FEM Simulation and Field

Monitoring of Depropping
Procedure of a Large-Span
Single-Layer Latticed Shell
Jun Chen1,*, Yixin Peng1 and Xudong Zhao2
1Department of Building Engineering, Tongji University, China
Email: [email protected]
2Shanghai Institute of Architectural Design & Research, Shanghai, China

Email: [email protected]

(Submitted on April, 19, 2010 - Reception of revised paper November, 09, 2010 - Accepted on December, 13, 2010)

ABSTRACT: This paper investigates the depropping procedure of a large span

single-layer dome-shape latticed shell roof structure through numerical analysis
and field measurements. The depropping procedure consists of six steps. For each
step, the joint displacements and member internal forces of the roof structure were
first predicated using a numerical model. Based on the simulations results,
instruments as vibrating wire strain gauge, total station, accelerometer and global
positioning system (GPS) were installed on the structure to monitor the actual
displacements and variations of structural members internal forces during the
depropping procedure. The GPS data are compared with total station
measurements to learn its applicability for monitoring depropping procedure of
large roof structure. The comparison between the numerical results and field
measurements demonstrates that the suggested simulation method is reasonable in
the sense that the variation trend of deflection and internal forces of the structure
can be captured. The GPS technology gives more accuracy measurements in
horizontal displacement than vertical deformation. A combination of GPS
technology and total station can increase the displacement measurement accuracy.

Key Words: single-layer latticed shell, depropping procedure, numerical

simulation, GPS technology.

1. INTRODUCTION may last for several hours or even a few days. During
More and more large-span steel spatial structures have the depropping procedure, the variation of
been built recently in mainland China such as the displacements and internal forces of the roof structure
Beijing Olympic Stadium, the Tianjin Olympic should be carefully controlled in order to ensure the
Stadium and the South Railway Station of Shanghai. safety of the structure. Moreover, there might be several
The installation or assembling of the steel roof depropping plans available for a certain roof structure.
structure is a key construction step of such structures Therefore, effective and efficient numerical simulation
(Cui, 2005; Du, 2008). In current practice, it is a of the depropping procedure in advance is required to
common way to first assemble roof structure members find out the maximum joint displacement and variation
on a temporary support system/structure and then of internal forces of the structural members. On the
remove the temporary support structure, through a other hand, field measurements in the depropping
procedure called depropping, to finally allow the roof procedure are always desirable to check the
structure to support itself. The depropping procedure methodology used in the numerical simulation, to
typically comprises staged construction processes that understand the actual performance of the roof
*Corresponding author: [email protected]

International Journal of Space Structures Vol. 26 No. 1 2011 45

 FEM Simulation and Field Monitoring of Depropping Procedure of a Large-Span Single-Layer Latticed Shell

