0 views

Uploaded by Anjneya Srivastava

cvxb

- MIT2_092F09_lec06.pdf
- 1-s2.0-S0022460X15010512-main.pdf
- 45909725-Mechanica-Tutorial
- Chapter 2_Finite Element Procedure
- Content AE675
- An Optimization of Rubber Bushing Material and Structural Parameters
- How we might be able to understand the brain 2004
- Finite Element Meshing
- Mitres2 002s10 Toc
- AcuSolve_Radioss_FSI
- Nisa Pro Shape
- TSAT_Model_Manual.pdf
- A comparison of neural network, statistical mathods, and variable choice for live insurrers' financial distress prediction
- Evaluation-Delft-1.pdf
- 4. DOE Language and Concepts
- ISPRS Cardaci Mirabella Versaci
- Prilog Na Spektrite Od Robot
- Introduction 1
- IRJET- Analysis of RCC Culvert by using Software
- DT091_Approvedmodules08[1]

You are on page 1of 11

Engineering Structures

journal homepage: www.elsevier.com/locate/engstruct

surface method

Wei-Xin Ren ∗ , Hua-Bing Chen

Department of Civil Engineering, Central South University, Changsha, 410075, China

National Engineering Laboratory for High Speed Railway Construction, Changsha, 410075, China

Article history: Fast-running response surface models that approximate multivariate input/output relationships of time-

Received 3 February 2009 consuming physical-based computer models enable effective finite element (FE) model updating analyses.

Received in revised form In this paper, a response surface-based FE model updating procedure for civil engineering structures in

6 April 2010

structural dynamics is presented. The key issues to implement such a model updating are discussed such

Accepted 7 April 2010

Available online 4 May 2010

as sampling with design of experiments, selecting the significant updating parameters and constructing

a quadratic polynomial response surface. The objective function is formed by the residuals between

Keywords:

analytical and measured natural frequencies. Single-objective optimization with equal weights of natural

Response surface frequency residual of each mode is used for optimization computation. The proposed procedure is

Finite element illustrated by a simulated simply supported beam and a full-size precast continuous box girder bridge

Model updating tested under operational vibration conditions. The results have been compared with those obtained from

Optimization the traditional sensitivity-based FE model updating method. The real application to a full-size bridge has

Design of experiment demonstrated that the FE model updating process is efficient and converges fast with the response surface

Regression analysis to replace the original FE model. With the response surface at hand, an optimization problem is formulated

explicitly. Hence, no FE calculation is required in each optimization iteration. The response surface-based

FE model updating can be easily implemented in practice with available commercial FE analysis packages.

© 2010 Elsevier Ltd. All rights reserved.

fine a mathematical model of a physical structure. Basically, FE

Nowadays the finite element (FE) method has become an model updating is an inverse problem to identify or correct the

important and practical numerical analysis tool. It is commonly uncertain parameters of FE models. It is usually posed as an op-

used in almost all areas of engineering analysis. However, the timization problem. Setting-up of an objective function, selecting

FE model of a structure is normally constructed on the basis of updating parameters and using robust optimization algorithm are

highly idealized engineering blueprints and designs that may not the three crucial steps in FE model updating. In a model updating

truly represent all the aspects of an actual structure. As a result, process, not only the satisfactory correlation is required between

the analytical predictions from a FE model often differ from the analytical and experimental results, but also the updated param-

results of a real structure. These discrepancies originate from the eters should preserve the physical significance. The updated FE

uncertainties in simplifying assumptions of structural geometry, models are used in many applications for civil engineering struc-

materials as well as inaccurate boundary conditions. It is often

tures such as damage detection, health monitoring, structural con-

required to update or calibrate the uncertain parameters of a FE

trol, structural evaluation and assessment.

model that leads to the better predictions of the responses of an

Finite element model updating is a topic of significant interest

actual structural.

in the field of structural dynamics. A number of FE model updating

Finite element model updating is such a procedure that mod-

ifies or updates the uncertainty parameters in the initial finite methods in structural dynamics have been proposed. The direct

element model based on the experimental results so that a more updating methods compute a closed-form solution for the global

realistic or refined model can be achieved [1]. In other words, FE stiffness and/or mass matrices using the structural equations

of motion and the orthogonality equations. These non-iterative

methods that directly update the elements of stiffness and mass

∗ Corresponding author at: Department of Civil Engineering, Central South matrices are one-step procedures [2,3]. The resulting updated

University, Changsha, 410075, China. Tel.: +86 731 82654349; fax: +86 731

matrices reproduce the measured structural modal properties well

85571736. but do not generally maintain structural connectivity and the

E-mail addresses: renwx@mail.csu.edu.cn, renwx@yahoo.com (W.-X. Ren). corrections suggested are not always physically meaningful.

0141-0296/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.

doi:10.1016/j.engstruct.2010.04.019

2456 W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465

The iterative parameter updating methods involve using the surface method for structural finite element model updating is

sensitivity of the parameters to find their changes. The sensitivity- somewhat new, especially with the civil engineering communities.

based FE model updating methods are often posed as optimization This paper is intended to present a response surface-based

problems. These methods set the errors of the structural response finite element model updating procedure in structural dynamics

features between analytical and experimental results as an and to take advantages of response surfaces for the FE model

objective function and try to minimize the objective function by updating of civil engineering structures in practice. Its purpose

making changes to the pre-selected set of physical parameters of is to estimate the values of structural parameters (moment of

the FE model. Link [4] gave a clear overview of the sensitivity- inertia, cross-sectional area and modulus of elasticity) based

based updating methods in structural dynamics. It is noted that on the measured response quantities (natural frequencies). The

the mathematical model used in the model updating is usually proposed procedure is based on constructing quadratic response

ill posed and the special attention is required for an accurate surfaces. Those surfaces represent an estimate for the relation

solution [5]. Jaishi and Ren [6–8] used either single-objective or between the unknown parameters of the finite element model

multi-objective optimization technique to update the FE models of and response quantities of interest. With the response surface at

civil engineering structures in structural dynamics. However, the hand, an optimization problem, whose solution is the estimate for

sensitivity-based method involving in the determination of local the values of the structural parameters, is formulated explicitly.

