decentralize damage detection algorithm
DESCRIPTION
Decentralize damage detection algorithm. Manuel Ruiz-Sandoval & Cesar Carpio. Outline. Motivation Types of detection Modal energy deformation POD (Proper Orthogonal Decomposition). Proposed method Numerical example Conclusions. Motivation. - PowerPoint PPT PresentationTRANSCRIPT
Decentralize damage detection algorithm
Manuel Ruiz-Sandoval & Cesar Carpio
Outline
Motivation Types of detection
Modal energy deformation POD (Proper Orthogonal Decomposition).
Proposed method Numerical example Conclusions
Motivation
►The occurrence of The occurrence of structural damage can structural damage can originate threaten life originate threaten life situationssituations
►Damage detection at Damage detection at early stages could early stages could prevent the loss of prevent the loss of human lives, as well as human lives, as well as reduce maintenance reduce maintenance costcost
- Visual inspection- Optical sensing- Acoustic methods- Modal analysis- Damage Localization Vector- Among others
Damage detection methods
Damage detection methods based on signal processing
Modal parameters Frequency changes Mode shape changes
function of mass stiffness damping
Comparison between undamaged and damage stages
Traditional data acquisition systems use a centralized scheme.
This system is required to be capable of manage all channels
Cost could hinder the use a great number of sensors
Centralized data information
Wire sensors
►New technologies New technologies availableavailable►Smart sensors: on Smart sensors: on board processing and board processing and wireless communication.wireless communication.►New paradigm to be New paradigm to be exploreexplore
Decentralized data acquisition system
Wireless sensors
Damage detection methods
Modal energy deformation methodModal energy deformation method
jTjj KU
2
1
Where Where
is the jis the jthth the mode shape vector, the mode shape vector, and and
is the stiffness matrixis the stiffness matrix
Modal energy deformation methodModal energy deformation method
The total amount of deformation energy can be The total amount of deformation energy can be visualized as the sum of the energies of all visualized as the sum of the energies of all structural elements.structural elements.
N
iijj UU
1
Where U is the contribution of the energy deformation method of element i in the j th mode, and N is the number of structural elements.
The change if the energy between an undamaged (u) and damage (d) case can be calculated with the following expression ijijij UdUuU
2D Truss
6
1 27
3 4 5
1 2 3
4 5
6 7 8 9 10 11
DOF 1,2 DOF 3,4
DOF 5,6
DOF 7,8
DOF 9,10
Two damage scenariosTwo damage scenarios
1) 60% Reduction of Young's modulus at element 2.1) 60% Reduction of Young's modulus at element 2.
2) 2) 60% Reduction of Young's modulus at elements60% Reduction of Young's modulus at elements 7 y 10.7 y 10.
61 2
7
3 4 5
1 2 3
4 5
6 7 8 9 10 11
61 2
7
3 4 5
1 2 3
4 5
6 7 8 9 10 1110
UndamagedUndamaged
CASE 1)CASE 1)
Undamaged stiffness matrix for each element 1, 2, … 11Undamaged stiffness matrix for each element 1, 2, … 11
K
……
DamagedDamaged
PeriodsPeriods
j
jTjj KU
2
1
ijijij UdUuU
CASE 2)CASE 2)
K
……
UndamagedUndamaged DamagedDamaged
PeriodsPeriods
Undamaged stiffness matrix for each element 1, 2, … 11Undamaged stiffness matrix for each element 1, 2, … 11
ijijij UdUuU
Planar frame
Element 4 with a 70% reduction of stiffness
Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6 Mode 7 Mode 8 Mode 9 Mode 10 Mode 11 Mode 12
undamagedj
gdl 1gdl 2gdl 3gdl 4gdl 5gdl 6gdl 7gdl 8gdl 9gdl 10gdl 11gdl 12
damagejgdl 1gdl 2gdl 3gdl 4gdl 5gdl 6gdl 7gdl 8gdl 9gdl 10gdl 11gdl 12
j
Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6 Mode 7 Mode 8 Mode 9 Mode 10 Mode 11 Mode 12
ijU
123456
jTjj KU
2
1
Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6 Mode 7 Mode 8 Mode 9 Mode 10 Mode 11 Mode 12
El POD is a tool for the dynamics and vibration (also known as Karhunen-Loève) that provides with a base for the modal response during a experiment. The POD based its results on sensor displacements over the structure. It compares an original stated with an unknown state.
Also, it can use eigenvectors and eigenvalues to determine the distribution of modal energy, as well as the energy participation of every mode.
