excitation („ Step 2“)
The Statistical Energy Analysis (SEA)
v
F
m f
i
i
i
2
2
2
2 4
2
=
· o
t ( )
no information about the modes necessary!
total mass
amplification function
force
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.5 Averaging over the frequencies of excitation („ Step 3“)
The Statistical Energy Analysis (SEA)
u
f
l
f
To simplify the mean square velocity
once again, we assume several similar
modes N in a frequency band
A
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.5 Averaging over the frequencies of excitation („ Step 3“)
The Statistical Energy Analysis (SEA)
) f f ( 2
1
2 ) f 2 ( m
F
v
u o
i
2
2
2
i
÷ t
·
q · t ·
t ·
~
force
total mass damping frequency band
by Michael Fischer JASS 2006 in St. Petersburg
6. Energetic approach:
6.5 Averaging over the frequencies of excitation („ Step 3“)
The Statistical Energy Analysis (SEA)
) f f (
N
) f 2 ( m 8
F
2
v m
E
u o m
2
2
i
f
÷
A
·
q · t · ·
~
·
~
A
force
total mass damping
frequency band
Energy within a certain frequency band:
centre frequency
by Michael Fischer JASS 2006 in St. Petersburg
6. Mean input power
The Statistical Energy Analysis (SEA)
force
total mass
frequency band
input power is independent from damping
) f f (
N
m 4
F
P
u o
2
_
÷
A
·
·
=
( ) t F
k
c
We are looking at one „subsystem“ (frequency band)
We assume a steady state vibration:
„the mean input power, which is introduced during
one cycle of vibration equals to the dissipated power
due to damping“ (compare SDOF system).
mean input power in a frequency band:
by Michael Fischer JASS 2006 in St. Petersburg
7. Balance of power hydrodynamic analogy
The Statistical Energy Analysis (SEA)
Mean input power P
Energy E in the subsystem
Dissipated energy
by Michael Fischer JASS 2006 in St. Petersburg
7. Balance of power hydrodynamic analogy
out in
P P =
The Statistical Energy Analysis (SEA)
q · · t = E f 2 P
m diss
every subsystem is considered as a
energy reservoir
The dissipated energy
is proportional to the absolute dynamic
energy E of the subsystem:
damping
by Michael Fischer JASS 2006 in St. Petersburg
7. Balance of power hydrodynamic analogy
i i m diss , i
E f 2 P q · · t =
The Statistical Energy Analysis (SEA)
Expansion to coupled systems:
For every subsystem holds:
out , i in , i
P P =
by Michael Fischer JASS 2006 in St. Petersburg
7. Balance of power hydrodynamic analogy
The Statistical Energy Analysis (SEA)
Expansion to coupled systems:
Energy flow between two subsystems:


.

\

÷ · · q · t =
j
j
i
i
i ij m ij
N
E
N
E
N f 2 P
modal energy
coupling loss factor
by Michael Fischer JASS 2006 in St. Petersburg
8. Equations of the SEA
The Statistical Energy Analysis (SEA)
The governing equations can be derived by considering:
the loss of energy by damping
the energy flow between every pair of subsystems (coupling)
j , i
i j , j
diss , i in , i
P P P
E
=
+ =
by Michael Fischer JASS 2006 in St. Petersburg
8. Equations of the SEA
The Statistical Energy Analysis (SEA)
q q q q
q q q q
q q q
1 1
1
1 12 1 1 1
21 2 2 2
2
2 2 2
1
1
1
2
2
1
2
+

\

.
 ÷ ÷
÷ +

\

.
 ÷
÷ +

\

.

¸
(
¸
(
(
(
(
(
(
(
(
(
·
¸
(
¸
(
(
(
(
(
(
(
(
=
=
=
=
¿
¿
¿
i
i
k
k
i
i
k
k
k k k ik
i k
k
k
k
k
N N N
N N N
N N
E
N
E
N
E
N
P
P
( ) ... ( )
( ) ... ( )
... ... ... ...
( ) ... ...
...
...
P
k
¸
(
¸
(
(
(
(
damping coupling
by Michael Fischer JASS 2006 in St. Petersburg
8. Equations of the SEA
The Statistical Energy Analysis (SEA)
Related to the different possible deflection patterns
(e.g. bending, shear, torsional waves):
each part of the structure might appear as various energy reservoirs
and thus described by various governing equations.
FE: usually a high dicretization of the structure is necessary
SEA: based on calculation of global values
computational costs are much smaller
interactive planning by the engineer is possible
by Michael Fischer JASS 2006 in St. Petersburg
8. Conclusions and look into the future
The Statistical Energy Analysis (SEA)
Energy methods have a huge impact on the methodology of noise
and vibration prediction
especially hybrid methods can carry out vibroacoustic investigations
with a good confidence
by Michael Fischer JASS 2006 in St. Petersburg
8. Conclusions and look into the future
The Statistical Energy Analysis (SEA)
example:
by Michael Fischer JASS 2006 in St. Petersburg
8. Conclusions and look into the future
The Statistical Energy Analysis (SEA)
RailImpedanceModel RIM
by Michael Fischer JASS 2006 in St. Petersburg
Thank you for your attention!
The Statistical Energy Analysis (SEA)
This action might not be possible to undo. Are you sure you want to continue?