ctbuh.org/papers
Title:
Monitoring of Structural Behavior of High-rise Buildings using GPS
Autho Au thors: rs:
Hyo Seon Seon Pa Park, rk, Associa Associate te Professor, Professor, Yonsei Yonsei Un Univ iversi ersity ty Hong Gyoo Shon, Assistant Professor, Yonsei University Ill Soo Kim, Graduate Student, Yonsei University Jae Hwan Park, Graduate Student, Yonsei University
Subject:
Structural Engineering
Keywords:
Structural Health Monitoring Structure Technology
Publication Date:
2004
Oriiginal Pu Or Publica cati tio on:
CTBUH 2004 Se Seoul Co Confe fere ren nce
Paper Type:
1. Book chapter/Part chapter 2. Journ Journal al paper 3. Conference proceeding 4. Unpu npublished blished confer conference ence paper 5. Magazi Magazine ne art article icle 6. Unpublished
© Council on Tall Buildings and Urban Habitat / Hyo Seon Park; Hong Gyoo Shon; Ill Soo Kim; Jae Hwan Park
Monitoring of Structural Behavior of High-rise Buildings using GPS Hyo Seon Park 1, Hong Gyoo Shon2, Ill Soo Kim3, Jae Hwan Park 3 1
Associate Professort, Dept. of Architectural Eng., Yonsei University 2 3
Assistant Professor, Dept. of Civ il Eng., Yonsei University
Graduate Student, Dept. of Architectural Eng., Yonsei University
Abstract A new displacement measuring system using GPS is introduced for monitoring of the lateral and torsional displacements of high-rise buildings. To develop the system, error ranges of the GPS measurement data are examined by varying the distance between a reference point and measuring points. Also, the feasibility a GPS displacement monitoring system is investigated through a physical model experiment. A GPS antenna was mounted on the model, and a laser displacement transducer was installed to measure the actual displacements. Displacements monitored by GPS were found to agree well with actual displacements. Finally, performance of the monitoring system is demonstrated in the full-scale monitoring of a 66-story high-rise multi-purpose building. GPS measurement system was able to accurately resolve movements of a high-rise building into static displacement and dynamic fluctuating displacement components . Keywords: Health monitoring, GPS, Lateral displacement, High-rise building
1. Introduction Structural design of high-rise buildings consists of strength design for assuring safety of the overall structure or its members, and stiffness design to satisfy the limits of the maximum lateral displacement at the top of the building and inter-story drifts. Owing to the tallness and high slenderness ratio of a high-rise building, the relative importance of stiffness design is increasing and structural systems tend to be determined by results of the stiffness design (Chan et al. 1995; Park et al. 2002). In case of a high-rise building that has a relatively high slenderness ratio greater than 5.0 the quality of the structural design depends more on satisfying the serviceability criteria than those of safety (Park and Park, 1997). Serviceability of high rise buildings against lateral loads such as wind loads is evaluated in terms of two types of structural responses: lateral displacement and horizontal acceleration level. Excessive lateral displacement can cause structural problems as well as other diverse problems on Contact Author: Hyo Seon Park, Associate Prof., Dept. of Architectural Eng., Yonsei Univ., Seoul 120-749, Tel: +82-2-2123-2794 Fax: +82-2-365-4668 e-mail:
[email protected]
1064
CTBUH 2004 October 10~13, Seoul, Korea
non-structural elements such as damages to finishing materials, while excessive horizontal acceleration level can bring feelings of unpleasantness to building occupants. For these reasons various researches have been conducted on methods of measuring and controlling relative lateral displacements and horizontal acceleration of high-rise buildings. Measurements of structural responses of high-rise buildings known so far are based on accelerometers (Li at al. 2000; Xu and Zhan 2001). The method of using accelerometers, which are small and light, allows relatively accurate measurement of horizontal accelerations under lateral loads. However, it has such difficulties as problems in displacement reference point setting, accumulated errors from double integration, noise introduction from extended cable connection to storage device, and maintenance problems of cable and measurement devices. Therefore, this method is considered rather difficult to apply for measuring relative displacements of high-rise buildings consisted of static and dynamic fluctuating displacements. As an alternative to the method based on accelerometers, numerous studies have been conducted