gradients at points may cause not only computational intensive, Hence, no further finite element simulations are required. The

but also convergence difficulty. objective function is formed by the residuals between analytical

If the structure of interest is represented by, e.g. a large finite and measured natural frequencies. Single-objective optimization

element model, the large number of computations involved can with equal weights of each natural frequency is implemented for

rule out many approaches due to the expense of carrying out optimization computation. The presented procedure is illustrated

many runs. For such a large FE model where so many elements by a simulated simply supported beam and a full-size precast

are involved, there are huge of both geometric and physical continuous box girder bridge tested under operational vibration

possible parameters to be updated. In addition, there are now conditions. The results have been compared with those obtained

many commercial finite element analysis packages available such from the traditional sensitivity-based FE model updating method.

as ANSYS, ABAQUS and SAP2000 et al.. The structural FE models The real application to a full-size bridge has demonstrated that the

are often constructed by using these packages. In all the iterative response surface-based FE model updating procedure is simple and

parameter updating methods, each iteration needs to go back fast so that it can be easily implemented in practice.

to run the finite element analysis package with any parameter

updated, which limits the popular applications of structural FE

2. Response surface-based finite element model updating

model updating in practice.

One way of circumnavigating the time-consuming and FE

Response surface-based finite element model updating is an

analysis package-related problems during the sensitivity-based

approach to achieve the global approximations of the structural

model updating is to replace the FE model by an approximate

response feature objectives and constrains based on functional

surrogate/replacement meta-model that is fast-running and less

evaluations at various points in the design space. It often involves

parameters involved [9]. Such a meta-model is easier to compute

experimental strategies, mathematical methods, probability and

with, because it is controlled only by a few explanatory variables.

statistical inference that enable an experimenter to make efficient

The FE model updating is carried out on the meta-model instead

empirical exploration of the structure of interest. The flowchart of

of the analytical FE model. Response surface is one of the

response surface-based finite element model updating is shown in

commonly used meta-models. Response surface methodology is

Fig. 1. The main steps include the following.

originally an experimental design approach for selecting design

parameters for experiments with the objective of optimizing • The selection of the structural parameters and the definition of

some function of a response [10–12]. It provides a mechanism a number of ‘‘level’’ for each selected parameters by using the

for guiding experimentation in search of optimal settings for design of experiments (DOE) techniques.

design parameters or optimal values of unknown response. Many • In design space, the response features are calculated by carrying

additional applications (largely a consequence of the increased out finite element analysis (FEA).

use of computational analyses) have broadened the range of • Performing the final regression followed by a regression error

application of response surface methods in the statistical and analysis to create the response surface model of the structure.

engineering literature. Recent literature has addressed more • The response features of the structure are measured and

flexible functional forms for modeling the response, new methods corresponding objective functions (feature residuals) to be

of response surface construction [13], alternate approaches to minimized are constructed.

updating the surface estimate [14], new sampling methods [15], • The finite element model updating (iteration) is carried

etc. out within the established response surface model. Updated

In many fields of engineering, the term ‘‘response surface’’ parameters are obtained and transferred to the original finite

is used synonymously with ‘‘meta-model’’ or ‘‘surrogate model’’, element model.

which refer to any relatively simple mathematical relationship

between parameters and a response, often based on limited

data [16]. In the case of structural finite element model updating, 2.1. Sampling and parameter selection

once the response surface of a structure has been constructed,

updating the model is reduced to the task of finding the To create a response surface that will serve as a surrogate for

smallest value on the response surface. The parameter values that the FE simulation model, the basic process is one of calculating

correspond to this smallest value are those that are used to update predicted values of the response features at various sample points

the model. Recently, the response surface that is represented by in the parameter space by performing a experiment at each of

a simple least-squares multinomial model has been adopted in those points. A number of feature values from the experiment

structural FE model updating, verification and validation [17–20]. ran across the parameter domain are fit with a response surface.

The response surface method for damage detection and The key is to select the parameters carefully, to minimize the

reliability analysis is not quite new [21,22]. However, the response number of dimensions in the parameter space, and then to select

W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465 2457

features of a structure is an important issue in FE model updating.

The structural parameters selected for updating should be able to

clarify the ambiguity of the model, and in that case it is necessary

for the model response to be sensitive to these parameters.

The problem always arises: How many parameters should be

selected and which parameters from many possible parameters

are used in the FE model updating? If too many parameters

are included in the FE model updating in structural dynamics,

the optimized problem may appear ill-conditioned because only

limited vibration modes can be effectively identified from the

field vibration measurements. The parameter selection requires

a considerable physical insight into the target structure, and

trial-and-error approaches are often used with different set of

selected parameters for complicated structures. In the frame of

response surface-based FE model updating, the selected updating

parameters can be evaluated from the sampled data by performing

a parameter effect analysis (hypothesis testing) based on statistical

variance (square of the standard deviation) analysis.

Compared with tests on means for a hypothesis testing, tests

on variances are rather sensitive to the normality assumption.

Variance analysis formulates all individual square of deviations of

sample data. The basic idea of variance analysis is to divide the total

square deviation of sampling features into two parts: SA —a square

of deviation caused by design parameter A (system deviation) and

Se —a square of deviation caused by the experiment.

The F-test method [12] can be used to carry out the hypothesis

testing to check the significance of parameter A:

Fig. 1. Flowchart of response surface-based FE model updating.

