Documente Academic
Documente Profesional
Documente Cultură
www.elsevier.com/locate/engstruct
Abstract
This paper describes the static, dynamic and long-term tests of bridges in situ, which have been performed in the Czech and
Slovak Republics since 1968. The standard methods are supplemented with the criteria for the elastic and permanent deformations,
natural frequencies and the dynamic impact factors. The monitoring of stresses under usual traffic loads provides important data
for the fatigue of bridges, for the estimation of their residual life and for the determination of inspection intervals. Modal analysis
and identification ascertain the characteristic properties of bridges from their response. The damage in bridges may be reflected in
the changes of their natural frequencies or modes of natural vibration. 2000 Elsevier Science Ltd and Civil-Comp Ltd. All
rights reserved.
1. Introduction to the first three versions of the standard [1], they were
obligatory for new bridges with spans greater than 18 m.
Load tests of bridges in situ are an important pro- Before testing, a programme is put together where the
cedure for checking the quality of structures. They make testing institute suggests the load (in accordance with
it possible to compare the theoretical assumptions with the designer), its distribution, the measured points, the
the actual behaviour of the bridge subjected to the test experimental method and other necessary data.
load. In this way they contribute to progress in civil Just before the tests, the bridge is inspected and sur-
engineering theory and practice. veyed by geodetic methods.
Being aware of the importance of bridges as structures During a static test it is obligatory to measure the ver-
serving public transport and being willing to contribute tical deflections at the points where the greatest effects
to better knowledge of their behaviour under load we are expected (in the middle of spans, in the quarter of
have standardized the load tests in situ in the Czech and arches etc.), the settlements of supports and the squeeze
Slovak Republics [1]. The first edition of the standard of bearings. It is recommended that measurements are
[1] was introduced in the year 1968 and revised in 1979, made of the temperature of the ambient air and of the
1990 and 1996. Several hundreds of both highway and structure, its strains and stresses, the deformations of
railway bridges of larger spans have been tested in the other parts of the structure, the development of cracks
past 30 years according to the standard [1]. The purpose and the stability of compressed elements of the bridge.
of the present paper is to describe the static, dynamic The basic data must be recorded, including: date of
and long-term tests of bridges, to submit our experience the test, weather conditions, experimental devices and
for discussion and to present the recent achievements in their sensitivity, mass of the loads and their dimensions
this field. and the test sequence.
The heavy vehicles usually applied for loading
2. Static tests include: locomotives, wagons, rail cranes, etc. for rail-
way bridges, and trucks, track vehicles, building
Static tests are carried out if demanded by government machines, water cisterns, etc. for highway bridges. The
authorities or by the designer of the bridge. According intention is to load the bridge by the heaviest load poss-
ible. Its efficiency is described by the equation
* Corresponding author. Tel.: +42-0246-1634; fax: +42-288-4634.
E-mail address: fryba@itam.cas.cz (L. Fryba). UNkUV (1)
0141-0296/01/$ - see front matter 2000 Elsevier Science Ltd and Civil-Comp Ltd. All rights reserved.
PII: S 0 1 4 1 - 0 2 9 6 ( 0 0 ) 0 0 0 2 6 - 2
L. Fryba, M. Pirner / Engineering Structures 23 (2001) 102109 103
where UN is the effect of the test load, UV is the effect of the width of cracks is limited for concrete bridges,
the standard load including the standard dynamic impact see Table 2.
factor, and k is the efficiency factor. For a static test it
should hold that Condition (5) may be modified for new bridges: if the
effects of the first loading comply with the condition
0.5kstat1.0. (2)
Sr
The static loads are so arranged on the bridge as to cause a1 a (6)
the greatest effects at measured points. Stot 2
The dead and long-term loads should be allowed to then the loading can be repeated, but the condition of
act on the bridge before the test, at least 72 h on concrete permanent deformation is now
bridges and 3 h on steel bridges. Concrete bridges may
be tested at least 3 months after the casting of their Sr
a . (7)
main parts. Stot 3
The minimum loading time is 30 min on concrete
bridges and 15 min on steel bridges. The measured data If condition (7) is not fulfilled, the loading may be
are recorded at least twice before loading, immediately repeated once more. Then, of course:
after loading, then at 10-min intervals (maximum), just
after unloading and at the same interval up to stabiliz- S r a1
. (8)
ation. The loading is usually repeated twice. Stot 6
The time rules mentioned above have been developed
from our experience because only a limited time period
In case the measured values do not correspond with
has been at our disposal for the load test. The rules are
the criteria (4) to (8), a special investigation and/or long-
a compromise between the scientific requirements and
term observation and/or dynamic tests are necessary. In
the practical possibilities.
any case, the decisive value provides the vertical down-
After the static test the total Stot, permanent Sr and
ward deflection at the midspan of the bridge or at the
elastic components Se of all measured values are calcu-
quarter-span of an arch.
lated, where the following relation is valid
StotSrSe. (3)
3. Dynamic tests
The first loading gives the decisive values for the evalu-
ation.
Dynamic tests are usually carried out on bridges of
The tested bridge satisfies the standard [1] if the fol-
large spans, unusual structural systems or new materials.
lowing three conditions are fulfilled (for the coefficients
The first three versions of the standard [1] prescribed the
a, ai, b see Table 1):
dynamic tests for bridges with spans greater than 50 m.
