Ascecf 1943-5509 0000419
Ascecf 1943-5509 0000419
Ascecf 1943-5509 0000419
net/publication/275183316
CITATIONS READS
26 766
4 authors, including:
Some of the authors of this publication are also working on these related projects:
A Verification Study on Bridges Supported by Pile Shaft Foundations with Non-Destructive Depth Inspection Technique View project
All content following this page was uploaded by Helsin Wang on 21 July 2015.
Abstract: In Taiwan, many bridges span rivers that exhibit severe channel variations in response to drastically increased flows and resulting
floods caused by typhoons and storms. Flood-induced scour can undermine the effective embedment depth of bridge foundations. The
Downloaded from ascelibrary.org by National Taiwan University on 01/22/14. Copyright ASCE. For personal use only; all rights reserved.
Shuang-Yuan Bridge, spanning a downstream section of the Kaoping Stream, was severely damaged during Typhoon Morakot in August
2009. A qualitative investigation was conducted to determine the possible causes of this bridge collapse. Relevant hydrologic, meteorologic,
structural, geotechnical, and terrain information from various sources both in written and image formats was collected, updated, and
reviewed. The disaster investigation included field investigations, nondestructive inspection of bridge remnants, and investigation of struc-
tural damage of upstream river crossings, hydrologic changes, damage to training structures/embankment, riverbed profile changes, and de-
bris during the flood event. Several inspections were applied to determine the embedment depths, positions of underwater remnants, and
changes in riverbed elevation. The investigation results identified multiple possible causes of failure, including flood flows, river-bend-
induced turbulent flows, and the joint effects of extended foundations and rafted wood. DOI: 10.1061/(ASCE)CF.1943-5509.0000419.
© 2014 American Society of Civil Engineers.
Author keywords: Forensic engineering; Scour; Bridges; Multiple natural hazards; Inspection; Floods; Migration.
Typhoon Morakot hovered around Taiwan for 3 days, and its suc-
cessively accompanying moist air current brought more than 2,500
mm of heavy precipitation to southern Taiwan in early August 2009.
This was the most severe and damaging typhoon encountered in the
past 50 years. The event left 704 people dead and 18 others missing
and caused approximately NT$110 billion (US$3.7 billion) in
damage [United Nations Office for the Coordination of Humani-
tarian Affairs (OCHA) 2009]. The typhoon also caused serious slope
failures, numerous debris flows, and floods with more than a 200-
year recurrence interval, along with damage to approximately 130
highway bridges [CECI 2009; Chang et al. 2009; FAMCI 2009;
Downloaded from ascelibrary.org by National Taiwan University on 01/22/14. Copyright ASCE. For personal use only; all rights reserved.
Archive Establishment
imum kinetic energy, have been marked using arrows in Fig. 4 to have been recorded for the period 1972–2009, as shown in Fig. 5.
determine the channel migration path. The shifting main channel Various exposed foundation depths ranging from 2 to 6 m were
directly undercuts the left-bank floodplain and expands the main- frequently observed on Piers P6–P18 located at the main channel.
channel width up to 1,200 m at the bridge site. The trend in local The most recent predisaster inspection revealed the continuous
hydrologic migration has subsequently caused the bend (erosion extension of the main channel from P6 to P4, as well as a maximum
position) on the right-bank floodplain to gradually move from an scour depth of up to 7 m at Pier P5 (CECI 2011b). Such bank
upstream position downward to the bridge site (from Points A to B degradation implied that lateral erosion had again been launched on
to C). The approaching bending vertex could have continuously the right bank. This could have led to elimination of the flood-
undermined the structural stability of the abutment and piers and resistance capacity of the bridge foundations during the typhoon. A
conservative measure of the maximum postdisaster scour depth was
13 m at the P15 position, measured using sonar techniques (details
are described in the section on the disaster investigation record).