structure, and to develop better design theory and of the roof structure and to increase its out-of-plan
construction plan. The simulation and field monitoring stiffness, (see Figure 1b). Besides, there is a 14-meter-
measures for deproppig procedure are of particular long opening as an entrance at the bottom of the
importance for single-layer latticed dome, whose reticulated shell. In order to reduce the horizontal force
safety and global stability are sensitive to defects and applied to the entrance girder by the upper roof
change of geometric parameters (Lopez et al., 2007, structure, two dendritic supports are also set in lateral
Gioncu,1995). direction to partially share the lateral force in
This paper presents numerical simulations and field conjunction with the lower supporting structure at one
measurements of the depropping procedure of a side of the reticulated shell along the longitudinal
special single-layer dome-shape roof structure. First, direction (see Figure 1b).
numerical model of the roof structure has been built The single-layer reticulated shell is in the form of
where gap element is adopted to simulate the three dimensional lattices whose size is approximately
temporary supports. The stresses of structural 2.5 meters. The joint of the shell, very rare in China, is
members and joint displacements are calculated by the a welded drum-shape node. A typical joint consists of
numerical model for each depropping step. Then, six rectangular steel pipes (220 140 66 mm) that
instruments as strain gauge, total station, are intersected in a drum-shape member of the size
accelerometer and global positioning system (GPS) 320 14 mm (for joint with dendritic support the
are installed on the structure to monitor the actual drum member size is 320 20 mm). Details of a
deformation and change of internal forces. The GPS typical joint are shown in Figure 2. The surrounding
technology is not very common in monitoring of roof supports of the latticed roof structure are hemisphere
structures. Thus, the GPS data are compared with total fixed supports connecting to the concrete platform.
station measurement to learn its applicability for The connections between the connecting plates and the
monitoring depropping procedure of large roof hemispheres use groove complete penetrate weld and
structure. Comparisons between numerical results and so do connections between the hemispheres and the
field measurements are performed leading to the main members.
observations of this study.
2.2. Construction procedure of the roof
2. THE ROOF STRUCTURE AND Using full-hall scaffold as temporary support system,
DEPROPPING PLAN the roof structures is constructed segment by segment
2.1. Overview of the roof structure along the longitudinal direction. Each segment has the
The main building is a water-drop-shape stadium same span as the roof structure in the lateral direction
covering an area of 3822.4 m2 (Figure 1a). The and is further divided into 4 to 5 parts. Each part is
building is located in Tieling City, China. Its steel roof welded on the ground first and then lifted to the design
structure is a special single-layer reticulated shell, of elevation. The segment is completed by welding all the
which the longitudinal span is 82.7 m, the lateral span parts together. After that the segment is laid on the
is about 55 m and the elevation is 17.5 m at the highest temporary support system by fastening segment pipes
point. Seven dendritic supports are arranged along the on the scaffold. Once available, connection members
lateral direction in order to reduce the longitudinal span between two adjacent segments are welded. Compared

Dendritic supports
Longitudinal

Lateral Dendritic
supports

Y
x
z

(a) Picture of the roof structure under construction (b) The roof structure system and locations of dendritic supports

Figure 1. Roof structure system

46 International Journal of Space Structures Vol. 26 No. 1 2011

Jun Chen, Yixin Peng and Xudong Zhao

220 140 6 6
10 mm thick 30 20 steel pipe 220 140 6 6
220 140 6 6 ribbed stiffener Sleeve
10 thick cover plate /2
220 140 6 6
Sleeve
220 140 6 6

120
260 0

14
60
220 140 6 6

R50
220 140 6 6 160
160

60
320
10 thick cover plate
140 140 6 6 R7
5
Pin shaft 30

320 20 steel pipe

220 140 6 6

Figure 2. Details of a typical joint (all dimension in mm)

to the traditional over-head assembling method of system. After having assembled the roof structure, the
reticulated shells, the above construction method temporary supports must be removed. The whole
could reduce significantly the amount of overhead structure becomes self-load-bearing and the internal
welding work, the accumulated assembling errors, the forces of the reticulated shell will redistribute among
stresses and deformations due to welding. Within the the members. Because the single-layer reticulated
construction period, the temporary system bears all the shell is sensitive to defects, once local unstable event
loads, including the self-weight of the reticulated shell or damage occurs during the depropping procedure, it
structure and the construction live load. After the roof is very likely to cause chain effects thus leading to
structure being completed, the temporary supporting instability of the whole roof structure. Therefore, the
system will be removed through the depropping depropping procedure must be symmetrical,
procedure. synchronized and steady in order to prevent any
excessive local internal forces or joint displacements
2.3. Depropping plan that may occur.
As mentioned above, members of the roof structure is As shown in Figure 3, the seven dendritic supports
actually fastened on the scaffold. There are in total 329 in the middle of the structural plan divide the plan into
connections, termed as temporary support, between southern part and northern part. The lateral span in
the roof structure and the temporary supporting southern part is short than that of the northern part.