to demonstrate the feasibility of GPS for measurement of lateral displacements. Case examples of using GPS for displacement measurement of high-rise building structures include that Loves et al. (1995) applied the method on Calgary Tower and Celebi (2000) applied a GPS measurement technique on simple test models and a 44-story building. Later, Tamura et al. (2002) applied RTK-GPS to perform an accurate experiment and analyzed that the method is useful at displacements of over 2 cm and natural frequency of 2 Hz or less. Also, Breuer et al. (2002) applied the GPS method to measurements of ambient vibrations of a TV-tower and the industrial chimney. High-rise buildings that are subjected to wind load experience lateral displacement accompanied by torsional displacements. Measurement of such distortion displacement has not been reported yet. For complete displacement monitoring of a high-rise building, torsional displacements as well as static and dynamic fluctuating lateral displacements must be measured. In this paper, we analyzed quantitatively error ranges from the baseline distance between a base station and rover stations, and tested the feasibility of the displacement measurement using GPS through a physical model experiment. Also, we applied the verified displacement measurement system on an existing high-rise building to monitor three-dimensional displacement history including torsional displacements under wind load. 2. DGPS DGPS (Differential Global Positioning System)
Fig. 1. Measurement by DGPS 1 DGPS Accuracy Test In developing a lateral displacement monitoring
system of high-rise buildings using GPS, we first identified the error range by the baseline distance between a base station and rover stations of DGPS. Also, we investigated the feasibility of the displacement measurement using DGPS by artificially generating displacements on the test model and comparing the GPS measured displacements against actual laser measurements. Error range by baseline distance In general the impact of ionosphere and the signal
delay through the atmospheric layer within 20 km range are considered as common error components. Here, for the sake of accurate observation we varied the baseline distance to 1 km, 2.5 km, and 4 km and measured errors for 40 minutes at 1 Hz each as shown in Fig. 2.
allows improved point positioning by establishing a reference station at a known position, compares the accurate position of the reference station and the observed position value by GPS receiver to identify errors, then broadcasts the error information to measuring station in the area (Fig. 1). DGPS allows accurate positioning by considering such error components as satellite orbit error, satellite clock error, ionosphere and troposphere time delay as common errors, and eliminating them.
Fig. 2. Accuracy test scheme for DGPS 1
CTBUH 2004 October 10~13, Seoul, Korea
1065
1.5
1.0
0.5
0.0 -1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
X (cm)
-0.5
-1.0
-1.5
Y (cm)
(a) Baseline distance 1 km
1.5
1.0
0.5
X (cm)
0.0 -1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
-0.5
-1.0
-1.5
Y (cm)
(b) Baseline distance 2.5 km
Geometric errors in GPS arise from the geometric arrangement of four or more observed satellites and are expressed as PDOP (Position Dilution Of Precision) values. Measurement data with PDOP of five or less can be trusted so that t he smaller the PDOP value, the less the error. Additionally, multipath caused by environmental conditions around rover stations causes an error component. In our experiment we obtained measurements at PDOP 3 or less in order to minimize the geometric error factor of GPS. To avoid the multipath error, we performed measurements from school grounds in downtown area for positions A and B and from a park away from downtown for position C (Fig. 2). Measured data were processed using PPK (Post Processed Kinematic). Then, we projected the three-dimensional displacement loci of GPS measurements onto a local X-Y plane to represent it as a two-dimensional displacement as shown in Fig. 3. Standard deviations of the data measured with Trimble 4700 system for baseline of 10 km or less show horizontal error of 1 cm+1 ppm and vertical error of 2 cm+1 ppm. Measurements of 40 minutes duration for baseline distance of 1 km, 2.5 km, and 4 km prove, as shown in Table 1, that their standard deviations fall within acceptable confidence intervals. As Table 1 shows, the variable range at position C according to its standard deviation was 0.26 cm on X-axis and 0.23 cm on Y-axis.