Fig. 3. Elevation evolution of longitudinal profile along main channel from Kaoping estuary (data from the Water Resources Planning Institute 2008)
Fig. 4. Progressive lateral erosion of main channel toward left-bank floodplain and shifting bending-vertex zone toward bridge on right bank
(schematics redrawn based on the satellite images from the Aerial Survey Office, Forestry Bureau, Taipei, Taiwan)
any direct pressure or suction effects on the flow velocity of the river
A comprehensive bridge inspection conducted in 2000 revealed
and consequently did not cause the bridge collapse.
structural safety issues such as concrete spalling and reinforcement
corrosion on the superstructure. Even though no significant scour
Structural and Geotechnical Information threat was found around the bridge piers, foundation underpinning
and relevant protection works were recommended to efficiently el-
evate the foundation bearing capacity and prevent unexpected flood
Original Structural and Geotechnical Design scour on the foundations.
In 2003, a structural retrofit project was launched to replace
The designated structural elements shown in Table 1, based on data compromised structural elements and underpin selected foundations
from the TBMS archives, reveal that the bridge was 2,083 m long and protection works at 30 piers, mostly located in the main channel
and 9.5 m wide and consisted of two parallel I-type prestressed- (Piers P5–P14). To increase the bearing capacity of the foundations,
concrete bridges built in different years. The older bridge on the four 0.9-m-diameter drilled shafts with a length of 50 m extended
upstream side was constructed in 1974. The newer bridge on the a 4.3-m-wide pile cap on each pier perpendicular to the flow di-
downstream side, an identical structural type to the old section, was rection (see Fig. 8). This extended pile cap could have led to dis-
built in 1981 in a bid to accommodate increasing traffic flows. The charge loss along the bridge site during the typhoon flood.
100-year recurrence flow rate was the basic design reference for the The archives of the TBMS also provide relevant information
bridge foundations. from regular bridge inspections. The routine inspections performed
The superstructure of the bridge was composed of 66 spans, each
30.6 m long, with two 31.6-m-long prestressed-concrete spans adja-
cent to the two abutments. The maximum horizontal acceleration for Table 1. Archive of the Shuang-Yuan Bridge (Data from the TBMS)
earthquake resistance was designed as 0:23g without any further Geometric dimensions
earthquake-resistance measures. The substructures of the bridges
comprised 67 cantilever-type prestressed-concrete piers with a max- Length 2,082.8 m
imum height of 9.2 m. The designated minimum clearance height Span number 68
was 7 m between the main-beam bottom and the still-water surface. Width 9.5 m
Span disposition 2 @ 31.60 m (two ends) 1 66
@ 30.60 m
Pier height 9.2 m
Pile length 33.1–50 m
Structural elements
I-Type main beam 94–95-m prestressed-
concrete beam
Transverse beam RC partition every 10 m
Deck RC structure paved with AC
Pier Prestressed-concrete
structure
Pile Reverse-circulation
installed-concrete structure
Abutment RC cantilever supported by
concrete piles
Bearing Rubber
Extension joint Sawed steel plate
Loading parameters
Live load HS20
Falling prevention length 0.7 m
Fig. 6. Stage hydrographs measured at gauging and tidal stations
Horizontal accelerationa 0.23g
during Typhoon Morakot (time scale begins at 00:00 on August 7, 2009)
Earthquake zoning B
(data from Central Weather Bureau and Water Resources Agency,
Note: AC 5 asphalt concrete.
Taiwan) a
No earthquake-resistant system is designed for the bridge.