Figure 3. Six steps of the depropping procedure

International Journal of Space Structures Vol. 26 No. 1 2011 47

 FEM Simulation and Field Monitoring of Depropping Procedure of a Large-Span Single-Layer Latticed Shell

Therefore, the depropping process starts from the about the temporary supports. Hence, all of the
southern part and consists of the following six steps. In temporary supports are simulated by spring element
Step 1, all the temporary supports along the three solid (Gap element). The spring is in compress when the
lines in Figure 3 are released. It is accomplished by joint moves downward and the spring bears no stress
loosening the connections between the roof and the when the joint moves upward, which means the
scaffold, and removing the supporting pipe in temporary support disengages from the structure.
the scaffold. In Step 2, all the remaining temporary Stiffness of the spring element k is critical for
supports in the southern part are removed. After that, successful simulation of the temporary support. All
Steps 3 to 5 are conducted in the northern part by the steel pipes of the scaffold are standard pipe of
removing temporary supports at certain locations. 48 3.5 mm. Assuming only axial deformation of
Finally, in step 6, all the temporary supports along the the supporting pipes, the stiffness k of the temporary
perimeter of the structure are removed. The number of support can be estimated by k = EA/L, where
temporary supports removed in Step 1 to 6 is, E = 201 105 N/mm2 and A = 489.3 mm2. The height
respectively, 33, 55, 55, 55, 40 and 91. of all the 329 temporary supports varies in the range of
7.2 to 17.5 m, the computational length is determined
3. NUMERICAL SIMULATION OF by the following way. All the temporary supports are
DEPROPPING PROCEDURE divided into 5 groups according to their heights, which
3.1. Numerical model and parameters are, respectively, within the height of 7.2 m (the height
Numerical simulation of the planned depropping of temporary supports on the outward circle); from 7.2 m
procedure has been conducted using the software SAP to 9.5 m; from 9.5 m to 12.5 m; from 12.5 m to 15.5 m;
2000 to learn the status of the roof structure in each and from 15.5 m to 17.5 m. An average height L = 7.2,
step. For double-layer or multi-layer latticed shell 8.35, 11, 14 and 16.5 m are assigned for each group.
structures, three-dimensional bar element is typically The corresponding stiffness k for each group are
adopted to model the structural member under the therefore 14271, 12305, 9341, 7339 and 6227 KN/m.
assumption that it bears mainly axial forces. For the Besides, a concentrated load of 0.8 kN is applied on
single-layer latticed shell structure, however, all each joint of the numerical model to account for the
the members are rigidly welded to the joint. Bending construction loads and the weight of the drum joint.
moment cant be neglected and it is more likely to 3.2. Simulation of depropping procedure
control the members performance. Therefore, three- With all the above modeling rules and assumptions,
dimensional beam elements are adopted to simulate all numerical simulation is performed to calculate the
the steel members of the roof structure. The joint is joint displacement and internal force of key members
assumed as a rigid joint, each joint has three linear for each step of depropping procedure. Figure 4a
displacements and three angular displacements. The through 4g shows, respectively, the deformation of
hemispherical fixed supports of the structure are the roof structure before depropping, after Step 1 to
modeled as rigid supports as well. after Step 6. The deformations are amplified for
As for modeling of the temporary support system, it demonstration purpose. The calculated member
could be integrated into the main building model. internal force will be presented in later section
However, as the number of elements increases, the together with measured value.
computational costs and possibilities of modeling error
will significantly increase. In this connection, 4. FIELD MEASUREMENT OF THE
simplified model for temporary support system is DEPROPPING PROCEDURE
desirable especially for comparison purpose when Based on the numerical simulation results, various
there are several different depropping plans available. kinds of sensors were arranged to monitor the
Considering the fact that temporary supports mainly depropping procedure. Numerical simulation shows
bear vertical loads, they could be simplified as a group the vertical displacements of four points, denoted as C,
of vertical springs and the coupling effects between D, E, F in Figure 5, are relatively large during the
different springs can be neglected. Moreover, the depropping procedure. In particular, Points C, D
temporary supports are working under different status; are the highest points of the roof structure (with the
some of them are right below structural joints while same altitude); Point C is where the largest vertical
others are below the structural members. It is hard to displacement happened under self-weight; Point E is
precisely simulate each of them; especially in the the 1/4 point of the lateral span, and Point F is the
design stage when there has limited detail information highest point in the southern part.