1.5
Table 1. Test results for different baseline distances 1.0
A
B
C
1 km
2.5 km
4 km
+
0.63
0.76
0.93
-
0.77
1.14
0.74
+
0.93
1.62
0.96
-
1.27
1.08
0.57
X
-0.01
0.01
0.00
Y
-0.02
0.15
0.12
X
0.22
0.28
0.26
Y
0.34
0.47
0.23
0.5
Baseline distance 0.0 -1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
X (cm)
-0.5
-1.0
X Displacement range (cm)
Y
-1.5
Y (cm)
(c) Baseline distance 4 km
Average displacement (cm)
Fig. 3. 40-min. measurement of a fixed point with
Standard deviation (cm)
different baseline distances
1066
CTBUH 2004 October 10~13, Seoul, Korea
Thus, position C had the least variations compared to positions A and B. It is probably because of the location being away from downtown area so that the effect of multipath was minimal.
1.0
Laser displacement meter GPS
0.8
) m c ( t n e m e c a l p s i D
0.6 0.4 0.2 0.0 0
5
10
15
20
25
-0.2 -0.4
Model experiment using GPS, laser displacement measurement device, and accelerometer As shown in Fig. 4, we installed a GPS receiver,
accelerometers, and a laser displacement meter to compare displacement histories and the level of accelerations by two measurement devices in free vibration.
Y +
-0.6 -0.8
Time (sec)
-1.0
(a) Initial displacement 1 cm 2.0
Laser displacement meter GPS
1.5
) m c ( t n e m e c a l p s i D
1.0 0.5 0.0 0
5
10
15
20
25
-0.5 -1.0 -1.5 -2.0
GPS
Time (sec)
Accelerometer Laser displacement meter Accelerometer
+
X
(b) Initial displacement 2 cm
Laser displacement meter
2
Fig. 4. Experimental model
) 1 m c ( t n e 0 m e c a l p s -1 i D
Laser displacement meter GPS
0
5
10
15
20
25
-2
Time (sec)
-3
(c) Initial displacement 3 cm 4
Laser displacement meter GPS
3 2
Fig. 5. Measurement devices and data acquisition coordinate convention
) m c 1 ( t n e 0 m 0 e c -1 a l p s i -2 D
5
10
15
20
25
-3 -4
The experiment body used was a wooden plate of 2.44 m by 1.24 m while the six vertical elements were of D10 rebars (Fig. 5). The model was set into free vibration along X-axis with a given initial displacement. In order to eliminate Y-axis oscillation component, we installed braces on vertical rebars and shock-absorbent pads on connection points to reduce energy loss at these points during free vibration.
Time (sec)
-5
(d) Initial displacement 4 cm
Fig. 6. Displacement measurement by GPS and laser
displacement meter
CTBUH 2004 October 10~13, Seoul, Korea
1067
The initial displacement of free vibration was varied from 1 cm to 2 cm, 3 cm, then to 4 cm increasingly and measurements were taken at frequency of 5 Hz. Installation of measurement devices and the data acquisition code convention were as shown in Fig. 4. The results of this experiment showed, as in Fig. 6, that both displacement histories using laser meter and GPS coincided regardless of initial oscillation amplitude. However, GPS measurements for displacement amplitude of 0.5 cm or less had a relatively low confidence level. Comparison of acceleration as differentiated from measured displacements using GPS against actually measured acceleration using servo-type accelerometer is shown as in Fig. 7. The GPS and laser meter accelerations are computed by numerically differentiating displacement values of three consecutive points. As shown in Fig. 6 for the case of displacement history graph, acceleration obtained using GPS receiver coincide quite well with actual acceleration measured with accelerometer.
30
Accelerometer Laser displacement meter GPS
20
) s 10 / m c ( n 0 o i t 0 a r e l e -10 c c A
2
5
10
15
20
25
-20
Time (sec)
-30
(c) Initial displacement 3cm
40
Accelerometer Laser displacement meter GPS
30
) s 20 / m c ( 10 n o i t a 0 r e 0 l e c -10 c A
2
5
10
15
20
25
-20 10
) 5 s / m c ( n o i t 0 a r 0 e l e c c A
-30
Accelerometer Laser displacement meter GPS
Time (sec)
-40
2
5
10
15
20
25
(d) Initial displacement 4cm Fig. 7. Acceleration measurement by GPS, laser displacement meter and accelerometer
-5
We then performed Fast Fourier Transform (FFT) on acceleration data using GPS, laser displacement meter, and accelerometer, respectively, to compute natural frequency of the physical model. The three results provide identical natural frequency of 0.6 Hz as shown in Fig. 8.