Fig. 7. Geologic profile along bridge site (from CECI 2011b, with permission)
Fig. 8. Design graphs of underpinned bridge foundation (unit: centimeters) (data from the Directorate General of Highways Bridge Archive)
in 2006, 2007, and 2008 indicated that the improvements had lifted failure, a temporary steel bridge supported with 50-m-long steel-
the Shuang-Yuan Bridge to a fine structural condition level. The pipe foundations was constructed over the downstream-side bridge
most recent precollapse inspection, carried out in March 2009, [Figs. 9(k and l)].
continued to classify the bridge safety level as in good condition. A resistivity image profiling (RIP) technique was applied to map
the electronic potential field of the damaged bridge site on the up-
stream side [Fig. 9(l)] (Wang et al. 2010). Line R1, 440 m long, was
Disaster Investigation Record extended from the position at Pier P1 to the position between Piers
P15 and P16 along the upstream side of the original bridge site. The
The conventional field investigations and subsequent inspections electrode probes were fixed on the tips of bamboo staffs and po-
were executed from August 12, 2009, to March 18, 2010, as shown sitioned on the river bottom by a diver. Line R2, 165 m long, was
in Fig. 9. The disaster investigation addressed six key items, in- positioned beside the original Pier P3 position, parallel to the flow
cluding bridge remnants, structural damage to upstream river cross- direction of the river.
ings, hydrologic changes, training-structure/embankment damage, The RIP results of Lines R1 and R2 are shown in the upper
riverbed profile change, and debris. This investigation provided portion of Fig. 10. The resistivity contour with a display mode of
crucial information about and evidence for determining the key visible-light spectrum, with corresponding resistivity values from
cause(s) of failure (Wu et al. 2014). 100 to less than 1 V-m, illustrates the resistivity intensity distributions
in space. The resistivity value for sediments composed of silt and sand
was more than 5 V-m. The steel-content structural pieces had lower
resistivity values, below 1 V-m.
Bridge Remnants
The RIP results are interpreted as follows (the lower portion of
The satellite and field images [Figs. 1(b and c), 4, and 9] indicated Fig. 10): (1) there was a positive identification of the foundation
that bridge sections between Piers P1 and P14 on the upstream side remnants at the Pier P2, P3, and P13 positions at depths of more
of the bridge were completely swept away, with similar conditions than 20 m, (2) the horizontal image configurations at depths from
between Piers P1 and P16 on the downstream side. On the right-bank 25 to 50 m and 15 m represented the scattering of bridge decks
section of the bridge, only one-span remnants from Abutment A1 along Piers P5–P8 and P9–P11, respectively, and (3) a high re-
and Pier P1 [Figs. 1(c) and 9(b, d, and j)] remained. One of the bridge sistivity distribution above 20–30 m deep along Piers P5–P11
slabs was askew and was suspended from Pier P14 of the upstream indicated the area of maximum scour or fluidization depths during
side of the bridge [Figs. 1(c) and 9(e)]. Five months after the bridge the typhoon.
Fig. 9. Postdisaster site records: (a) broken right upstream embankment; (b) remnant of bridge; (c) deposition of rafted wood; (d) remnants of Pier P1
and Abutment A1; (e) ground-penetrating radar inspection around Pier P14; (f) remnant Piers P14 and P16; (g) drifting wood rafted on Piers P39 and
P40; (h) broken right upstream embankment; (i) damage at Kaoping Weir; (j) sonar inspection over main channel; (k) construction of temporary bridge;
(l) RIP inspection along upstream side of bridge site (from CECI 2011b, with permission)
Fig. 10. RIP results along upstream side of bridge site (Line R1) and beside former Pier P3 position (Line R2) and possible underwater foundation or
structural remnants
GPR inspection with antenna frequencies of 200, 450, and 900 MHz elements from Piers P4 to P12 in the main channel. Foundation
was conducted along the main channel to map the streambed ele- remnants indicated the possibility of Stage I collapse (underpinned
vation. The measurable zone was spatially limited to the region of Piers P4–P16) under such severe flood conditions.
Piers P2–P20. However, because the high salt content in the water
significantly lowered signal resolution, the riverbed profile could
not be accurately determined. Accordingly, sonar techniques were Conclusions
employed to conclusively survey the riverbed elevation after the
typhoon flood. The new riverbed elevation profile is plotted in Fig. 5. The typhoon-induced extreme precipitation caused slope failures,
A maximum depth of 13 m was graphically found at the Pier P15 debris flows, barrier lakes, and unprecedented floods in southern
position, with actual scour depth deeper than 13 m during the bridge Taiwan in early August 2009. Despite its good maintenance record,
collapse. the Shuang-Yuan Bridge, one of 130 damaged highway bridges, still
collapsed. This also highlights possible improvements that could be
made to current maintenance mechanisms used by bridge agencies.