48 International Journal of Space Structures Vol. 26 No. 1 2011

Jun Chen, Yixin Peng and Xudong Zhao

(a) Before depropping (b) After step 1

(c) After step 2 (d) After step 3

(e) After step 4 (f) After step 5

(g) After step 6

Figure 4. Deformation of the roof structure at each step of depropping procedure

Displacement transducer is not applicable since it is and GPS are adopted. Observe prisms were installed
difficult, if not impossible, to find an available fixed on Point C to F and total station (type: TCA2003) were
reference point on the temporary structure. Therefore, set on the roof of a nearby office building. The
non-contact measurement techniques as total station measurement accuracy is 0.5 second for angle and

International Journal of Space Structures Vol. 26 No. 1 2011 49

 FEM Simulation and Field Monitoring of Depropping Procedure of a Large-Span Single-Layer Latticed Shell

Total
E station

C Y

y F
E

z D
x

C F

X
A B
A B

Figure 5. Measurement points and arrangement of total station

1 mm + 1 ppm for distance. During the depropping

procedure, the total station scans all the measurement
points every 2~3 minutes to record the variation of
their three dimensional coordinates.
Though the measurement accuracy is very high,
total station has its shortcomings that the target points
must be visible and it is not an all-weather technology.
Moreover, displacements at different points cant be
simultaneously recorded by total station. For dome-
shape roof structure, it is very likely that some parts
(a) Reference station
are not visible from a fixed observation station. To
tackle this problem, the global positioning system
(GPS) technology is adopted to measure the joint
displacement of the structure. One more reason for
using GPS is that it is an important part of structural
health monitoring system (SHM). Installation of SHM
on large space structures has become a worldwide
trend in order to assess structural safety during
structures service life. Therefore, whether the GPS
technology (or SHM system) can be utilized in the
construction stage is worth investigating. To this end,
two reference GPS stations were set at Point A and B (b) Rover station
as shown in Figure 5, and four rover stations were set Figure 6. Setup of GPS measurement system
at Point C to F on the structure. For comparison
purpose, GPS receivers (type: Leica GX1230) were
placed near the total station observe prisms at each from which the final deflection of the roof structure
measurement point. Setup of GPS stations on-site are could be calculated. During the depropping operation,
shown in Figure 6. Before and after the depropping real-time kinematic (RTK) continuous dynamic
operation, the stationary global coordinates of Point C observation was carried out by GPS to continuously
to F were recorded by GPS through static observation, measure displacement of each point with a sampling

50 International Journal of Space Structures Vol. 26 No. 1 2011

Jun Chen, Yixin Peng and Xudong Zhao

Table 1. Time frame of each depropping step

Depropping Number of temporary

Step Time supports removed
Step 1 09:4010:03 33
Step 2 10:0310:27 55
Step 3 10:2710:42 55
Step 4 10:4210:57 55
Step 5 10:5711:13 40
Step 6 11:1311:43 91

