New Civil Engineering magazine has published a summary of the official investigation report into the 26 November 2011 collapse of the Kutai Kartanegara suspension bridge in Indonesia (image below courtesy of Ezagren at Wikipedia). You may need to be an NCE subscriber to see their report, unfortunately.
The 470m long bridge collapsed while maintenance work was ongoing, killing at least 20 people and injuring many more. Only the bridge towers and main cables were left in place after the collapse (image courtesy of Katakutu on Wikipedia).
The NCE report isn't completely clear, but it indicates that the bridge was originally built with a 3.7m rise in the deck at midspan, which by late 2011 had reduced to 3m. Maintenance work was instructed with the aim of restoring the original profile, apparently by jacking at each vertical hanger. NCE states that the sag was the result of a 200mm horizontal movement of one tower foundation, but this appears incorrect, and movement of one of the main cable anchorage blocks seems the more likely cause (see below).
The report notes that collapse was initiated by shear failure of one of the hanger-to-deck connections, presumably near the tower where the hanger loads will have been increased most severely by any sag. The loads were then exacerbated when the maintenance contractor jacked on one side of the bridge deck only - while traffic continued to run on the other side. Reportedly, no calculations were made as to what effect this jacking sequence would have, although it seems intuitively obvious that it would lead to greatly increased hanger loads.
The bolts connecting the hangers to the deck were also found to be very brittle, almost like cast-iron in quality. This is quite remarkable, as use of ductile material in bridge works is just a commonplace in modern construction - to use something different is almost perverse.
The entire bridge apparently collapsed in about 30 seconds - after the first hanger connection failed, its load was redistributed to adjacent hangers, which also failed, unzipping the entire bridge deck very quickly. Although the initial failure may have been at deck level, the photos clearly imply that further failures occurred at the connection of the hangers to the main cable.
A little more light is shed on matters by an unofficial report from Professor Sohei Matsuno of IBA University, available online as a PDF. This notes that the anchor blocks had moved 100mm horizontally, causing the towers to tilt and the deck to sag. Matsuno blames this on the absence of raking piles from the anchor block foundations, which strikes me as a reasonable complaint.
Matsuno's report notes that the main cable sag will have redistributed forces in the hangers, such that some will have lost most or all of their tension, and become prone to significant movement under live load. Alternate loading in the cables may have damaged the joints. In any event, failure at a joint is normally something which is prevented in design: joints are generally designed to be stronger than the parts that they connect, which does not appear to have been the case on this bridge.
If a hanger was sufficiently overloaded to fail in a well-designed bridge, you would expect the cable to fail (probably right next to the joint), not its connection. You would also expect the bridge deck and adjacent hangers to have been designed to cope with loss of one (and possibly more) hangers. This is normal design practice both to safeguard against disproportionate collapse due to material failure, impact damage, or sabotage, as well as to make hanger replacement more straightforward.
To summarise, the bridge deck failed because the hanger joints at top and bottom were too weak, possibly exacerbated by the use of brittle components; hanger loads were increased by the bridge deck sag and by the irrational jacking sequence; hanger movement may have caused damage; and there was insufficient ductility and redundancy in the bridge to cope with the loss of one or more local components.
Although the reports available don't entirely clarify what went wrong technically (whether it was poor design or poor construction, for example), I think the non-technical reasons for failure may be more interesting.
The UK's Standing Committee on Structural Safety (SCOSS) promotes the view that structural failures all result from the "3 Ps" - people, process and product. An obsession with the technical mode of failure (the product) may obscure wider issues. The questions for the Kutai Kartanegara collapse therefore include: Were the designers, original contractors, and maintenance contractors competent? Did they understand the structural behaviour of the bridge, and how the anchorage movement had altered this behaviour? How was competence ensured - was it a condition of the procurement documents, and how was competence tested, monitored or audited? What was the technical approvals process - was there an independent body scrutinising the original design and construction proposals, or scrutinising the maintenance methodology, as would be the case in the UK? Were the permanent or temporary proposals adequately checked?