Proper Orthogonal Decomposition (POD)
_
iii dda
POD (Proper Orthogonal Decomposition)
di(t) = (di(t1), di(t2), di(t3),…. di(tM))T
Displacement history at different point over the structures are needed
These values are normalized by the mean value
With matrix A, a correlation matrix R can be calculated
R matrix is real and symmetrical of N x N order
)t(a)t(a)t(a
)t(a)t(a)t(a
)t(a)t(a)t(a
A
MNMM
N
N
21
22221
11211
AAMR T/1
Matrix A is constructed of M x N (time Vs Number of sensor positions)
POD (Proper Orthogonal Decomposition)
The correlation matrix represents the behavior of the structure at certain points.
The second step for damage identification is to do a ratio between R undamaged and damage for every corner sensor. Those point with a great difference will be indicative of presence of damage at that place
POD Method
Decentralization
Sensors are collocated at joints and midsections
L
L
L L
L/2 L/2 L/2 L/2
L/2
L/2
Smart sensors
L/2L
Beam element
Column element
L/2L/2
L
Sensor placement
1 2 3
16
5
4
6 7 8 9 10 11
12 13 14 15
17 18 19 20 21 22
23
27 2928
24
30
25
31 3332
26
First level of sharing
Sensor used for damage detection
33312927
222018
11`975
16
Cluster head sharing
CBA
FED
Master nodes
CBA
FED
Master nodes
Numerical example
Numerical example
A 20% reduction of stiffness is place at element between nodes 5, 12 and 16
1 2 3
16
5
4
6 7 8 9 10 11
12 13 14 15
17 18 19 20 21 22
23
27 2928
24
30
25
31 3332
26
Ruido blanco
-15
-10
-5
0
5
10
15
0 5 10 15 20 25 30 35 40 45 50Tiempo (seg)
Ac
ele
rac
ión
(c
m/s
eg
2 )Excitation: white
noise
Displacement matriz with/without damage is presented
(a) (b)
Tiempo (seg) Sensor 1 Sensor 5 Sensor 6 Sensor 121 0.01 -1.00E-04 -1.00E-04 -1.00E-04 -1.00E-042 0.02 -0.0002 -0.0003 -0.0003 -0.00043 0.03 -1.00E-04 -2.00E-04 -2.00E-04 -2.00E-044 0.04 0.00E+00 1.00E-04 1.00E-04 1.00E-045 0.05 1.00E-04 1.00E-04 1.00E-04 1.00E-046 0.06 1.00E-04 1.00E-04 1.00E-04 2.00E-047 0.07 1.00E-04 1.00E-04 1.00E-04 1.00E-048 0.08 0.00E+00 0.00E+00 0.00E+00 -1.00E-049 0.09 -0.0001 -0.0002 -0.0002 -0.000310 0.1 -1.00E-04 -2.00E-04 -2.00E-04 -2.00E-04: : : : : :: : : : : :
4991 49.91 0.00E+00 -1.00E-04 -1.00E-04 -1.00E-044992 49.92 -1.00E-04 -1.00E-04 -1.00E-04 -2.00E-044993 49.93 -1.00E-04 -1.00E-04 -1.00E-04 -2.00E-044994 49.94 0.00E+00 0.00E+00 0.00E+00 0.00E+004995 49.95 -1.00E-04 -1.00E-04 -1.00E-04 -1.00E-044996 49.96 -1.00E-04 -2.00E-04 -2.00E-04 -2.00E-044997 49.97 -1.00E-04 -1.00E-04 -1.00E-04 -1.00E-044998 49.98 0.00E+00 -1.00E-04 -1.00E-04 -1.00E-044999 49.99 -0.0001 -0.0002 -0.0002 -0.00035000 50 -0.0001 -0.0002 -0.0002 -0.0003
-1.61E-08 -3.22E-08 -3.22E-08 -3.95E-08
Sensor 5
M =
5000
d i sin daño
N = 4
d i promedio
Sensor 1 Sensor 5 Sensor 6 Sensor 12-1.00E-04 -1.00E-04 -1.00E-04 -1.00E-04-0.0002 -0.0003 -0.0003 -0.0004
-1.00E-04 -2.00E-04 -2.00E-04 -2.00E-040.00E+00 1.00E-04 1.00E-04 1.00E-041.00E-04 1.00E-04 1.00E-04 1.00E-041.00E-04 1.00E-04 1.00E-04 2.00E-041.00E-04 1.00E-04 1.00E-04 1.00E-040.00E+00 -1.00E-04 -1.00E-04 -1.00E-04-0.0001 -0.0002 -0.0002 -0.0003
-1.00E-04 -2.00E-04 -2.00E-04 -2.00E-04: : : :: : : :
0.00E+00 0.00E+00 0.00E+00 -1.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -1.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -2.00E-040.00E+00 -1.00E-04 -1.00E-04 -1.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -2.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -2.00E-040.00E+00 -1.00E-04 -1.00E-04 -1.00E-040.00E+00 -1.00E-04 -1.00E-04 -1.00E-04-0.0001 -0.0002 -0.0002 -0.0003-0.0001 -0.0002 -0.0003 -0.0003
-1.72E-08 -3.42E-08 -3.44E-08 -4.28E-08
d i promedio
Sensor 5
d i con daño
N = 4
Aplication of POD at sensor 5 (node 9).