Time (sec)
-10
(a) Initial displacement 1cm
15
10
) 2 s / m 5 c ( n o 0 i t 0 a r e l e -5 c c A
0.06
Accelerometer Laser displacement meter GPS
0.05
) s / 2 m c ( x S 3
5
10
15
20
25
0.6Hz
Accelerometer Laser displacement meter GPS
0.04
0.03
0.02
0.01
-10 0.00
-15
0
Time (sec)
1
2
Frequency (Hz)
(b) Initial displacement 2cm Fig. 8. Natural frequency comparison by GPS, laser displacement meter, and accelerometer
1068
CTBUH 2004 October 10~13, Seoul, Korea
Measuring point
m 9 . 3 3 2
The symbol convention for data acquisition was determined so that the lateral direction was set as X-axis for convenience of data acquisition as shown in Fig. 10. In order to measure torsional displacements we installed GPS_1 at point B of Fig. 11 and GPS_2 at a distance of 16,182 mm from point B. The GPS base station as the reference point was installed on the roof of a five-story apartment building about 600 m away from the measurement building to ensure no lateral displacement (Fig. 10). Y direction (+) 0° N Anemometer2
Reference point
GPS_1 Accelerometer m Anemometer1 1
B
GPS_2
. 5 3
X d i 9 e r 0 c ° t i o n ( + )
42.6 m Base floor plan
600 m
42.6 m
Fig. 9. A high-rise building for full-scale test
3. Full-scale Measurement Building summary
The building used for real measurements is a 66-story multi-purpose facility of reinforced concrete structure with shear wall and outrigger system (Fig. 9). Its height to the heliport where the GPS equipments were installed is 233.9 m and the slenderness ratio is 6.63. Measurement equipments
The measurement equipments included GPS antennas for measuring the building's displacement, anemometers, which consist of a wind meter (Model 05103) and a display meter (Model 04503) by Young Inc. The measurable range of wind speed was 0 m/s ~ 60 m/s, wind speed accuracy of ±0.3 m/s and wind direction accuracy ±3º. GPS equipments used were Trimble 4700 with dimensions of W 11.9 cm x H 6.6 cm x L 20.8 cm, weight 1.22 Kg, C/A code and reflection wave L1 and L2 receiver with automatic OTF initialization from five SV. The antenna was micro-centered antenna (P/N 14553-01) and the antenna cable was a 10 m low-loss, dedicated antenna cable (P/N 14553-01). Measurement summary
Fig. 10. Measurement equipment installation positions and symbols
Both the base station and rover stations used Trimble's software to take measurements at 5 Hz and the data was stored in the computer. The measurement data was then processed using Trimble Geometric Office software at 1 Hz. Processed 3-dimensional data displayed in WGS84 (World Geodetic System 1984) coordinates was then transformed into two-dimensional coordinates by establishing a local X-Y coordinate system on the building roof. Data measurements were performed for ten minutes continuously from 16:10 to 16:20 on 22 March 2002.
4. Analysis of Measurement Results Lateral displacements Since displacements measured with GPS could be obtained from comparison of reference point coordinates and measured coordinates, the reference
point coordinates for relative displacement measurement were set before de-facto experiment at a time period with wind speed less than 4.3 m/s. Displacement measurements were performed on a day when the wind speed exceeded 10 m/s with yellow dust storm blowing. The ten-minute measurement of displacement history at the top of the building is
CTBUH 2004 October 10~13, Seoul, Korea
1069
displayed in three-dimensional displacement loci in global coordinates as shown in Fig. 11.
40
20
) m m0 ( Z
Fig. 12 and Fig. 13 show time histories of lateral movements of the building for each direction upon processing of measured displacement data using GPS. As seen in the measured data the X-axis displacement during the ten minutes varied in the range -11.7 ~ 20.9 mm with a mean of 3 mm, while the Y-axis displacement varied in the range 31.8 ~ 61 mm with a mean of 44 mm. Acceleration measured with GPS and accelerometer
-20
80
-40 -40
60
-20
) m m Y ( 40
X ( 0 m m )
20
20 40
0
Fig. 11. 3-dimensional movement measured by GPS
25
We installed two servo-type accelerometers at position B as shown in Fig. 10 in order to compare computed acceleration from GPS measured displacement data against acceleration from accelerometers. We performed two stage differentiation of X-axis and Y-axis direction GPS measured displacements to obtain accelerations. Comparison of these computed accelerations against those from accelerometers for 10 seconds within an hour is as shown in Fig. 14. The result shows that these two sets of accelerations are almost identical.
