Observations of the November 12, 1999, Düzce Earthquake

Y. Xiao, T.T. Yaprak

While in the mid of recovering from the devastating earthquake of Mw=7.4 on August 17, 1999 with tremendous casualties and damage, the country of Republic Turkey was hit by another new and almost as powerful earthquake on November 12, 1999 at 18:58 local time. The initial magnitude assigned to this event is Mw=7.2, and it was centered near the town of Duzce in the province of Bolu. Duzce is about 60km east of Adapazari, where the August 17, 1999 earthquake had occurred. As of November 26, 1999, the casualties caused by the Duzce earthquake are estimated at 755 along with 4958 injuries. Two structural engineers (Xiao and Yaprak) from the USC Center for Research in Earthquake and Construction Engineering (CRECE) visited the affected site from November 20 to 24, under partial funding support from the Pacific Earthquake Engineering Research Center (PEER). The following brief reconnaissance report is focused on the damage to a modern viaduct bridge. Although the viaduct suffered extensive damage to the bearings and energy dissipation devices with off-siting due to strong shaking and direct crossing of ground rupture through portion of the bridge, the viaduct was able to manage escaping narrowly from collapse. With most of the structural elements completed and without much cosmetic features such as pavement, guardrails, etc. at the moment of the November 12, 1999 event, the viaduct bridge may have probably provided the performance of a "bare" structure, most suitable for detailed studies.

1. General

The Trans European Motorway (TEM) is a toll road that eventually will reach Ankara, the capital of Turkey and shorten the traffic significantly between Istanbul and the capital. While the approximate 300km long portion between Istanbul and Kaynashli is in service, the extension from Kaynashli through the mountainous region to reach Bolu, the capital of the north central province is still under construction. This portion is closely aligned with the Duzce branch of the North Anatolian Fault, which was responsible for the November 12, 1999 event.

The stretch between Kaynashli and Bolu includes the following three major structures:

Viaduct 1:

Length: 2313m

Width: 2x 17.5m

Total number of piers: 58

Max. Pier height: 49m

Max. Span: 39.6m

Pier foundation: 1.8m diameter friction piles

Tunnel:

Length: 3321m

Width: 12m

Height: 8.6m

Clearance Envelope: 110m2

Excavated volume: 220-280m2

Support: 40-60 cm SFRS shotcrete

40-70 cm B30 Unreinforced Concrete Lining

60 cm B40 Unreinforced intermediate Concrete Lining

Maximum Overburden: 250m above tunnel crown

Viaduct 2:

Length: 3559m

Width: 2x 17.5m

Total number of piers: 78

Max. Pier height: 48m

Max. Span: 39.6m

Pier foundation: 1.8m diameter friction piles

At the time of the November 12, 1999 earthquake, Viaduct 1 that crosses the Kaynashli valley (Fig.1) and the tunnel were in their final stage of the construction, while the Viaduct 2 was in the stage of foundation construction. Although the authors could not access to the tunnel, discussions with the site engineers revealed that the tunnel suffered partial collapse near its eastern end where the lining was not fully completed.



Fig. 1





2. Structural Design Features of Viaduct-1:

The viaduct-1 has two bridges each supporting 3-4 lane traffic of eastbound or westbound.

2.1. Substructure

As shown in Fig.2, the monolithic column footings are supported on 12 cast in drilled hole friction piles with a diameter of 1.8m. The depth of the piles ranges from 20m to 30m. The typical size of the footing is 3m deep, 18.7m long (in the bridge longitudinal direction) and 16m wide (in the bridge transverse direction). The footings were designed with both top and bottom reinforcement mats and sufficient shear reinforcement, that was checked by using the strut-and-tie analogy (per conversation with the site engineer).



Fig. 2





The height of the columns ranges from 10m to 49m. The pier columns were essentially designed as vertical cantilevers. The column section is a hollow octagon with outmost length of 8m and width of 4.5m and average wall thickness of 0.6m. The columns were reinforced with two layers of f36 longitudinal bars with f20 transverse reinforcement and f12 cross ties spaced at 10cm. The concrete had a design strength of 30MPa by cubic specimen testing and 25MPa by cylinder testing. Based on the author’s discussion with the site engineer representing the design firm, the columns were designed to respond in elastic for a peak ground acceleration of 0.4g.

2.2. Supper Structures

The superstructure includes PC hollow T-girders with a depth of 1.8m and a 24cm cast in-situ monolithic concrete topping. The PC girders were simply supported by flat stainless steel-PTFE sliding bearings (Fig.3) on top shoulders of the piers with the cast in-situ concrete topping providing some composite effects to tie typically 10 spans together to form a module unit (Fig.4). Extensive amount of energy dissipation devices are also used in the bridge. The central support of the 10-span unit, considered as fixed point of the 10-span unit for thermal expansion, is provided with a multidirectional crescent moon-type steel energy dissipation device (Fig.5). Other supports of the 10-span unit are provided with viscous connecting devices (Fig.6). At the expansion joints between two 10-span units, steel strands restrainers are also provided (Fig.7). At the two abutments, the girders are supported on bearings with bi-directional steel energy dissipation devices. The maximum vertical load of the bearings is 1500kN, and the allowable strokes are ± 410mm for longitudinal direction and ± 210mm for transverse direction. The design elastic limit for the energy dissipation devices is 38mm and the plastic limit is 280mm.



