Modal Strain Energy Approach for Localization of

Damage in Structures

Abstract
The area of damage detection and localization based on measured vibration data
received considerable attention in recent years. Different methodologies were
reported in the literature for identification and localization of damage using
optimization techniques and modal based approaches. In this paper, an attempt has
been made to localize the damage using modal strain energy approach. This is very
useful for further identification and quantification of damage in large scale
structures using optimization approaches as only the identified damaged elements
need to be considered as parameters for the objective function. To demonstrate the
proposed methodology, numerical studies have been carried out on selected
structures. The approach of localization of the damage using modal strain energy
change ratio is found to be very efficient.

INTRODUCTION

Deterioration and degradation of structures is a major problem due to large

number of ageing structures. It is often becoming necessary to evaluate the
integrity of such structural systems. Generally, structural damage may be
considered as weakening of the structure that reduces its performance.
Damage may also be defined as any deviation in the structure’s original
geometric or material properties that may cause undesirable stresses,
displacements, or vibrations in the structure. Damage of the structures has to
be identified to evaluate safety and to prevent potential catastrophic events,
to offer the possibility of extending its service life by repairing/retrofitting
components of the structure. Early damage detection and eventual estimation
of damage is important, as it forms the basis for decision regarding structural
repair and/or part-replacement.

Because of the reasons stated above, the methods of identification of damage

in structures is gaining importance in the recent past due to safety and
economic aspects. The technological advancements in computing and sensor
monitoring have motivated extensive research activity in this area. Studies
have been carried out by several researchers to detect and identify the
damage directly through the changes in measured vibration responses of the
structure. When a structure is under distress, either in local or global, there
would be reduction in stiffness and a decrease in the free energy stored in the
body (Salawu and Williams, 1995). Since the dynamic response is governed
by system parameters viz., stiffness, mass, and damping; changes in these
parameters would lead to changes in the vibrational response as
characterized by the modal parameters viz. natural frequencies, mode
shapes, and modal damping values. Due to the fact that each vibration mode
has a different energy distribution, any localized damage will affect each
mode differently depending on the location and severity of the damage.
Modal parameters are also sensitive to boundary conditions, i.e., physical
constraints of the structure. The dynamic testing allows the engineer to use
the structural responses in conjunction with appropriate analysis and
modelling techniques to assess the integrity of a structure.

Due to the relative ease of instrumentation and development of new

powerful system identification techniques, non-destructive identification of
structural damage through the changes in vibration characteristics has gained
increased world-wide attention over the past few years (Doebling 1996;
Doebling et.al., 1996, 1997). Early damage detection approaches use the
natural frequency changes as damage indicator (Salawu, 1997); this is
because the modal frequencies can be measured easily and accurately.
However, the frequencies are not spatially specific and are not very sensitive
to local damages. Hence, its application is severely limited to simple
structures. The mode shapes have the advantage of being spatially specific
(Salawu & Williams, 1994), but its measurement is more complex and
difficult and also not accurate.

Different methodologies have been proposed in the literature for

identification and localization of damage based on optimization techniques.
The drawback in using the optimization approaches is that they require the
evaluation of the objective function at each stage of generation for the total
population, which is computationally very intensive and problems of
convergence may occur. In this paper, an attempt has been made to identify
the location of damage first by using modal strain energy approach so as to
reduce the number of parameters for the objective function in the
optimization approach. Further, these identified damaged elements can be

2
used for exact identification and quantification of the damage using
optimization approach. This approach is very useful for identifying damage
in large scale structures. To demonstrate the proposed methodology,
numerical studies have been carried out on selected structures. The approach
of localizing the damage using modal strain energy change ratio is found to
be very efficient.

EFFECT OF STRUCTURAL DAMAGE

Structural damage may cause loss of stiffness in one or more elements of a

structure, but not a loss in the mass. In the finite element model (FEM) of the
skeletal structure, damage is assumed to affect only the stiffness matrix of
the system. Small changes in the stiffness of the system produce small
changes in the square of the modal frequencies i and the mode shapes  i.
When damage occurs in the structure, it can be represented as a small
perturbation in the original system. Thus, the global stiffness matrix Kd, the
ith modal eigenvalue id, and the ith mode shapes  id of the damaged
system can be expressed as

L L
Kd  K  j 1
ΔK j  K  α K
j 1
j j (-1 α j  0) (1a)

id  i  Δi (1b)

id  i  Δi (1c)

where superscript d denotes the damaged case,  j is the coefficient defining

a fractional reduction in the jth elemental global stiffness matrices; and L is
the total number of elements in the system.