The use of brittle components, the odd jacking sequence, the seeming failure to properly consider the altered structural behaviour, these are all symptoms of inadequate people and processes. It is possible that the design was entirely satisfactory, only compromised by deliberate or accidental "errors" in construction, but these still point to issues with supervision or approval, not simply with poor materials. There are few, if any, structural failures which can be attributed to a single cause - factors of safety in design, and safeguards of quality in construction, are normally such that a number of separate errors have to coincide to cause failure, particularly disproportionate collapse as happened in Indonesia.
I imagine some engineers have the "it couldn't happen here" mentality. Does an incident in the developing world really hold lessons for the developed world? I've read several comments from Indonesian sources suggesting their initial reaction is to attribute the blame to corruption or incompetence, at a level which you may not expect to find in countries like the UK. But there are plenty of first-world examples where an unanticipated combination of errors in design and construction lead have led to structural failure.
In my own experience, I can think of cases where temporary works and operations have not had the same level of care devoted to them that the design of permanent works receives. I have also seen cases where poor materials or construction errors have either been deliberately introduced on site or deliberately covered up.
I think a key issue is what the incentives are on a project which might encourage or discourage safe practice. These generally relate to procurement methods, and I have to say that my experience is that safety is not an issue procurement specialists think about, certainly not in the context of incentives, which are generally devoted to programme (liquidated damages for delay) or cost (pain/gain sharing). It strikes me as obvious to anyone other than a procurement specialist that such incentives must be a positive disincentive to enhance safety (or quality), as safety is positively related to taking time to be careful, and to spending money on training, checking and supervision. The increasingly common view seems to be that the only incentive to ensure satisfactory quality and safety is the fear of civil or criminal prosecution.
It's not that long ago that these were understood to not always be effective as incentives - the degree of fear present relates to the perception of risk, and structural collapse is generally seen as such a remote risk that the incentive to prioritise actions against it may be weak. As a result, clients were in the past prepared to pay to put additional safeguards in place: independent checking of permanent and temporary works designs, and independent inspection and auditing of construction quality. These safeguards have been steadily eroded, with increased reliance on a contractor's self-certification of their own compliance with quality control requirements.
The Construction Design and Management Regulations (CDM) were one attempt to make sure that all parties in construction recognised their role in ensuring safety. However, in practice, they seem to have deflected attention away from the parties responsible for most construction accidents (contractors and their employees), added to the burdens on designers (the case that CDM in design has saved many lives seems at best unconvincing), and been seen by clients often as an unwelcome and unnecessary cost. However, it is the clients, through their approach to procurement, who can have the greatest opportunity to incentivise safety. Imagine a world where there were contract penalties for hours lost due to accidents - would this be more or less effective than relying only on the threat of prosecution for serious safety lapses? Would an Indonesian CDM system have prevented the Kutai Kartanegara collapse? I doubt it.
Another missed opportunity to improve the role of people and processes in structural safety came with the introduction of the Eurocodes - not just the structural design standards, but also the accompanying European execution standards ("execution" here means "construction", for anyone not yet versed in Eurospeak). These standards say quite a bit about how requirements for competence, supervision, checking and quality control can all be varied to suit the probability and consequences of failure. In general, they imply the adoption of more onerous procedures for more complex or significant structures. In practice, the design and execution standards are not properly tied together, and implementation of both by national authorities has only muddied the waters. What could have been a great opportunity to focus attention on the role of design and construction management in enhancing safety, and to make that more prominent in the procurement process, is being missed.
The technical specifics of the Kutai Kartanegara failure are unlikely to be repeated in the UK or similar countries. We are, however, yet to learn what really went wrong, particularly in regard to the people and process issues. I suspect that when we do, we should not be in too great a hurry to assume that some of the same problems "could not happen here".