Matrix A for sensor 5
(a) (b)
Sensor 1 Sensor 5 Sensor 6 Sensor 12
1 -1.00E-04 -1.00E-04 -1.00E-04 -1.00E-042 -2.00E-04 -3.00E-04 -3.00E-04 -4.00E-043 -1.00E-04 -2.00E-04 -2.00E-04 -2.00E-044 1.61E-08 1.00E-04 1.00E-04 1.00E-045 1.00E-04 1.00E-04 1.00E-04 1.00E-046 1.00E-04 1.00E-04 1.00E-04 2.00E-047 1.00E-04 1.00E-04 1.00E-04 1.00E-048 1.61E-08 3.22E-08 3.22E-08 -1.00E-049 -1.00E-04 -2.00E-04 -2.00E-04 -3.00E-0410 -1.00E-04 -2.00E-04 -2.00E-04 -2.00E-04: : : : :: : : : :
4991 1.61E-08 -1.00E-04 -1.00E-04 -1.00E-044992 -1.00E-04 -1.00E-04 -1.00E-04 -2.00E-044993 -1.00E-04 -1.00E-04 -1.00E-04 -2.00E-044994 1.61E-08 3.22E-08 3.22E-08 3.95E-084995 -1.00E-04 -1.00E-04 -1.00E-04 -1.00E-044996 -1.00E-04 -2.00E-04 -2.00E-04 -2.00E-044997 -1.00E-04 -1.00E-04 -1.00E-04 -1.00E-044998 1.61E-08 -1.00E-04 -1.00E-04 -1.00E-044999 -1.00E-04 -2.00E-04 -2.00E-04 -3.00E-045000 -1.00E-04 -2.00E-04 -2.00E-04 -3.00E-04
Sensor 5
Matriz A sin dañoM
=50
00
Sensor 1 Sensor 5 Sensor 6 Sensor 12
-1.00E-04 -1.00E-04 -1.00E-04 -1.00E-04-2.00E-04 -3.00E-04 -3.00E-04 -4.00E-04-1.00E-04 -2.00E-04 -2.00E-04 -2.00E-041.72E-08 1.00E-04 1.00E-04 1.00E-041.00E-04 1.00E-04 1.00E-04 1.00E-041.00E-04 1.00E-04 1.00E-04 2.00E-041.00E-04 1.00E-04 1.00E-04 1.00E-041.72E-08 -1.00E-04 -1.00E-04 -1.00E-04-1.00E-04 -2.00E-04 -2.00E-04 -3.00E-04-1.00E-04 -2.00E-04 -2.00E-04 -2.00E-04
: : : :: : : :
1.72E-08 3.42E-08 3.44E-08 -1.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -1.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -2.00E-041.72E-08 -1.00E-04 -1.00E-04 -1.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -2.00E-04-1.00E-04 -1.00E-04 -1.00E-04 -2.00E-041.72E-08 -1.00E-04 -1.00E-04 -1.00E-041.72E-08 -1.00E-04 -1.00E-04 -1.00E-04-1.00E-04 -2.00E-04 -2.00E-04 -3.00E-04-1.00E-04 -2.00E-04 -3.00E-04 -3.00E-04
Matriz A con daño
Sensor 5
Undamaged Damaged
3.15E-08 6.28E-08 6.29E-08 7.75E-086.28E-08 1.25E-07 1.25E-07 1.54E-076.29E-08 1.25E-07 1.26E-07 1.55E-077.75E-08 1.54E-07 1.55E-07 1.90E-07
R Sensor 5 sin daño
3.07E-08 6.11E-08 6.13E-08 7.67E-086.11E-08 1.22E-07 1.22E-07 1.53E-076.13E-08 1.22E-07 1.23E-07 1.53E-077.67E-08 1.53E-07 1.53E-07 1.92E-07
R Sensor 5 con daño
1 2 3
16
5
4
6 7 8 9 10 11
12 13 14 15
17 18 19 20 21 22
23
27 2928
24
30
25
31 3332
26
R undamaged R damaged
3.16E-08 6.30E-08 6.30E-08 6.31E-08 7.77E-086.30E-08 1.26E-07 1.26E-07 1.26E-07 1.55E-076.30E-08 1.26E-07 1.25E-07 1.26E-07 1.55E-076.31E-08 1.26E-07 1.26E-07 1.26E-07 1.55E-077.77E-08 1.55E-07 1.55E-07 1.55E-07 1.91E-07
R Sensor 7 sin daño
3.11E-08 6.18E-08 6.20E-08 6.21E-08 7.72E-086.18E-08 1.23E-07 1.23E-07 1.23E-07 1.53E-076.20E-08 1.23E-07 1.23E-07 1.24E-07 1.54E-076.21E-08 1.23E-07 1.24E-07 1.24E-07 1.54E-077.72E-08 1.53E-07 1.54E-07 1.54E-07 1.92E-07
R Sensor 7 con daño
1 2 3
16
5
4
6 7 8 9 10 11
12 13 14 15
17 18 19 20 21 22
23
27 2928