SA /fA

FA = ∼ F (fA , fe ) (1)

Se /fe

the combinations of parameter values where the experiment is

performed. where fA and fe are the degrees of freedom of SA and Se respectively.

The term experiment herein refers to either physical experi- For a given significance level p, if FA ≥ F1−p (fa , fe ), the effect of

ments or computer experiments. The planning of experimentation parameter A is significant.

is further referred to design of experiments (DOE). The selection of

sample points is related to the accuracy and cost of a response sur- 2.2. Response surface regression

face to be constructed. Less sample points may reduce the surface

accuracy, while more sample points may improve the surface accu- The family of response surface forms selected to represent

racy but increase the work load. In the real application, the sample a response can have a substantial impact on the results of an

points mainly depend on the problem to be solved, the response analysis. The selected response surface form should be capable

feature values of interest and the selected method of DOE. of attaining surfaces that meet specific smoothness requirements

DOE plays an important role in constructing a response surface. of an application. There is often a balance between assumptions

Two efficient DOE methods are commonly used. They are central and data requirements. Different surface families may be preferred

composite design (CCD) method and D-optimal design method. for different applications. In the case of structural finite element

There are several other DOE methods that can be used in model updating in structural dynamics, polynomials are popular

constructing a response surface such as box-behnken design, Latin forms representing a response surface because the calculations

Square design, fractional factor design [12]. For the purpose of are simple and the resulting function is closed-form algebraic

structural finite element model updating in structural dynamics, expression. For example, a quadratic polynomial response surface

Guo and Zhang [18] found that both CCD and D-optimal design has the form:

methods can achieve almost the same accuracy in the creation of

k k X

k

polynomial surfaces. In the current study, the CCD method in DOE X X

is used in the paper as it is simple in constructing the response y = β0 + βi xi + βij xi xj (2)

i =1 i=1 j=1

surfaces of a polynomial type.

Central composite design uses the orthogonal table to perform where β0 , βi , βij are the regression coefficients to be estimated

the experimentation to determine the sample points of selected from the experimental data. In such a way, a response surface

parameters. It contains a fractional factorial design 2k (levels y is represented as a function of the parameters or variables x

are ±1 and k is quantity of factors) with central points that of the design space (generally some subset of Euclidian k-space,

are augmented with a group of 2k star points that allow for Rk ). In the case of k = 2, the vector of surface parameters is six

the estimation of curvature. 2k star points are (±α, 0, . . . , 0), dimensional.

(0, ±α, . . . , 0), . . ., (0, 0, . . . , ±α ). For the purpose of constructing The number of sampling points n must be greater than or

high-order surfaces, another 2k star points (±α1 , 0, . . . , 0), equal to the number of terms in the polynomial. When there are

(0, ±α1 , . . . , 0), . . ., (0, 0, . . . , ±α1 ) are appended. The precise more sampling points, the equation is over-determined and the

values of α and α1 rely on certain properties such as orthogonality regression techniques are required to fit the response surface to the

and rotatability desired for the design and on the number of factors sampled data. In this case, the surface does not, in general, exactly

involved [12]. match the response values at every sample points. The method

2458 W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465

Table 1

Natural frequency differences of simply supported beam before and after updating.

Mode Undamaged beam (Hz) Damaged beam (Hz) Difference (%) Updated (Hz) Error (%)

2 35.92 33.24 8.04 33.25 0.04

3 80.63 79.14 1.89 79.15 0.01

4 142.93 128.04 11.63 128.09 0.04

5 149.14 146.22 2.00 146.07 −0.10

process to create a response surface.

xlk ≤ xk ≤ xuk , k = 1, 2, 3, . . .

Before the regressed response surface is put forward to be used

in structural FE model updating, it should be verified to check where xlk and xuk are the upper and lower bounds on the kth

whether the regressed surface has enough accuracy. If not, the design variable to be set. N is the total number of design variables

parameters are adjusted (based on the response data) to achieve (updated parameters). Single-objective optimization with equal

a balance between variability and bias of the regressed surface. R2 weight of each natural frequency is used in this paper for

(ranged from 0.0 to 1.0) criterion and empirical integrated squared optimization computation.

Error (EISE) criterion can be used in response surface verification:

N 3. Numerical verification

[yRS (j) − y(j)]2

P

i=1 A simulated simply supported beam without damage and with

R2 = 1 − (3)

N an assumed damage element is considered to demonstrate the

[y(j) − ȳ]2

P

procedure of response surface -based FE model updating. The beam

j =1

of 6 m length is equally divided into 15 two-dimensional beam

1

N

X elements as shown in Fig. 2. The density and elastic modulus of

EISE = ∗ [yRS (j) − y(j)]2 (4) the beam are 2500 kg/m3 and 3.2 × 1010 Pa, respectively, while

N j =1 the area and moment of inertia of cross-section are 0.05 m2 and

where yRS is the response value of the confirmation samples; y is 1.66 × 10−4 m4 , respectively.

the true value of the confirmation samples; ȳ is the mean of all true FE modal analysis is first carried out on the beam without

values. The larger the value of R2 , the more accurate the regressed damage to get the analytical natural frequencies. To simulate the

response surface. On the contrary, the smaller the value of EISE, the actual (target) beam, one damage location is assumed at beam

closer the fit is to the data. element 10 where the element bending stiffness is reduced by 50%.