Figure 7. Arrangement of strain measurement locations

frequency of 20 HZ. All the GPS sensors are calibrated

before the field measurement. Details of the GPS
technology and the calibration procedure can be found
in Chen (2010).
Based on numerical simulation results, vibrating
wire strain meters were mounted on the surface of steel
members at five locations as shown in Figure 7.
Location 1# to 3# are members connecting to Point C Figure 8. Inside-view of the roof structure after the
where the vibrating wire strain meters were installed depropping procedure
close to the joint and on the upper surface of
rectangular steel pipe. Location 4# is lower chord deflection of the roof structure is around 45 mm. It is
member of the dendritic support and the sensor was compatible with the design results reported in Jin
installed close to the support (i.e the end away from the (2008) that for all the concerned cases of design load
joint). Location 5# is the upper diagonal member of the combinations, the maximum deflection varies in a
dendritic support and the sensor was installed close to range of 4 to 34 mm. In particular, under load case of
the joint of shell. Moreover, an accelerometer is placed Dead Load + Live Load + Raising temperature (30
on Point C to measure the vertical acceleration degrees) the maximum defection is only 4 mm. The
responses during the depropping procedure. relative low defection is mainly due to the arrangement
of dendritic supports in both longitudinal and lateral
5. MEASUREMENT RESULTS AND direction that divide the structure into four parts of
COMPARISON which the actual span is significantly reduced.
5.1. General Figure 9 shows that the numerical results in general
The whole depropping procedure took two and half agree with the variation trend of the field
hour to complete. Time frame of each step is shown in measurements. Point C and D, the highest point of the
Table 1. Figure 8 shows the final inside-view of the roof structure, has a theoretical/measured
roof structure. displacement of 4.8/2.4 mm and 3.5/4.0 mm
respectively. The 1/4 span point E and the highest
5.2. Joint displacement point of lateral span direction F has measured
Figure 9 compares the simulated vertical displacement
deflection of 5.2 mm and 2.0 mm, which are
at Point C to F with the field measured values by total
comparatively apart to the theoretical calculated
station. In each step of depropping procedure, the
values of 2.1 mm and 3.3 mm. Besides, the
measured displacement varies in a range of 1~2 mm
displacement of Point F increases rather than
because the temporary supports cant be
decreases after Step 4 because of the existence of the
simultaneously removed, as did in the numerical
central dendritic supports and the depropping process
simulation. It is seen from Figure 9 that final
is in northern part. Nevertheless, the variation of the

International Journal of Space Structures Vol. 26 No. 1 2011 51

 FEM Simulation and Field Monitoring of Depropping Procedure of a Large-Span Single-Layer Latticed Shell

3 3
Measured (C) 2 Measured (D)
2
Numerical Numerical
1
1
0
Displacement (mm)

Displacement (mm)
0 1
1 2
2 3
4
3
5
4
6
5 7
Step 1 Step 2 S3 S4 S5 Step 6 Step 1 Step 2 S3 S4 S5 Step 6
6 8
09:30 10:00 10:30 11:00 11:30 12:00 09:30 10:00 10:30 11:00 11:30 12:00
Time (hh:mm) Time (hh:mm)

5 2
4 Measured (E)
3 Numerical 1
2
Displacement (mm)

1 0

Displacement (mm)
0
1 1
2
3 2
4
5 3
6
7 4 Measured (F)
8 Numerical
Step 1 Step 2 S3 S4 S5 Step 6
9 5
09:30 10:00 10:30 11:00 11:30 12:00 09:30 10:00 10:30 11:00 11:30 12:00
Time (hh:mm) Time (hh:mm)

Figure 9. Vertical displacement of points C to F during the depropping procedure

reticulated shell structures deflection is within 6 30

Total station
reasonable controlled range, and is far lower than Gps
4 20
1/400 of the lateral span, regulated by Chinese design
Total station (mm)

code for such a structure. 2 10

Gps (mm)
5.3. GPS measurements 0 0

Figure 10 compares the vertical displacement of point 2 10

C measured by GPS and total station. Note that the
GPS and total station measurement has similar trend, 4 20

but the amplitude of GPS data is about five times 6 30

09:30 10:00 10:30 11:00 11:30 12:00 12:30
larger than total station.
Time (hh:mm)
Figure 11a and b shows the horizontal displacement
of Point C in X and Y direction at each depropping Figure 10. Comparition of GPS and total station displamcne
step by GPS. The figure shows the horizontal measurement (point C)
displacement amplitude of Point C is about 2 cm and
there has no significant change during the depropping total station, GPS has two distinct features in
procedure. Figure 11c compares the total station displacement measurement. First, it can simultaneously
measurements with GPS data. Note that variation range measure displacements of all the target points with a
of GPS and total station measurements are similar. relative high frequency, which is very important for
Compared with results in Figure 10, the accuracy of real-time monitoring. Second, GPS needs the target
GPS data in horizontal direction is much higher. The point be visible from the top that is much easier to
conclusion for Point D to F is the same. Compared with achieve than from a fixed horizontal viewpoint.