24
30
25
31 3332
26
R undamaged R damaged
Sensor point i R undamged / R damage
5 0.98137 0.98649 0.99311 0.996116 0.979418 0.984322 0.986227 0.981329 0.984531 0.984733 0.9835
0.9794 16 0.9843 18 0.9835 33
0.9813 27 0.19 0.9845 29 0.02 0.9847 31 0.12
0.9843 18 0.31 0.9847 31 0.02 0.9862 22 0.15
0.9845 29 0.02 0.9863 20 0.16 0.9863 20 0.01
0.9794 16 0.9843 18 0.9862 22
0.9813 5 0.19 0.9863 20 0.20 0.9863 20 0.01
0.9843 18 0.31 0.9864 7 0.01 0.993 9 0.68
0.9864 7 0.21 0.993 9 0.67 0.9961 11 0.31
C
FE
BA
D
16
5 7 9 11
18 20 22
27 29 31 33
A B C
D E F
18
16
16
5
16
Beam
BeamBeam
0
0.2
0.4
0.6
0.8
1
30%40%50%60%70%80%90%100%
Porcentaje de daño
Ind
ice
Nodo 1,5,9
Nodo 2,6,11
Nodo 9,16,20
Nodo 13,18,24
Nodo 20,27,31
Nodo 26,30,37
Índice de daño estructural para cada barra del marco I.
Inde
x
Stiffness reduction
Frame II.
300 cm
300 cm
400 cm
400 cm600 cm400 cm
31 373635343332
26252423222120
27 302928
19181716
1514131211109
5 6 7 8
4321
1 2 3
16
5
4
6 7 8 9 10 11
12 13 14 15
17 18 19 20 21 22
23
27 2928
24
30
25
31 3332
26
37
43 4442
36
41
35
39 4038
34
39 40 4138
43 44 45 46 47 4842
300 cm
0
0.2
0.4
0.6
0.8
1
1.2
30%40%50%60%70%80%90%100%
Porcentaje de daño
Ind
ice
Nodo 1,5,9
Nodo 2,6,11
Nodo 9,16,20
Nodo 11,17,22
Nodo 20,27,31
Nodo 22,28,33
Nodo 31,38,42
Nodo 33,39,44
300 cm
300 cm
400 cm
400 cm600 cm400 cm
31 373635343332
26252423222120
27 302928
19181716
1514131211109
5 6 7 8
4321
1 2 3
16
5
4
6 7 8 9 10 11
12 13 14 15
17 18 19 20 21 22
23
27 2928
24
30
25
31 3332
26
37
43 4442
36
41
35
39 4038
3439 40 4138
43 44 45 46 47 4842
300 cm
Inde
x
Stiffness reduction
Frame III
32
39 4038
31
37
30
35 3634
29
262524232221
18171615
11109876
4
20
321
1 2 3 4
8765 10
11 12 13 14 15
20 21 22 23
35 36 3734
25 26 27 28 30 31
40 41 42 43 44 4539
500 cm 500 cm 500 cm
400 cm
300 cm
300 cm
500 cm
4746
3332
38
24
1716
5
5
13 14
19
27 28
4241
33
12
18 19
29 W18x35Vigas
ColumnasW24x55
0
0.2
0.4
0.6
0.8
1
1.2
30%40%50%60%70%80%90%100%
Porcentaje de daño
Ind
ice
Nodo 1,6,11
Nodo 2,7,13
Nodo 3,8,15
Nodo 11,20,25
Nodo 13,21,27
Nodo 15,22,29
Nodo 25,34,39
Nodo 27,35,41
Nodo 29,36,43
The use of smart sensor can allow implementing a decentralization of damage detection method.
Most the actual methods are use in a centralized fashion.
This works explore some of the existent methods and how to decentralize them.
CONCLUSIONS
Method is applied to planar shear deformation frame.
Modal Energy method was not able to detect damage for this specific case.
Proper Orthogonal Decomposition method was able to detect damage.
CONCLUSIONS
A proposal to decentralized POD methods is presented.
Only information of cluster head are required to determine damage.
This method detects damage for small stiffness changes at low level columns.
Damage detection for upper columns is achieve for relatively large change of stiffness.
CONCLUSIONS