Before testing the levelling of rails or the road pave-
the condition of elastic deformation (Scal is the calcu-
ment in front of, on and beyond the bridge is necessary.
lated value)
The dynamic response at important points of the struc-
Se ture, speed of vehicles or input characteristics of other
b a (4) loads are recorded during the test. Further, it is rec-
Scal
ommended that measurements are made of the strain
or stresstime history at other points, the horizontal
the condition of permanent deformation
Sr Table 2
a (5) Width of cracksa
Stot 1
Bridge Environment class Crack width (mm)
Reinforced 1 0.4
2, 3 0.3
Table 1 4, 5 max 0.1
Coefficients a, ai, b Partially prestressed 1 0.2
2, 3 0.1 (6)
Bridge a a1 a2 a3 b 0 (7)
4, 5 0
Prestressed 1.05 0.2 0.5 0.1 0.7 Full prestressed any 0
Reinforced 1.10 0.25 0.5 0.125 0.6
a
Steel 1.05 0.1 0.3 0.05 0.8 1 dry, 2, 3 humid, 4, 5 aggressive, (6) post-stressing,
(7) prestressing.
104 L. Fryba, M. Pirner / Engineering Structures 23 (2001) 102109
viv(j)vid(j) 2
COMAC(j)
i1
(18)
N N
application to big engineering structures has not been where viv(j) is the displacement of the ith vibration mode
developed until recently. This method is able to recog- of the first (virgin) state of the structure at point j, vid(j) is
the displacement of the ith vibration mode of the second
(damaged) state of the structure at point j, and N is the
number of excited vibration modes.
If the COMAC value is equal to one, then no differ-
ence appears between the deflection coordinates in the
first (virgin) state and the second (damaged) state. The
lower COMAC values signify the differences in coordi-
nates and, thus, possible damage. The COMAC method
has been applied to the identification of imperfections,
damage or cracks at the Institute of Theoretical and
Applied Mechanics of the Academy of Sciences of the
Czech Republic in Prague.
The COMAC method was used for checking the qual-
ity of a three-span prestressed concrete segment bridge
across a motorway, see Fig. 6. The length of the bridge
is 64.4 m.
The left superstructure crashed by sliding spon-
taneously off its bearings. This movement caused defor-
mation and opening of most joints between segments.
After the structure had been lifted back on its bearings
and the cracks sealed, both the damaged (left) and
undamaged (right) superstructures were subjected to
static and dynamic loading tests. The purpose of the tests
was to verify whether the repaired left superstructure
Fig. 5. Amplitudes of vibration of columns of the bridge shown in was comparable with the unimpaired right superstruc-
Fig. 4 due to the wind at speeds v. ture.
L. Fryba, M. Pirner / Engineering Structures 23 (2001) 102109 107
Fig. 6. The cross-section of one of the bridges and a plan view of both structures together with movements of the bearings.
Fig. 7. The first four excited modes of the left- and right-hand side structures shown in Fig. 6.
108 L. Fryba, M. Pirner / Engineering Structures 23 (2001) 102109
Table 4 much from the right one and, consequently, that the
COMAC values for the bridge at points j of the mesh 14 and m to M repair of the crashed structure was well done.
Mesh 1 2 3 4 Another example concerns a 14-year-old prestressed
concrete footbridge (Fig. 8). The principal aim of the
m 0.9997 1.0000 0.9999 1.0000 dynamic tests was to verify the sensitivity and reliability
l 0.9995 0.9999 0.9999 0.9997 of the footbridge after 14 years of traffic. At that time,
only the natural frequencies (not natural vibration
k 0.9993 1.0000 0.9999 0.9999
modes) were measured and, therefore, the present natural
j 0.9982 1.0000 0.9997 0.9994 frequencies were used to compare the two states of the
i 0.9500 0.9990 0.9998 0.9996 structure. An electrodynamic exciter at point L1 or K1
h 0.9721 0.9844 0.9969 0.9474 (Fig. 9) was applied to provide the forced (or natural)
g 0.9998 0.9998 1.0000 1.0000 modes and natural frequencies. Fig. 9 shows five natural
f 0.9994 0.9997 0.9998 0.9999
e 0.9992 0.9996 0.9997 0.9999 modes and frequencies.
d 0.9994 0.9997 0.9999 0.9999 A comparison of natural frequencies before and after
c 0.9999 0.9999 1.0000 1.0000 14 years is presented in Table 5 together with the calcu-
b 1.0000 1.0000 1.0000 1.0000 lated results. Table 5 shows that the structure does not
differ much from the virgin state 14 years ago. The small
A 0.9999 0.9999 1.0000 1.0000
differences may be explained by the different frequency
B 1.0000 1.0000 1.0000 1.0000 steps of the applied analyzers (Bruel-Kjaer in the year
C 0.9999 1.0000 1.0000 1.0000 1983 and Onosoki in the year 1997). The maximum error
D 0.9998 0.9999 0.9999 1.0000 in frequency measurements could be expected to be
E 0.9997 0.9999 1.0000 1.0000 within the range of 1% in both cases.
F 0.9995 0.9999 0.9999 1.0000
G 0.9998 1.0000 1.0000 0.9998
H 0.9653 0.9946 0.9979 0.6679
I 0.9866 0.9995 0.9999 0.9997 6. Conclusions
J 0.9980 0.9998 1.0000 0.9997
Fig. 9. The first five modes and frequencies of natural vibration of the footbridge shown in Fig. 8.
Table 5 Acknowledgements
Measured and calculated natural frequencies (Hz) of the footbridge
depicted in Fig. 8
The support of grants 103/97/0139, 103/96/K034 and
Measured in the Measured in the Calculated, Calculated, 103/98/1479 of the Grant Agency of the Czech Republic
year 1983 year 1997 deformation finite element is gratefully acknowledged.
method method