52 International Journal of Space Structures Vol. 26 No. 1 2011

Jun Chen, Yixin Peng and Xudong Zhao

(a) (b)
34.245 79.750

34.240 79.745
Coordinate in X direction (m)

Coordinate in Y direction (m)

34.235 79.740

34.230 79.735

34.225 79.730

34.220 79.725

34.215 79.720

34.210 79.715
09:40 10:00 10:20 10:40 11:00 11:20 11:40 09:40 10:00 10:20 10:40 11:00 11:20 11:40
Time (hh:mm) Time (hh:mm)
(c)
0.020
Total sation
GPS
0.016
Displacement (m)

0.012

0.008

0.004

0.000

09:40 10:00 10:20 10:40 11:00 11:20 11:40

Time (hh:mm)

Figure 11. Horizontal displacement of point C at different depropping step (by GPS)

Table 2 shows the coordinates, obtained through in displace (plan) and vertical (height) dynamic
GPS static measurements, of each measurement points measurement with reporting sampling frequency of
of the reticulated shell before and after depropping. 10~20 Hz, whilst the measurement accuracy for static
The changes of horizontal displacement of each displacement is around 3 mm. In the case, the final
measurement points before and after depropping is vertical displacement of the roof is lower than 1 cm.
5 mm on average and 8 mm in maximum, which are Thus, the accuracy of GPS is comparatively low in the
slightly larger than total station results. vertical direction and is acceptable in the horizontal
Nowadays, the nominal accuracy of GPS technology direction. A combined scheme of total station and GPS
has been updated to 10 mm and 20 mm resolution is preferable for roof structure monitoring.

Table 2. Coordinates of the point C to F before and after the depropping

X-coordinate (m) Y-coordinate (m) = X 2 + Y 2 (m)

C before 34.2272 79.7328 0.0045
after 34.2283 79.7284
D before 25.0951 87.3634 0.0079
after 25.0968 87.3557
E before 33.8748 92.6733 0.0055
after 33.8876 92.6686
F before 55.6812 74.7755 0.004
after 55.6784 74.7727

International Journal of Space Structures Vol. 26 No. 1 2011 53

 FEM Simulation and Field Monitoring of Depropping Procedure of a Large-Span Single-Layer Latticed Shell

5.4. Stresses of structural members values and measured values are converted into relative
Figure 12 shows the measured stresses at each location values, to which the initial stress value is zero.
during the depropping procedure. Theoretical stress Through comparison, it can be found that the stress
values are calculated through the axial force and change is not significant before and after the depropping
moment at the location where the vibrating wire strain procedure. The computational results are close to the
meter installed. Measured values are calculated through measured values for Point #1, 2, 3, 5, but not for Point
stress-strain relationship. Positive value means tensile 4#, which was installed far from the joint. The variation
stress and negative value means compressive stress. range is about 10 Mpa for Point 1#, 2# and 3#.
For comparison purpose, all the theoretical stress Measuring point 3#, on the upper side of the diagonal
member, is mainly under compress and has a smooth

6 4
4 Measured (1#) 3 Measured (2#)
Numerical Numerical
2 2

0 1
Stress (MPa)

Stress (MPa)
0
2
1
4
2
6
3
8 4
10 5
12 6
09:30 10:00 10:30 11:00 11:30 12:00 09:30 10:00 10:30 11:00 11:30 12:00
Time (hh:mm) Time (hh:mm)

2 8
Measured (3#)
1 Numerical 4
0
0
1
Stress (MPa)
Stress (MPa)

4
2
3 8

4 12
5 16 Measured (4#)
6 Numerical
20
7
09:30 10:00 10:30 11:00 11:30 12:00 09:30 10:00 10:30 11:00 11:30 12:00
Time (hh:mm) Time (hh:mm)

8
6
4
2
Stress (MPa)

0
2
4
6
8 Measured (5#)
Numerical
10
09:30 10:00 10:30 11:00 11:30 12:00
Time (hh:mm)

