Analyzing the collapse of bridges in Indonesia

Sohei Matsuno, Palembang

It was Dec. 5, 1995 when the Air Beliti Bridge along the Trans-Sumatra Highway collapsed in South Sumatra, some eight and a-half years before Cipunagara Bridge along the North Coast Highway of West Java collapsed on July 23.

The two bridges are of the same Callendar-Hamilton (C-H) design with steel pony trusses. The loads the bridges carried at the time of collapse (six trucks on the Air Beliti Bridge and seven trucks on the Cipunagara bridge) are also similar.

The probable reason given for the most recent collapse is also similar to the one initially given in 1995: An aging bridge overloaded with heavy trucks.

The C-H type bridge was designed to meet engineering conditions in Indonesia, and played an important role in the 1970s and 1980s, with hundreds built throughout the country.

Many adjustments to the bridges have also been made since their initial construction, including efforts to widen the C-H bridge's applicable span with minimum adjustments in parts and components.

One of the enduring qualities of the C-H bridge is that it is well-designed, manufactured and assembled. Hence, it is suppposed to be stable, workable and bankable. The bridges are designed to be durable, with a life span of more than 100 years under current traffic conditions, and all the parts and components of the bridge are galvanized. The structures are also covered layers of zinc that prevent them from rusting.

When reviewing the investigation into the bridge collapse, several important questions should be asked:

If the collapse was caused by aging, why didn't other similar bridges which are even older also collapse?

If overloading was cited as the cause, why didn't other bridges the trucks pass along the North Coast Highway also collapse?

Why was there no warning of the dangers?

If any of the above causes are proven, who should the victims of the accident claim compensation from?

None of the these answers can exclusively explain the accident.

If the accident was caused by aging, there must have been significant warning signs, known in the trade as ominous deflection, and these could have been detected and a warning could have been issued.

The investigation into the 1995 collapse focussed on corrosion and metal fatigue as these are known to be common causes of bridge collapse.

It was confirmed the bridge had shown insignificant ominous deflection before the collapse. It was also found the bridge was fully galvanized. Hence, there was little indication of corrosion and consequential sectional loss throughout the structure.

After these discoveries, corrosion was deleted from the list of possible causes. Aging as a cause of the collapse disappeared at this stage of the investigation.

However, what was observed were apparent beach marks (evidence of fatigue) on a sheared section of the bridge, which proved a bridge support, or member, had broken because of fatigue. However, this kind of fatigue occurrs only when a member is subjected to alternating deformation -- repetitious bending up and down, expanding and contracting.

And alternating deformation does not occur as a rule in this kind of truss design. Therefore it was essential to determine how this sheared member was subjected to the deformation.

The puzzle was solved when it was discovered that the bridge was erroneously supported by fixed shoes (hinges) at both ends. It should have been commonsense to support one end by a hinge and the other end with a movable shoe (rollers). In this way, changes of size due to temperature variations could have been accommodated.

Because of the erroneous shoe arrangement, the upper part of the hinge (upper shoe) nodded up and down due to the longitudinal thermal expansion and contraction of the bridge. After a certain number of cycles, the bridge developed a fatigue crack at a point where it was connected to the upper shoe. The crack gradually grew as a result of the alternating deformation until the member broke. It was subjected to a deformation angle of about plus or minus 1.15 degrees and took an estimated 30,000 cycles to fail.

The next question was to identify if the erroneous shoe arrangement was a design flaw or merely an implementation flaw.

In studying the bridge construction manuals, it was found that instructions in the manual dated Aug. 9, 1977, identified the need to support one end by a hinge and the other with rollers. However in bridge drawings dated Jan. 25, 1977, it was indicated that both ends could use a hinged shoe arrangement.

As it is, it is reasonable to suppose the construction of the bridge -- which began in 1976 -- used the January drawings.

Future corrections in the scheme of the bridge did not take effect after it was completed in 1978.

Investigation of some of the 22 other C-H type bridges in the South Sumatra Province and on the North Coast Highway between Jakarta and Surabaya found similar errors in construction.

While overloading may have been a factor in triggering the bridges' collapse, there is a strong likelihood the cause of the collapse of the Cipunagara Bridge is similar to that of the Air Beliti Bridge, given its similarity of construction and type.

The writer (ypnt@indo.net.id) was a leader of the ad hoc team at the Regional Improvements Office II (RBO-II), Bina Marga that conducted the investigation into the collapse of the Air Beliti Bridge in December 1995. He is currently a senior lecturer at Tridinanti University, Palembang.