Debris A forensic investigation into the bridge collapse was conducted
Typhoon-induced damage caused drifting woody debris to be rafted to identify the main causes of failure. Based on the investigation
along the riverbank and bridge site [Figs. 9(a, c, and g)]. The highest results, the following conclusions can be drawn:
standing elevation of the wood debris implied that the flood-stage 1. A qualitative assessment process for flood-induced bridge
elevation was at least 6 m during the typhoon flood. The observed failure was developed to review relevant evidence and effec-
elevation was identical to the official proclaimed flood stage of 6 m. tively evaluate all possibilities through the establishment of an
This evidence also provides further insight into the various possible archival database. The hydrologic, meteorologic, structural,
causes for discharge loss, flow-velocity distribution, and rafting geotechnical, and damage information was collated to estab-
elevation that would be helpful for later scenario evaluation. lish a forensic archive to support the investigation into this
flood-induced bridge failure.
2. The relevant information covered related long-term hydro-
Identification of Possible Causes logic migration, hydraulic engineering countermeasures, me-
teorologic events, whole-life bridge structural records, and site
The establishment of an archival database assisted investigators in geotechnical characteristics from diverse sources existing in
their comprehensive review of the hydrologic, meteorologic, struc- the current TBMS archive and field investigation.
tural, geotechnical, and disaster information leading to the bridge 3. The satellite images and field scour measurement indicated
collapse event. Several possible causes of failure were identified as that the bending vertex (erosion position) on the right-bank
unprecedented flood flows, river-bend-induced skewed flows, tidal floodplain approached the bridge, induced by local hydrologic
effects, the effects of extended foundations, and the effects of rafted migration, and continuously undermined the structural stabil-
woody debris. ity of the bridge. The unexpected effect of extended pile caps,
The investigation results reveal that the rainfall that exceeded which were used to underpin the bridge foundations, was to
2,500 mm in the watershed upstream induced flow rates of over 200- cause discharge losses and increased flow velocity along the
year recurrence, as observed at the three upstream gauging stations. bridge site during the typhoon flood.
The tidal variation did not appear to directly contribute to the bridge 4. The disaster investigation recorded the conditions of bridge
collapse event, given the low tidal elevation. The fierce typhoon remnants, structural damage to upstream river crossings, hy-
flood also prompted the local hydrologic migration of the river. The drologic changes, training-structure/embankment damage, riv-
rapid movement of the upstream bending vertex (main erosion po- erbed profile changes, and debris. The resistivity image profiling
sition) approached Abutment A1 and the piers of the bridge adjacent indicated the likely scour and fluidization, which presents the
to the right bank during the typhoon. The bending-vertex zone temporarily deposited soil performing an aquiform mechanical
gradually shifted onto the bridge site and directly induced Stage behavior, not susceptible to any loading, and vulnerably af-
II collapse (unimproved Piers P2–P4) and failure of the right fected by local flood conditions at depths of around 20–30 m
embankment. during the typhoon. All structural elements of Piers P4–P12
Foundation underpinning with extended pile caps also could were completely swept away at the main channel. The rafted
have caused discharge loss along the bridge site, speeded up the flow woody debris formed a discharge blockage and sped up the flow
velocity, deepened the exposed depths of the foundations at the main velocity along the bridge site during the typhoon flood.
channel, and eliminated the flood-resistance capacity of the bridge 5. Multiple causes, including overflooding, river-bending-induced
subjected to the typhoon flood thrust. Likewise, a great amount of flow, and the joint effects of extended foundations and rafted
woody debris removed from the upstream watershed could have wood, induced the Shuang-Yuan Bridge collapse. The effects of