MODAL STRAIN ENERGY METHOD FOR DAMAGE

IDENTIFICATION

Elemental modal strain energy (MSE) is defined as the product of the

elemental stiffness matrix and the second power of its mode shape
component. For the jth element and the ith mode, the MSE before and after
the occurrence of damage is given as
T
MSEij   Ti K j  i ; MSE ijd   id K j  id (2a,b)

3
 d
where MSEij and MSEij are the MSE of the jth element for the ith mode
shape for undamaged and damaged cases respectively. Because the damaged
elements are not known, the undamaged elemental stiffness matrix Kj is used
d
instead of the damaged one as an approximation in MSEij , and the modal
strain energy change ratio (MSECR), defined as follows, has been found to
be a good indicator for damage localization (Law et al. 1998, Shi et al.1998,
2002, 2000):

MSEijd  MSEij
MSECRij  (3)
MSEij

where j and i denote the element number and mode number, respectively.

The MSECR will be used to quantify damage. Here, the MSEC of the jth
element for the ith mode could be obtained from the mode shapes as
T
MSECij   id K j  id   Ti K j  i (4)

Substituting Eq.1(c) into Eq. (2) and neglecting the second-order terms, the
MSECij becomes
MSECij  2 iT K jΔ i (5)

For a small perturbation in an undamped n-degrees of freedom (n-DOFs)

dynamic system, the equation of motion becomes

[(K  ΔK) - (i  Δi )M] ( i  Δ i )  0 (6)

Neglecting second-order terms, Eq.(6) leads to

(K - i M)Δ i  Δi M i  ΔK i (7)

The term Δ i can be expressed as a linear combination of mode shapes of

the original system.
n
Δ i  d
k 1
ik  k (8)

4
where d ik =scalar factors; and n= total number of modes of the original n-
DOFs system. Substituting Eq.(8) into Eq. (7), and premultiplying Tr to
both sides of Eq. (7),
n

d
k 1
T
ik  r (K - i M ) k  Δi  Tr M i   Tr ΔK i (9)

With the orthogonal relationship, Eq. (9) can be simplified into the following
formula when r is not equal to i:

 Tr ΔK i
d ir   (r  i ) (10)
 r  i

For the case of r = i, it is easy to know that drr = 0 by the orthogonal

relationship  Tr M i  I . Therefore, Eq. (8) is rewritten as follows:

n
 Tr ΔK i
Δ i  
r 1

 r  i
r (r  i) (11)

Substituting Eq. (11) into Eq. (5) and neglecting higher-order terms, the
MSECij becomes

 n  T ΔK i 
MSEC ij  2 Tr K j  
 r
 r  i
 r  (r  i ) (12)
 r 1 

Substituting Eq. (1a) into Eq. (12),

L n  Tr K p  i
MSEC ij  
p 1
 2 T
pi K j 
r 1

 r  i
r (r  i ) (13)

The term on the left hand side of Eq. 13 is the elemental modal strain energy
change of the jth element in the ith mode, which can be calculated from
Eq.(4) by using the mode shape of the undamaged and damaged states. All
the terms on the right-hand side of Eq.13 except p are all known
information of the undamaged system. Solution of Eq.13 can be used to
identify the damage.

5
Application of modal strain energy change ratio

The damage localization method using modal strain energy change ratio
described above has been verified for localizing damage in a plane truss
structure shown in Fig. 1. The geometric details of the truss are given in
Fig.1 along with the node and element numbers used for simulation in finite
element model. The properties of the elements are: (i) area of cross-section
(A) = 1100 mm2, (ii) weight density () = 7860 kg/m3. It is well known that
the damage can significantly reduce the stiffness or increase the flexibility of
the structure. Hence, in the present study, the damage has been simulated by
changing the stiffness of the structure. Different damage scenarios have been
simulated on pre-selected elements by reducing the stiffness of these
elements. Free vibration analysis has been carried out to obtain the
frequency and mode shape information of the structure. The first five modes
have been considered in the damage identification processes which are
shown in Fig.2. The damage detection process has been carried out for
various damage scenarios viz., single as well as multi-damage of different
magnitudes.