Another criterion is the root mean squared error (RMSE). It is The FE modal analysis is again carried out on this damaged beam to

the square root of the mean square error (MSE) where the distance get the simulated measured modal parameters. The initial values

vertically from the point is taken to the corresponding point on of the first 10 natural frequencies selected and corresponding

the curve fit (the error). The squaring is done so negative values differences are shown in Table 1. The maximum and minimum

do not cancel positive values. The RMSE is thus the distance, on errors that appeared in the natural frequency are 15.63% and 1.89%

average, of a data point from the fitted line, measured along a respectively.

vertical line. The smaller the value of root mean squared error, with Updating of the FE model of the undamaged beam is to achieve

0 corresponding to the ideal, the closer the fit is to the data. In the a goal to correlate the natural frequencies with the damaged beam.

real application, the R2 criterion and EISE or the RMSE criterion In this study, moment of inertia I10 and elastic modulus E10 of

can be used in a complementary way to check the accuracy of the individual element 10 are chosen as the updating parameters that

regressed response surface. affect the independent variable, natural frequencies. The range

selected for each parameter should reflect the change that one

expects to observe for the domain of the prediction of interest with

2.3. Model updating

1.920 ≤ E10 ≤ 3.465(×1010 Pa) and 0.863 ≤ I10 ≤ 1.797

(×10−4 m4 ).

In the finite element model updating in structural dynamics, the

The central composite design (CCD) method in design of

structural response features of interest are often eigen solutions

experiments is then used to get the sample points of the parameter

related to such as natural frequencies and mode shapes. In this

values selected. A total of 9 runs of experiments are carried out. The

study, the structural natural frequency is employed as a response

sampled parameter values and corresponding natural frequencies

feature. Therefore, the optimized objective function is formulated

calculated from FE model are listed in Table 2. By using least-square

in terms of the residuals between analytical and measured natural

fitting with these sample values, a quadratic polynomial response

frequencies

surface (Eq. (2)) can be regressed where parameter x1 refers to

m

X E10 while parameter x2 refers to I10 . Fig. 3 illustrates the regressed

Π (x) = wi (λai − λei ) 0 ≤ wi ≤ 1 (5) response surfaces of the first and second natural frequencies. The

i =1 horizontal axes are parameters selected while the vertical axis

where λai and λei are the analytical (finite element calculated) gives the response (natural frequency) at any point or location.

and measured natural frequencies of the ith mode respectively. To evaluate the above-selected parameters of the FE model, a

wi is the weight factor to impose to the different order of natural statistics-based parameter effects analysis is performed. Parts of

frequencies. m is the number of modes involved and x is the design the general regression significance of selected parameters (x1 and

set. x2 ) calculated from Eq. (1) on each mode of natural frequencies

The FE model updating can then be posed as a constrained comparing with the significance level of p = 0.05 are shown

minimization problem to find the satisfied design set such that in Fig. 4. The results show that both elastic modulus E10 and

moment of inertia I10 are high significance on the structural natural

Min kΠ (x)k22 (6) frequencies. It is observed that the cross-quadratic term x1 x2 in

W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465 2459

Table 2

Central composite design and sample values.

Run E10 (1010 Pa) I10 (10−4 m4 ) Freq 1 (Hz) Freq 2 (Hz) Freq 3 (Hz) Freq 4 (Hz) Freq 5 (Hz)

2 1.920 1.660 8.67 35.12 80.23 137.86 147.19

3 3.200 1.000 8.68 35.13 80.24 137.91 149.14

4 3.200 1.660 8.99 35.92 80.63 142.93 149.14

5 1.655 1.330 8.36 34.39 79.83 133.72 146.41

6 3.465 1.330 8.92 35.72 80.54 141.62 149.37

7 2.560 0.863 8.36 34.40 79.85 133.77 148.40

8 2.560 1.797 8.91 35.72 80.53 141.60 148.40

9 2.560 1.330 8.72 35.24 80.29 138.57 148.40

9.5

8.5

Frequency

38

8

36

7.5

Frequency

34

7

6.5 32

3.5

3 30

1.5

2

2.5

x1

2

1.5

2 0 x1 2.5

×1

1

0.5 ×10 1 –4

0

×10

10

1 0 3

1.5 1.5 0.5

2 x2 ×10–4 3.5 0

x2

(a) Surface of first natural frequency. (b) Surface of second natural frequency.

Table 3 Now it is time to carry out the FE model updating where the

RMSE values for each mode. FE model is replaced by the regressed quadratic polynomial. The

Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 residuals between analytical (undamaged beam) and measured

RMSE 0.0028 0.0002 0 0.0090 0.0028

(damaged beam) natural frequencies are used as the optimized

objective function. Single-objective optimization algorithm with

equal weight of each natural frequency is implemented to

achieve the best minimization of natural frequency residuals. The

the surface expression is less significant on the structural natural optimization algorithm used to minimize the objective function is

frequencies. For the simplification of a quadratic polynomial a standard Trust Region Newton method in MATLAB. The tuning

response surface, the cross-quadratic term can be omitted in minimization process is over when the tolerances are achieved or

practice to increase the efficiency of structural FE model updating. pre-defined number of iterations is reached. The updated natural

To check the accuracy of the regressed surface, the RMSE is frequencies and differences of the simulated simply supported

calculated for each mode as shown in Table 3. It is demonstrated beam are also summarized in Table 1. It can be observed that a

that all RMSE values are close to zero, which indicates that the good agreement of natural frequencies has been achieved after

created response surface has a high regression accuracy. carrying out the response surface-based FE model updating. The

2460 W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465

(a) Parameter significance on the first natural frequency. (b) Parameter significance on the second natural frequency.

Fig. 5. Convergence of objective function of response surface-based updating. Fig. 6. Convergence of objective function of sensitivity-based updating.

(true value E10 = 1.60 × 1010 Pa) and I10 = 0.85 × 10−4 m4 (true

value I10 = 0.83 × 10−4 m4 ).

Fig. 5 illustrates the convergence of the objective function

during each optimization iteration. The horizontal axis is the

number of iteration while the vertical axis is the square sum of

natural frequency residuals of each mode. To demonstrate the

efficiency of current response surface-based FE model updating,

the objective function convergence of the traditional sensitivity-

based FE model updating is shown in Fig. 6. For such a numerical

simple beam, it is shown that the convergence of the objective

function is fast and iteration number is dramatically reduced to

reach the same residual level of the objective function by using the

Fig. 7. Potion of Hongtang bridge.

response surface method.