Figure 12. Stresses of points 1#~5# during depropping procedure

54 International Journal of Space Structures Vol. 26 No. 1 2011

Jun Chen, Yixin Peng and Xudong Zhao

change of stress. The theoretical value agrees well with shells supports during depropping procedure. Its
the measured value. Measuring points 1# and 2#, compressive stress decreases at the first step, increases
located on the main rib and the main load-bearing at the following three steps, and decreases again
member of the structure, is under compression and the during the fifth step to the end. This is because the
stress increases at first and decreases later on. The stress process of depropping procedure is from the south part
changing trend is close to the theoretical value in each to the north part and 4# is located on the north of the
step and can reflect the regularity of the elements stress dendritic supports. At the beginning of depropping
during depropping procedure. Moreover, measuring procedure in the southern, part of the load is borne by
point 1# and 2#, setting on symmetric members, have the peripheral supports on the south, and the stress of
similar stress change trends to each other. the dendritic support 4# located on may decrease.
Measuring point 4#, which is on the lower side of As the depropping procedure gradually moving to the
the second central dendrictic supports and mainly north, the stress increases; and after the last stepthe
under compression, reflects the stress of the reticulated last outward circle of braces are removed, the peripheral

16.4 38 2.5 37
16.6
37 2.0 36
16.8

Temperature (degree)
Temperature (degree)

17.0 36 1.5 35

Stress (Mpa)
Stress (Mpa)

17.2 35 1.0 34
17.4
Point #2
34 0.5 33
17.6 Stress
Temperature
17.8 33 0.0 32
18.0 Point #1
Stress 32 0.5 31
18.2 Temperature
18.4 31 1.0 30
15:00 15:20 15:40 16:00 16:20 15:00 15:20 15:40 16:00 16:20
Time (hh:mm) Time (hh:mm)

4.2 40 0 30.8
Point #3 39 Point #4
4.4 Stress 5 Stress 30.6
Temperature (degree)

Temperature (degree)
Temperature 38 Temperature
4.6 10
Stress (Mpa)

30.4
Stress (Mpa)

37
4.8 36 15 30.2
35
5.0 20 30.0
34
5.2 25 29.8
33

5.4 32 30 29.6
15:00 15:20 15:40 16:00 16:20 15:00 15:20 15:40 16:00 16:20
Time (hh:mm) Time (hh:mm)

10.0 41
40
7.5
Temperature (degree)

39
5.0 38
Stress (Mpa)

Point #5
Stress 37
2.5
Temperature 36
0.0 35
34
2.5
33
5.0 32
15:00 15:20 15:40 16:00 16:20
Time (hh:mm)

Figure 13. Variation of stress after depropping procedure

International Journal of Space Structures Vol. 26 No. 1 2011 55

 FEM Simulation and Field Monitoring of Depropping Procedure of a Large-Span Single-Layer Latticed Shell

Table 3. Comparison of theoretical and test fundamental frequency in each step

Before Step 1 After Step 1 After Step 3 After Step 4 After depropping
Theoretical result 10.4 Hz 8.34 Hz 6.94 Hz 6.91 Hz 6.88 Hz
Test result 13.67 Hz 11.33 Hz 8.20 Hz 7.71 Hz 7.52 Hz
Error 31% 36% 18% 12% 9%