The MSECR obtained using the above formulations are shown in Fig. 3 to 5
for identifying damages in different locations. The figures show the
identification of damage for multiple damage cases. The figures clearly show
the damaged elements with higher MSECR values for the three cases.
Further, damage quantification can be made by using these damaged
elements as solution parameters in genetic search optimization method
(Mares and Surace, 1996) for obtaining the stiffness reduction factors of the
damaged elements. Damage localization using this method is very useful for
detection of damage in large scale structures as number of solution
parameters required for genetic search optimization would get reduced for
damage detection.

CONCLUDING REMARKS

During the last two decades, vibration-based methods were developed and
applied to detect structural damage in many engineering applications.
Structural damage usually decreases structural stiffness, which produces
changes in the vibration characteristics of the structure. Identifying the

6
structural damage with the measured vibration data is an inverse problem.
The usual damage detection methods minimize the objective function, which
is defined in terms of the discrepancies between the vibration data identified
by modal testing and those computed from the analytical model. But, these
optimization methods are computationally intensive particularly for large
scale structures such as bridges, because of drastic increase in number of
solution parameters. In the present study, modal strain energy method has
been used to reduce the number of solution parameters by identifying the
damaged elements. Studies have been carried out on truss structure and the
method is found to be very effective in identifying the damaged elements
with sufficient accuracy. The technique can further be combined with
optimization procedures for quantifying the damage.

REFERENCES

Salawu, O.S., and Williams, C. (1995), “Review of full-scale dynamic

testing of bridge structures”, Engineering Structures, 17(2), 113-121.
Doebling, S.W. (1996), “Minimum-rank optimal update of elemental
stiffness parameters for structural damage identification”, AIAA Journal,
34(12), 2615-2621.
Doebling,S.W., Farrar, C.R., Prime,M.B. and Shevitz, D. (1996), “Damage
identification and health monitoring of structural and mechanical systems
from changes in their vibration characteristics: A literature review”, Los
Alamos National Laboratory, USA, Research report LA-13070-MS, 1996.
Doebling, S.W., Hemez, M.S., Peterson, L.D., and Farhat, C. (1997),
“Improved damage location accuracy using strain energy based mode
selection criteria”, AIAA Journal, 35(4), 693-699.
Salawu, O.S. (1997), “Detection of structural damage through changes in
frequency: A review”, Engineering Structures, 19(9), 718-723.
Salawu, O.S. and Williams, C. (1994), “Damage location using vibration
mode shapes”, Proceedings of 12th International Modal Analysis
Conference, 933-939.
Law, S. S., Shi, Z. Y. and Zhang, L. M. (1998), “Structural Damage
Detection from Incomplete and Noisy Modal Test Data”, Journal of
Engineering Mechanics, 124(11), 1280-1288.
Mares, C. and Surace, C. (1996), “An application of genetic algorithms to
identify damage in elastic structures”, Journal of Sound and Vibration,195,
195–215.
Shi, Z.Y., Law, S.S., and Zhang, L.M., (1998), “Structural damage localization
from modal strain energy change”, Journal of Sound and Vibration, 218(5),
825-844.

7
Shi, Z.Y., Law, S.S. and Zhang, L. M. (2000), “ Damage Localization by
Directly Using Incomplete Mode Shapes”, Journal of Engineering
Mechanics, 126(6), 656-660.
Shi, Z.Y., Law, S.S., and Zhang, L.M., (2002) “Improved damage
quantification from elemental modal strain energy change”, Journal of
Engineering Mechanics-ASCE, 128(5), 521-529.

Note : i - joint number; - element number

Fig. 1 Plane truss structure considered for simulation

Mode shape - 1 Mode shape - 2

Mode shape - 3 Mode shape - 4

Mode shape - 5

Fig. 2 First five mode shapes considered for damage identification

8
 0.25
Damage in Elements 7,14,15

0.20

0.15
MSECR

0.10

0.05

0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Element No.

Fig.3 MSECR for damage in elements 7,14 and 15

0.25

Damage in Elements 1,8,15,16

0.20

0.15
MSECR

0.10

0.05

0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Element No.

Fig.4 MSECR for damage in elements 1,8,15 and 16

9
 0.3

Damage in Elements 1,7,8,14,15,16

0.25

0.2
MSECR

0.15

0.1

0.05

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Element No

Fig.5 MSECR for damage in elements 1, 7, 8, 14, 15 and 16

10