4. A precast concrete bridge tested in the field types of spans: simply supported spans, precast truss supported

spans and precast continuous girder spans. The photograph of

A real case study is carried out on the Hongtang Bridge, located the part of bridge at present condition is shown in Fig. 7. For

in Fuzhou city, the capital of Fujian province, China. The bridge is a the purpose of dynamics-based condition assessment, one portion

multi-span continues-deck precast concrete highway bridge. The of six continuous girder spans with a span length of 40 m were

construction was completed in December 1990. The total length tested on the site under operational vibration conditions. On-site

of the bridge is 1843 m, with a span layout of (16 + 27 + 4 ∗ dynamic testing of a structure provides an accurate and reliable

30 + 60 + 120 + 60 + 31 ∗ 40 + 8 ∗ 25) m. It includes three description of its current dynamic characteristics.

W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465 2461

(a) Plan and longitudinal elevation of bridge showing sensor placement for the ambient vibration test.

(b) Details near the connection of the abutment and piers with deck (c) Details of the bearings used in support.

showing the bearing.

Fig. 8. Details of the Hongtang bridge and bearings with measurement points for the ambient vibration test.

The equipments used for the operational vibration measure- concrete is considered to be homogeneous with an initial value for

ments include accelerometers, signal cables, and a 32-channel Young’s modulus of 3.50 × 104 MPa and density of 2500 kg/m3

data acquisition system with signal amplifier and conditioner. The (C50 grade of concrete).

force-balance (891-IV type) accelerometers and INV306 data ac- It is well known that the modeling of bridge boundary condi-

quisition system were used. Accelerometers convert the ambient tions will heavily influence the calculated dynamic properties. The

vibration responses into electrical signals. Cables are used to trans- deck of tested bridge is connected to the supporting piers and abut-

mit these signals from sensors to the signal conditioner. A signal ments by neoprene bearings that allow both vertical and rotational

conditioner unit is used to improve the quality of the signals by displacements of the bridge deck as shown in Fig. 8(b) and (c). To

removing undesired frequency contents (filtering) and amplifying simulate the behavior of the translational and rotational springs,

the signals. The amplified and filtered analog signals are converted each neoprene support is modeled by means of additional con-

nected beam elements. These elements consisted of small length

to digital data using an analog to digital (A/D) converter. The sig-

(0.01 m) beams that connect the deck to abutments or piers. The

nals converted to digital form are stored on the hard disk of the

roller supports can be simulated in the numerical model by choos-

data acquisition laptop computer.

ing a cross-section of the connected beam element with a large sec-

Forty-nine measurement points were chosen at one side of the

tion area and small inertia moment, whereas the rigid supports can

bridge as shown in Fig. 8 (a). The accelerometers were directly

be simulated in the numerical model by choosing a cross-section of

installed on the surface of the bridge deck in the vertical direction. the connected beam element with a large value of both section area

Five test setups were conceived to cover the planned testing and inertia moment. In this work, the translational and rotational

locations of the bridge. One reference location was selected near behavior of the neoprene supports are simulated by choosing the

one side of abutment. The sampling frequency on site was 300 Hz. suitable values of the section areas and inertia moments of these

An operational modal identification procedure will need to base connected elements. Each neoprene bearing as shown in Fig. 8(c)

itself on output-only data. In this study, the natural frequencies has seven layers reinforced with steel plates between them. The

of the bridge are obtained by picking the peaks on the graphs of equivalent Young’s modulus of a single-layered composite element

the average normalized power spectral densities (ANPSDs) from all can be calculated with the following formula:

acceleration measurements. More details about the field ambient Ecomp = 0.1(530S − 418) (MPa) (7)

vibration measurements and modal parameter identification can

be found in Jaishi and Ren [8]. in which

The initial finite element model has been developed according A

S= (8)

to the blue print of the bridge aimed at simulating the dynamic 2t (b + c )

behavior in the vertical direction. The deck of the considered where S is the shape coefficient of the neoprene bearing; b, c , t are

spans has the form of the hollow-core precast concrete girder with the width, length and thickness of one layer of neoprene bearing

a overall width of 11 m. The deck and piles of the bridge are respectively; and A is the area of bearing. The vertical spring

modeled by two-dimensional beam elements without the shear stiffness of the multi-layered composite bearing is then equal to

deformation consideration. The equivalent values for the cross-

sectional area and inertia moment of the girder and piers are pre- Ecomp A Ecomp bc

Kv = P = P N/m. (9)

calculated and given as the inputs for the beam elements. The t t

2462 W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465

Table 4

Natural frequency differences of six span bridge before and after updating.

Mode Initial FE model (Hz) Field test (Hz) Error (%) Updated FE model (Hz) Error (%)

2 3.352 3.291 −1.86 3.188 3.13

3 3.957 3.542 −11.72 3.730 −5.31

4 4.780 4.149 −15.21 4.431 −6.80

5 5.650 4.611 −22.52 5.173 −12.18

Table 5

Parameters and their bounds.

Parameter Min. Mid. Max.

x2 (10−3 m2 ) 1.050 2.099 3.149

x3 (10−3 m2 ) 3.240 6.480 9.720

Table 6

R2 values with cross-quadratic terms.