20
) m m ( t n e m e c a l p s i D
15 15
10 5
Accelerometer GPS
10
0 -5 -10 -15 0
100
200
300
400
500
600
Time (sec)
) 2 s / m m ( n o i t a r e l e c c A
5
0
-5
-10
Fig. 12. Time history of movement of a 66-story building in X-direction
-15 60
62
64
66
68
70
Time (sec)
Fig. 14. Comparison of accelerations form GPS and accelerometers in X-axis acceleration
65 60
) m m ( t n e m e c a l p s i D
55
Estimation of vibration level Since this building's acceleration response of cross-windward direction is greater than that of
50 45 40 35 30 0
100
200
300
400
500
600
Time (sec)
Fig. 13. Time history of movement of a 66-story building in Y-direction
1070
CTBUH 2004 October 10~13, Seoul, Korea
windward direction, we used Y-axis acceleration for assessment of the serviceability. Fig. 15 presents comparison of maximum acceleration response on average wind speed of ten minutes measured with GPS and accelerometers against the forecast equation in the Japanese habitability evaluation guideline. As seen in the figure the maximum computed acceleration values
Research Foundation under grant KRF-2001-042 -E00137, which is gratefully acknowledged.
10
Full scale measurement of GPS Full scale measurement of Accelerometer AIJ
) s / m c 1 ( n o i t a r e l e c c a 0.1 n u m i x a M
2
Reference Breuer P., Chmielewski T., Górski P., and Konopka E. (2002). “Application of GPS technology to measurements of displacements of high-rise structures due to weak winds” Journal of Wind Engineering and Industrial Aerodynamics, 90, 223~230. Çelebi M. (2000). “GPS in dynamic monitoring of long-period structures” Soil Dynamics and Earthquake Engineering, 20,
0.01 1
10
100
Mean wind speed (m/s)
477~483. Chan C.M., Grierson D.E., and Sherbourne A.N. (1995). “Automatic
Fig. 15. Relation between wind speed and acceleration response in Y-direction
optimal
design
of
tall
steel
building
frameworks” Journal of structural engineering, ASCE, 121(5), 838-847. Li Q.S., Wong C.K., Fang J.Q., Jeary A.P., and Chow Y.W.
from the two measurement techniques are either similar or less than that of the estimation equation.
(2000). “Field measurements of wind structural responses of a 70-storey tall buildings under typhoon conditions” The Structural Design of Tall Buildings, 9, 325-342.
5. Conclusions As a basic research for development of an objective and rational high-rise building maintenance and management system to secure safety of a structure and to improve serviceability of a building, we proposed a three-dimensional displacement monitoring system based on GPS. For the system development we analyzed error ranges of the GPS measurement data in terms of base station distance. We then verified the feasibility of a GPS displacement monitoring system using a physical model experiment. We applied the developed system on a 66-story high-rise building to perform translational and torsional displacement history monitoring under wind load.
Loves J.W., Teskey W.F., Lachapelle G., and Cannon M.E. (1995). “Dynamic deformation monitoring of tall structure using GPS technology” Journal of Surveying Engineering, 121(1), 35~40. Park H.S. and Park C.L. (1997), “Drift control of high-rise buildings with unit load method”, The Structural Design of Tall Buildings, 6, 23-35. Park H.S., Hong K.P., and Seo J.H. (2002). “Drift design of steel-frame shear-wall systems for tall buildings” The Structural Design of Tall Buildings, 11, 35-49. Tamura Y., Matsui M., Pagnini L.C., Ishibashi R., and Yoshida A. (2002). “Measurement of winnd-induced response of buildings using RTK-GPS” Journal of Wind Engineering and Industrial Aerodynamics, 90, 1783~1793. Xu Y.L., and Zhan S., (2001). “Field measurement of Di Wang
Acknowledgement This material is based on work sponsored by Korea
Tower during typhoon York” Journal of Wind Engineering and Industrial Aerodynamics, 89(1), 73-93.
CTBUH 2004 October 10~13, Seoul, Korea
1071