supports bears the main vertical load, the dendritic structure during and after the depropping procedure.
support may have a lower level of stress. The whole After completion of each step, all the construction
process discussed above clearly shows the regularity of activities were temporarily suspended and the vertical
the redistribution of the internal force during acceleration responses of Point C were measured
depropping procedure. under environment excitation. The fundamental
Measuring point 5# is on the upper side of the frequencies of the roof structure in each step are then
dendritic supports diagonal member, whose stress is identified by a method suitable for closely-spaced
similar to 4#s. The measured stress increased at the modes (Chen, 2002). Table 3 shows the measured and
beginning and decreased later and then increased on computed fundamental frequency for every step.
the final stage, which is tally with the theoretical Note that with gradually removing of temporary
analysis. In addition, 5# is near the support, and hence supports, the fundamental frequency decreases
its stress is not as stable as 3#, which is also on a accordingly. The difference between the measured
diagonal member but far away from the support. It is values and the theoretical values reduced from 31% to
clear that the stress is comparatively stable and has not 9% as the depropping procedure goes on. The boundary
abrupt change in the depropping procedure. The conditions and modeling of the temporary supports has
measured values accurately reflect the regularity of the significant influence on the theoretical value.
stress redistribution in the process of the unloading.
It is found that the redistribution of internal forces 6. CONCLUDING REMARKS
in the shell structure dont cease after the temporary Numerical simulations and field measurement of the
supports have been completely removed. Actually, it depropping procedure of a single-layer reticulated shell
might take quite a long time for the redistribution structure are reported in this paper. The following
procedure. As a result, the internal forces in the observations are made from this study. The maximum
structures still gradually increase or decrease after the horizontal displacement and deflection of the roof
depropping procedure. Figure 11 shows the measured structure are measured as 7.9 mm and 6 mm after
stress values at Point 1# to 5# three hours after the depropping. Stress measurement results demonstrate
depropping procedure together with the measured that the variation of the stress is smooth during the
member surface temperature. Note that the depropping and the measured values can reflect the
temperature decreased in the recording period about shells redistribution of the internal forces during and
6 7 degrees, which is not very large. The compress after depropping. The variation trend of deflection,
stress at Point 1# increases from 8 Mpa to 18 Mpa internal forces and fundamental frequency can be well
within the three hours. The variation of stresses at captured by the numerical model, and for certain points
Points 2# and 3# are not significant, but that of Points and structural member the numerical results agree quite
4# and 5# are significant. The stress at Point 4 and 5, well with the field measurements. The numerical
on the dendritic support, continuously and steadily model and simplification method suggested in this
increased but the temperature randomly varied, paper is proper and applicable for analyzing such kind
implying that the increase of stress is mainly due to of structure. GPS and total station are used to monitor
redistribution of internal forces. the displacement of structural system transformation
process. It is found that at current technology level the
5.5. Variation of fundamental frequency of dynamic GPS measurement accuracy is not acceptable
the shell in vertical direction. However, the combined utilization
It is well known that the modes of vibration of latticed of GPS and total station can exert each others
shell structure are generally very close-spaced. It is thus advantages, not only overcomes the disadvantage of
interesting to learn the vibration properties of the shell GPS in low precision of measuring vertical

56 International Journal of Space Structures Vol. 26 No. 1 2011

Jun Chen, Yixin Peng and Xudong Zhao

displacement, but also handles the non-synchronization mode decomposition, Proceedings Earth and Space 2010:
problem of each point when using total station. Engineering, Science, Construction and Operations in
Challenging Environments, ASCE, March 1417, 2010,
Honolulu, HI, pp. 24122421.
ACKNOWLEDGEMENT Cui, X.Q., Gao, Z.F., Li, Z.X. and Wu, X.P, Simplified Analysis
The finical support to this research project from Fok Method of Large Complex Temporary Supports, Spatial
Ying Tung Education Foundation trough 11th Young Structures, 2005, 11(1): 4549. (in Chinese)
Teacher Research Fund (No. 111077) to the first Du, W.F. and Zhang, H., Space Structure, China Electric Power
Press, China. (in Chinese)
author is highly appreciated. Gioncu V. Buckling of reticulated shells: state-of-the-art. Int J
Space Struct 1995; 10(1): 1 46.
REFERENCES Jin G. et al., (2008), Structure design of steel roof and a large
Chen Jun and Xu Youlin, (2002), Identification of modal overhang platform in Tieling gymnasium, Building
damping ratios of structures with closely spaced modal Structure, 38(9): 3739. (in Chinese)
frequencies, Structural Engineering and Mechnics, 14(4): Lopez, A., I. Puente, and M.A. Serna, Numerical model and
417 434. experimental tests on single-layer latticed domes with
Chen Jun, Shang Xiaodong and Zhao Xudong, (2010), GPS semi-rigid joints. Computers & Structures, 2007.
multipath effect mitigation algorithm based on empirical 85(78): p. 360374.

International Journal of Space Structures Vol. 26 No. 1 2011 57