Mode 1 Mode 2 Mode 3 Mode 4 Mode 5

on the selection of updating parameters. In practice, the discrep-

ancies between the experimental and analytical models are due

to many parameters and uncertainties. The updating parameter

selection should be made with the aim of considering the uncer-

tainties in the model. Physical and geometric properties of finite

elements as well as the stiffness of the connection are normally

chosen as the updating parameters. In this study, three important

parameters are selected for updating on the basis of the created fi-

nite element model. These parameters are elastic modulus of con-

crete material of bridge deck standing for x1 , cross-sectional area

Fig. 9. Diagram to calculate the equivalent rotational stiffness of support. of connection element at abutment bearings standing for x2 and

cross-sectional area of connection element at interior pier bear-

One bridge support consists of several neoprene bearings ings standing for x3 . The initial values of selected parameters and

arranged as shown in Fig. 8(b). The equivalent spring stiffness in their low and upper bounds are listed in Table 5 where the support

the vertical direction at one support is given by the summation stiffness of bearings is represented by varying the cross-sectional

of all individual bearing springs stiffness obtained from Eq. (9) area of corresponding connection elements. Although it is very dif-

at that support. The equivalent rotational spring stiffness of the ficult to accurately estimate the variation bound of the parameters

support around the transverse direction of the bridge is derived during updating, a quite large of variation to the parameters of con-

with reference to Fig. 9 using the value of vertical stiffness. Due nection elements (bearings) is expected due to more possible un-

to the application of bending moment M1 at support, both vertical certainties involved in the bridge bearings.

springs with stiffness K1 and K2 , which are separated L distances Once the model structural parameters and their bounds are

apart, are compressed with deflection ∆1 and ∆2 respectively. selected, the sampled parameter points can be determined by

Using simple statics, it can be obtained that design of experiments (DOE). Again, the central composite design

(CCD) method is used to obtain the sample points of the parameter

values. With these sampled parameter values, corresponding

M1 1 1

θ= + . (10) structural natural frequencies (responses) are calculated from

L2 K1 K2

FE modal analysis. By using least-square fitting, a quadratic

In the case of K1 = K2 = Kv , the equivalent rotational stiffness polynomial response surface (Eq. (2)) can be constructed with

of the support can be written as regard to parameters x1 , x2 and x3 .

For the construction of a quadratic polynomial response surface

M1 Kv L 2 in FE model updating in structural dynamics, as discussed in the

Krot = = N m/rad. (11) case study of a simulated simply supported beam, the cross-

θ 2

quadratic term (interaction effect) in the surface expression is

The equivalent section area and second moment of area of less significant on the structural natural frequencies. To check

the connected beam elements calculated from these vertical and the correlation of the regressed surfaces, the R2 criterion is

rotational stiffness are used as the initial values in the FE model used. For the first five natural frequencies, the R2 values of the

updating. Subsequently, the established FE model of tested six regressed quadratic polynomial surfaces with and without the

continuous spans of the bridge, as shown in Fig. 10, is composed cross-quadratic terms are listed in Tables 6 and 7 respectively.

of 87 nodes, 159 elements (79 beam elements, 73 mass elements, It can be seen that the correlation accuracies of both quadratic

7 connected beam elements). The natural frequencies calculated polynomial surfaces with and without the cross-quadratic terms

from the initial FE model and identified from operational vibration are almost the same. Therefore, the cross-quadratic terms in the

measurement are shown in Table 4. quadratic polynomial can be omitted to increase the efficiency

W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465 2463

Fig. 11. Created response surfaces for the first natural frequency.

Table 7

R2 values without cross-quadratic terms.

Mode 1 Mode 2 Mode 3 Mode 4 Mode 5

Table 8

Surface regression coefficients for first natural frequency.

β0 β1 β2 β3 β11 β22 β33

0.698 0.742 0.041 0.061 −0.041 −0.006 −0.002

of the created quadratic polynomial surface without the cross-

quadratic terms are shown in Table 8. Fig. 11 illustrates the created

response surfaces of the first natural frequency. The horizontal

axes are the parameters (x1 , x2 and x3 ) selected, while the vertical

axis gives a value of the first natural frequency (response) at any

Fig. 12. Parameter significance on the first natural frequency.

point or location when one parameter is changed and other two

parameters are fixed.

To demonstrate the significance of three selected parameters FE calculated and experimental natural frequencies have been

on the structural responses (natural frequencies), the test on improved after performing the response surface-based FE model

variances (F-test) is carried out to check the significance of selected updating. However, the error of the second mode natural frequency

parameters. The regression significance of selected parameters increases after the model updating process. That implies that it

calculated from Eq. (1) on the first mode of natural frequency is not always guaranteed that the updated finite element model

compared with the significant level of p = 0.05 is shown in Fig. 12. gives an improved prediction for every order of frequencies as the

It can be observed that the parameter x1 , the elastic modulus natural frequencies reflect the overall structural stiffness.

of concrete material of bridge deck, has the most significant After updating, the changes in the selected three updating

effect on the structural natural frequencies among three selected parameters of the bridge are shown in Table 9. It can be seen that

parameters. the initial value for Young’s modulus of the concrete is quite good

Subsequently, an optimization procedure for model updating is such that a small correction is needed, but a significant correction

carried out on the created quadratic polynomial response surface. on the initial estimation of the stiffness of the neoprene supports

The first five natural frequency residuals between analytical turns out as the stiffness of neoprene supports depends much on

and measured natural frequencies are used as the optimized the hardness of the rubber that is difficult to determine exactly.

objective function. Single-objective optimization algorithm with Therefore, the simulation of support conditions in the FE model

equal weight of each natural frequency is implemented to achieve updating of a real bridge is very important, and replacing the

the best minimization of the natural frequency residuals. The support simply with rollers cannot actually simulate the dynamic

optimization algorithm used to minimize the objective function behavior of the bridge. The neoprene supports in general provide a

is a standard Trust Region Newton method in MATLAB. The final significant value of the rotational stiffness.

correlation of natural frequencies after the FE model updating is Fig. 13 illustrates the convergence of the objective function

shown in Table 4. This shows that the overall correlations between during each optimization iteration. The horizontal axis is the

2464 W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465

Selected parameters of Hongtang bridge before and after updating.

