Fischer SEA

S E A
The Statistical Energy Analysis (SEA)

by Michael Fischer JASS 2006 in St. Petersburg
1. Methods used for vibration problems:
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
wave-lengths, eigenmodes etc.

-The following introductory example shows, that at higher frequencies the
reliability of the result of calculation might be considerably reduced.
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
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

- 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

- 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 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

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)
...
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
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.
4. Motivation for SEA:
5. Deterministic approach: modal superposition
+ =
i
i i
pressure velocity,
contribution of the i.th mode
mode shape (point of observation)
system response
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 >
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)
6.2 Averaging in the SEA
- Now we increase the prediction accuracy by appropriate averaging
in several steps

6.3 Averaging over the points of observation ( Step 1)
- by this step the phase information gets lost

2
i
i i
2
v
|
|
.
|
\
|
+ =

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)
Orthogonality of modeshapes
6.4 Averaging over the points of excitation ( Step 2)
-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 dont have to be calculated!

2
4
2
2
2 2
2
2
1
i
i
V
i
V
i
i
f dV
dV F
V
o
t

|
|
.
|
\
|
+
+
=
}
}
) (
modal
mass
mean modal force
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
6.5 Averaging over the frequencies of excitation ( Step 3)
u
f
l
f
-To simplify the mean square velocity
once again, we assume several similar
modes N in a frequency band

A
) 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
) f f (
N
) f 2 ( m 8
F
2
v m
E
u o m
2
2
i
f
q t
~
~
A
force
total mass damping
frequency band
Energy within a certain frequency band:
centre frequency
6. Mean input power
force
total mass
frequency band
input power is independent from damping
) f f (
N
m 4
F
P
u o
2
_
=
( ) t F
k
c
-We are looking at one sub-system (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:
7. Balance of power- hydrodynamic analogy
Mean input power P
Energy E in the sub-system
Dissipated energy
out in
P P =
q t = E f 2 P
m diss
-every sub-system is considered as a
energy reservoir

-The dissipated energy
is proportional to the absolute dynamic
energy E of the sub-system:
damping
i i m diss , i
E f 2 P q t =
Expansion to coupled systems:

-For every sub-system holds:
out , i in , i
P P =
Expansion to coupled systems:

-Energy flow between two sub-systems:
|
|
.
|
\
|
q t =
j
j
i
i
i ij m ij
N
E
N
E
N f 2 P
modal energy
coupling loss factor
8. Equations of the SEA
The governing equations can be derived by considering:

the loss of energy by damping

the energy flow between every pair of sub-systems (coupling)
j , i
i j , j
diss , i in , i
P P P
E
=
+ =
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
-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
8. Conclusions and look into the future
-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
-example:
Rail-Impedance-Model RIM
Thank you for your attention!

Fischer SEA

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fischer SEA

Uploaded by

Copyright:

Available Formats

S E A

The Statistical Energy Analysis (SEA)

The Statistical Energy Analysis (SEA)

You might also like