Fig. 3







Fig. 4







Fig. 5







Fig. 6







Fig. 7





2. Effects of the August 17, 1999 Earthquake on Viaduct-1

The Turkish engineers at the construction site have provided some studies on the performance of the viaduct-1 during the previous August M7.4 event. Ground accelerations were recorded at the Duzce station, based on which, the maximum longitudinal accelerations are estimated as 0.39g for longitudinal, 0.31g for transverse and 0.5g for vertical directions, respectively. The relative movement of the beam bearings after the August 17, 1999 event was about 3-8 cm in longitudinal direction and 3-7cm in the transverse direction.

3. Observed Damage to the Viaduct by the November 12, 1999 Earthquake

1. Ground Motion

Unfortunately, it appears that no ground motion record was taken at the site during the November 12, 1999 event. It was said that the strong motion station in Duzce near the epicenter and about 5km to the west of the viaduct was shut off due to the excess of the ground motion above the set limit of 1.0g. The strong motion record in Bolu, about 25km to the east of the viaduct has a maximum acceleration of 0.8g. Thus, it appears to be not unreasonable to believe that the maximum ground acceleration at the bridge site was at least above 1.0g.

After touring the areas affected by the November 12, 1999 earthquake, particularly in Kaynasli where the fault rupture went through the township, the authors had a general impression that there were more collapses of buildings located on the north side of the fault rupture. Another observation was that the majority of the collapsed buildings were typically laying towards the directions of west and east.

3.2. Geotechnical Features

The ground rupture was observed nearby the south side of the bridge, and intersected with the bridge at the Pier No. 45 (counted eastwards) of the south bridge of the viaduct. Soils around the footing of the Pier No. 45 were disturbed significantly, with a pile of about 0.8m high near the northwest corner of the footing (Fig.8). Although there was no visible inclination observed, the Pier No. 45 rotated about 12 degrees about its vertical axis (measured by Dr. J.P. Bardet of USC). Signs of localized liquefaction were also observed near some footings, but did not seem to be a major concern (Fig.9)

Fig. 8

Fig. 9

3.3. Structural Damage:

Damage to the pier columns of the viaduct was minimal. The authors were only able to identify the fine horizontal flexural cracks near the bottom portion of several piers. Structural damage to the viaduct bridges was mainly occurred on the supper structure and the abutments.

The superstructures of the two bridges were observed having a westward permanent displacement relative to the piers, leaving all the ends of the girders off-set from their supports (Fig.10). Stainless sheets, bearing steel plates and the PTFE bearings fell down from the supports were seen on the ground around and near almost all the piers. The danger can be imagined considering the fact that a local road under-crosses the viaduct. Observation of the scratch signs on the surface of the fell down stainless steel sheets indicated that the bearings slid off probably in very early stage before any cyclic movement (Fig.11). This supports the scenario that the bridge encountered a near-fault (more accurately, "on-fault") pulse-type motion on November 12, 1999.

Fig. 10

Fig. 11

The maximum displacements of the superstructure were observed for the supports of the girders on pier No. 44 to No. 50 of the bridges. Some of the outmost PC girders slid off or right on the edge of the concrete pedestals, leaving the bridge supported only by the cast in-situ concrete topping (Fig.12). Cracking of concrete was visible for many PC girders near their ends. Severe concrete spalling and crushing were observed for at least one PC girder supported on Pier No. 46 (Fig.13). These indicated that the bridge was in a very critical stage with the danger of collapse of superstructure occurring at any moment due to aftershocks.

Fig. 12

Fig. 13

Damage to the energy dissipation devices at the fixed supports (on center piers of the 10-span units) was not able to examine due to the difficult accessibility. The observed damage to the energy dissipation device at the east abutment includes the distortion of the steel yielding elements, rupture of the anchor bolts of the device (Fig.14).

Fig. 14

The south side wing walls of the east abutments of the two bridges were damaged due to the pounding by the ends of the bridge girders, leaving a permanent southward transverse offset of about 250mm at the ends of both the south and the north bridges (Fig.15). At the west end of the viaduct, the bridges appeared to have been pushed towards the abutment pounding the back wall and caused concrete crushing and the inclination of the west abutments (Fig.16).

Fig. 15

Fig. 16

4. Hypothesis of Damage Scenario

The above observations appear to support the following hypothesis for the damage scenario. The November 12, 1999 earthquake probably caused the substructure (at least from abutment 1 at the west end to Pier 45) to move eastward suddenly along with the ground on the north side of the fault rupture. Giving the nature of the structure with isolation between the superstructure and the substructure, the isolated superstructure of the bridge did not follow the move of the substructure due to its inertia, thus generating a sudden relative movement (pulse) between the superstructure and the substructure, causing the failure of the bearings and energy dissipation devices, as well as the pounding damage to the west abutment.

5. Concluding Remarks

The viaduct bridge clearly encountered a shaking beyond the original design level in during the November 12, 1999 earthquake. Might be just lucky, but the engineers responsible for the design and construction should be congratulated for the fact that the viaduct did escape from collapse. Further study of the performance of the viaduct is needed, because such study may provide valuable information for future design of bridges subject to near-fault or "on-fault" effects.

6. Acknowledgements

The authors would like to thank the following gentlemen for their kind help,

- Assoc. Prof. Dr. Ahmet Yalciner, Visiting Professor at USC, Civil Engineering Department, Los Angeles, CA

- Hasan Ali Adiguzel, Chief of Design Office, Astaldi S.p.A, Bolu, Turkey

- Recai Yilmaz, Chief Control Engineer, General Directorate of State Highways, Bolu District, Bolu, Turkey