Selected parameters Initial values Updated values Change (%) The paper presents an application of the response surface

Elastic modulus of deck 3.500 3.118 −10.91 method for finite element model updating of civil engineering

(1010 Pa) structures in structural dynamics. A quadratic polynomial re-

Connection element area 2.099 2.686 27.96

sponse surface is constructed using the finite element model sim-

at side bearings (10−3 m2 )

Connection element area 6.480 12.909 99.10 ulation data and then employed to improve the computation

at interior bearings efficiency in the model updating. The proposed procedure is il-

(10−3 m2 ) lustrated on a simulated simply supported beam and a full-size

precast continuous box girder bridge tested under operational vi-

bration conditions. With the response surface to replace the origi-

nal FE model, the model updating process becomes efficient and

converges fast compared with the traditional sensitivity-based

model updating method. Once the response surface is constructed,

there is no finite element calculation involved in each optimization

iteration. As a result, the FE model updating can be easily imple-

mented in practice with available commercial finite element anal-

ysis packages.

It is demonstrated that updating parameter selection from all

possible physical or geometrical parameters in the FE model and

a good approximation of response surface construction are the

keys to the success of the method. A small number of structural

parameters chosen in the examples and an estimation of the

response surfaces by a quadratic function may not be far from the

truth for more general problems. It is still a challenge to test the

response surface method for the finite element model updating

of more complex civil engineering structures where the relation

between the unknown parameters and the response quantities of

Fig. 13. Convergence of objective function of response surface-based updating. interest is more complex and a much large number of parameters

of the finite element model are involved.

Acknowledgements

2009AA11Z101 and Natural Science Foundation of China (NSFC)

under grant number 50678193 are greatly acknowledged.

References

[1] Friswell MI, Mottershead JE. Finite element model updating in structural

dynamics. The Netherlands: Kluwer Academic Publishers; 1995.

[2] Baruch M, Bar-Itzhack IY. Optimal weighted orthogonalization of measured

modes. AIAA J 1978;16(4):346–51.

[3] Berman A, Nagy EJ. Improvement of large analytical model using test data.

AIAA J 1983;21(7):1168–73.

[4] Link M. Updating of analytical models-Review of numerical procedures and

application aspects. In: Structural Dynamics forum SD2000. Los Alamos;

1999.

Fig. 14. Convergence of objective function of sensitivity-based updating. [5] Fritzen CP, Jannewein D, Kiefer T. Damage detection based on model updating

methods. Mech Syst Signal Process 1998;12:163–86.

[6] Jaishi B, Ren WX. Structural finite element model updating using ambient

number of iteration, while the vertical axis is the square sum vibration test results. J Struct Eng, ASCE 2005;131(4):617–28.

of relative natural frequency residuals of each mode. For the [7] Jaishi B, Ren WX. Damage detection by finite element model updating using

comparison as the first example, the convergence of the objective modal flexibility residual. J Sound Vib 2006;290(1–2):369–87.

[8] Jaishi B, Ren WX. Finite element model updating based on eigenvalues and

function using the traditional sensitivity-based finite element strain energy residuals using multiobjecive optimization technique. Mech Syst

model is shown in Fig. 14. It is clearly shown that the convergence Signal Process 2007;21(5):2295–319.

error of the objective function after the same 30 iterations is less [9] Lewa TL, Spencera AB, Scarpaa F, Wordena K, Rutherfordb A, Hemezb F.

Identification of response surface models using genetic programming. Mech

by using the current response surface-based finite element model

Syst Signal Process 2006;20:1819–31.

updating method. [10] Khuri AI, Cornell JA. Response surface designs and analysis. Marcel Dekker, Inc;

It should be noted that the real bridge problem has serious 1987.

scouring and erosion at the submerged part of the structure while [11] Myers RH, Montgomery DC. Response surface methodology: process and

product in optimization using designed experiments. New York: John Wiley

the parameters selected are those at the bearings. An equivalent

& Sons, Inc; 1995.

damage at the supports of the bridge deck is assumed with this [12] Montgomery DC. Design and analysis of experiments. 3rd ed. New York: John

assumption and this would lead to non-realistic results though Wiley & Sons, Inc; 2006.

the final natural frequencies may have good match with the [13] Romero VJ, Swiler LP, Giunta AA. Construction of response surface based

on progressive-lattice-sampling experimental designs with application to

measured ones. Hidden damages cannot be noticed by engineers uncertainty propagation. Struct Saf 2004;26:201–19.

and they may be unaccounted for the application of the proposed [14] Batmaz I, Tunali S. Small response surface designs for metamodel estimation.

approach. European J Oper Res 2003;145:455–70.

W.-X. Ren, H.-B. Chen / Engineering Structures 32 (2010) 2455–2465 2465

[15] Simpson TW, Williams BJ, Notz WI. Sampling strategies for computer [19] Hemez FM, Doebling SW, Anderson MC. A brief tutorial on verification

experiments: design and analysis. Int J Reliab Appl 2002;2(3):209–40. and validation. In: Proceedings of the 22nd international modal analysis

[16] Rutherford BM, Swiler LP, Paez TL, Urbina A. Response surface (meat-model) conference. 2004. Paper No. 250.

methods and applications. In: Proceedings of the 24th international modal [20] Steenackers G, Guillaume P. Finite element model updating taking into

analysis conference. 2006. Paper No. 184. account the uncertainty on the modal parameters estimates. J Sound Vib 2006;

[17] Doebling SW, Hemez FM, Schultze JF, Cundy AL. A metamodel-based approach 296:919–34.

to model validation for nonlinear finite element simulations. In: Proceedings [21] Faravelli L, Casciati S. Structural damage detection and localization by

of the 20th international modal analysis conference. 2002. p. 671–8. response change diagnosis. Struct Saf Reliab 2004;6:104–15.

[18] Guo QT, Zhang LM. Finite element model updating based on response surface [22] Rutherford AC, Inman DJ, Park G, Hemez FM. Use of response surface

methodology. In: Proceedings of the 22nd international modal analysis metamodels for identification of stiffness and damping coefficients in a simple

conference. 2004. Paper No. 93. dynamic system. Shock Vib 2005;12:317–31.

- MIT2_092F09_lec06.pdfUploaded byDaniel
- 1-s2.0-S0022460X15010512-main.pdfUploaded bypalash22
- 45909725-Mechanica-TutorialUploaded byrithehbk
- Chapter 2_Finite Element ProcedureUploaded byIvan Klyuchka
- Content AE675Uploaded byBrajesh kumar
- An Optimization of Rubber Bushing Material and Structural ParametersUploaded byRajaSekarsajja
- How we might be able to understand the brain 2004Uploaded byLilya Butorina
- Finite Element MeshingUploaded byNalaka Santhajeewa
- Mitres2 002s10 TocUploaded byDaniel
- AcuSolve_Radioss_FSIUploaded byRaghu Samraat
- Nisa Pro ShapeUploaded byseethaviji
- TSAT_Model_Manual.pdfUploaded bySuterm Seccion
- A comparison of neural network, statistical mathods, and variable choice for live insurrers' financial distress predictionUploaded bykaixun90
- Evaluation-Delft-1.pdfUploaded byrezapou
- 4. DOE Language and ConceptsUploaded byguillermo2000
- ISPRS Cardaci Mirabella VersaciUploaded byAntonella Versaci
- Prilog Na Spektrite Od RobotUploaded byevgenija89
- Introduction 1Uploaded bySeng Heang
- IRJET- Analysis of RCC Culvert by using SoftwareUploaded byIRJET Journal
- DT091_Approvedmodules08[1]Uploaded byDeepankumar Athiyannan
- 2D_FEUploaded bysqaiba_g
- RM Project (Complete)Uploaded byPrakash Naik
- 1stUploaded bybasa rk
- 20120006035Uploaded byhetrodit
- creo01a-160515062445Uploaded byBoro Grozdanić
- 2007 - Modal Acoustic Transfer Vector Approach in a FEM BEM Vibroacoustic AnalysisUploaded byGuilherme Henrique Godoi
- High order hybrid finite element schemes for Maxwell’s equations taking thin structures and global quantities into accountUploaded byJoseph Smith
- Full Text 01Uploaded byMario Barbato
- Creation of the Finite Element Model of Complicated Structure Cylindrical Involute Internal Spur GearUploaded bygapinganus
- Biological Neural Systems-PresentationUploaded byapi-26548538

- Introduction to Finite Elements in Engineering, 3rd Ed, t.r.chandrupatla_2Uploaded byRohitash Nitharwal
- Yuan_et_al-1998-Earthquake_Engineering_%26_Structural_Dynamics.pdfUploaded byAnjneya Srivastava
- Hogs and PitsUploaded byAnjneya Srivastava
- 200 Best Ques Nov With AnswersUploaded byAnjneya Srivastava
- 1-s2.0-S0888327016304964-mainUploaded byAnjneya Srivastava
- new paperUploaded byAnjneya Srivastava
- Damage detection using response surface methodology in shear building.pdfUploaded byAnjneya Srivastava
- Damage detection using response surface methodology in shear building.pdfUploaded byAnjneya Srivastava
- Shokrani_et_al-2018-Structural_Control_and_Health_Monitoring.pdfUploaded byAnjneya Srivastava
- Shokrani_et_al-2018-Structural_Control_and_Health_Monitoring.pdfUploaded byAnjneya Srivastava

- NPWS 2009 AA GuidanceUploaded byliamo8888erere0
- ASD Social EmotionalUploaded byCiera Corca
- Simple Linear RegressionUploaded byHicham Tou Nsi
- الإندماج الأوروبي ونظريات التكامل.pdfUploaded byFA R Ha
- Ewing's Analytical Instrumentation Handbook 2005Uploaded byNirmal Kumar Pandey
- Qualitative SamplingUploaded byAndi Ulfa Tenri Pada
- Jeffers PEAR Proposition Critique PostUploaded byBryan J Williams
- ScWk 240 Week 5 2nd Set Slides Internal and External ValidityUploaded byRam Mohanreddy
- Maori wellnessUploaded byfiserada
- Unit-1 Comparative Politics- Nature, Significance and EvolutionUploaded bySuhailAhmed
- TopographyUploaded byIuliana Si Mihai Niculita
- Detection LimitsUploaded byMohd Alif
- Strategic management of information systems in healthcareUploaded bymrtnfs
- Manual Sensor SickUploaded byOliver Paramo
- Timefor Final Results 6Uploaded byJohanna Alumbro
- UV_VisUploaded bylenthobakar
- Lec_3Uploaded byALIKNF
- Homework+_1+Additional+Questions+SolutionsUploaded byclairejaeger
- Biology IntroductionUploaded byWelkom Biologyproject
- PROBABILITY BASED CLUSTER EXPANSION OVERSAMPLING TECHNIQUE FOR IMBALANCED DATAUploaded byCS & IT
- LIkert analysisUploaded bymilind_bs_2000
- Research MethodsUploaded byJanti Undari
- ch2-3-4510Uploaded byLeia Seungho
- Scope of SociologyUploaded bySwati Agarwal
- Bio StatisticsUploaded bygaoyutailang
- Chapter 12Uploaded byDavizas
- Lecture 01Uploaded byHéctor Flores
- System Modelling and SimulationUploaded byArvind Kumar
- Why Time Travel is PossibleUploaded byLaron Clark
- Imagination: a Legacy of Romantic LiteratureUploaded byJudyth Vary Baker