West Gate Bridge Collapse
Hi and welcome to Failurology; a podcast about engineering failures. I’m your host, Nicole, and I’m from Calgary, Alberta.
In Alberta, the path to being a professional engineer can vary. Previously, the only way you could stamp was with an engineering degree, four years experience and to pass an ethics exam. About a decade ago, the Association of Science and Engineering Technology Professionals of Alberta, also known as ASET, and the Association of Professional Engineers and Geoscientists of Alberta or APEGA, got together and new designations were formed. Through ASET, you can apply to be a Professional Technologist of Engineering and through APEGA you can apply to be a Professional Licensee of Engineering. Both require transcripts from an engineering diploma at an accredited school, a minimum of 6 years of experience and a passing grade on an ethics exam. As part of the application process you must define and submit documentation to support your scope of work within your field. Once accepted, you are only allowed to practice within that define scope. That said, if you wish to expand your scope in the future, you can always re-apply.
I am mostly sharing this information to show that there are more than one ways to stamp designs in Alberta and you should explore all options and pick a path that works best for your skill set and learning style.
But I also want to lowkey humble brag that I received my Professional Technologist of Engineering designation last week. I was, and still am, so excited to have received that email; I could barely hold myself together. For years I went back and forth on whether I wanted to apply. Since I work more on the construction side of engineering, I don’t necessarily NEED to stamp drawings. But over the years, I’ve become more involved in reports and mechanical system assessments and it will be really nice to be able to stamp those going forward. Anyway, this is a huge milestone for me in my career, and I just wanted to share it with all of you. Don’t let anyone crush your dreams, if you want it, go get it!
This week’s engineering news is all about heat recovery from wastewater. Yes, you heard that right, wastewater. Seems kind of weird right? I mean, wastewater? I didn’t even realize this was possible, let alone so interesting.
The National Western Center just north of downtown Denver Colorado, will be a urban hub for food and agriculture discovery. And it’s expected to be complete in 2024. The campus has a sustainable design, saving energy in two forms. First, the site has a centrally located district energy plant, meaning that the entire campus is fed from one heating and cooling plant which distributes hot and cold water, depending on the season, to each building via a series of underground pipes. We have a similar system in Calgary that is used to heat City Hall, Telus Sky, and Bow Valley College, just to name a few. Secondly, to create a more efficient district energy plant, they’ll use outgoing wastewater to either pre-heat the heating system, or reject heat from the cooling system. The combination of these two designs will allow the 250 acre campus to avoid emitting 2,600 metric tons of carbon every year.
So how does one use waste water for heating and cooling plans. Well, after the water from showers, sinks, tubs, dishwashers, washing machines and toilets flows down the drain, it all collects into underground piping system before leaving site. Don’t worry, the waste water never interacts with the clean water. They each pass through a heat exchanger which allows the wastewater to transfer heat to the clean water. What about solids you ask, they are diverted around the heat exchanger. Fun fact; even toilet water, which is fed from the cold water system, because it often sits for long periods between flushing is traditionally somewhere near room temperature when it leaves the building.
In heating applications, the heat from the waste water is transferred to the heating system. And as I mentioned in the Channel Tunnel episode, cooling is not the addition of cold, it’s the absence of heat. Most of the challenge in cooling design, is where do you reject that heat. You can use a similar concept to the heating design and transfer the excess heat from the cooling return water into the wastewater.
It might not be the Greatest Outdoor Show on Earth, but this is still a really cool project. Just a little Stampede joke for you Calgarians listening. I can’t wait to check out the National Western Center once we get back to whatever normal is going to look like. Hopefully they offers tours of the district energy plant and wastewater heat recovery facility.
This is just one example of a heat recovery from wastewater application. This could be scaled up to towns and cities, or scaled down to apartment buildings or even houses.
If you want to read more on the National Western Center, check out the link in the show notes.
Now on to this week’s engineering failure; the West Gate Bridge Collapse in Melbourne Australia in 1970. This is a tragic story, that scarred Melbourne for decades.
The west gate bridge crosses the Yarra River, just north of where it opens into Port Phillip Bay. The main bridge span across the river is 336m, and the deck is 58m above the water. Including the pre-stressed concrete approaches on the east and west sides of the river, and the steel box girder, cable stayed span across the river, the total length of the bridge is 2,582m. The west gate bridge is the fifth longest bridge and the highest road deck of any bridge in Australia. The bridge provides a vital link between the central business district and western suburbs with the industrial suburbs and city of Geelong to the west. Its also one of the busiest road corridors in Australia. While it started out with eight lanes, today it carries 10 lanes of traffic, 5 inbound and 5 outbound, servicing up to 200,000 vehicles per day. Even with 10 lanes, its often congested during morning and afternoon commuting peaks.
The west gate bridge opened on Nov 15, 1978, after 10 years of construction and a cost of $202 million. The bridge is often windswept due to a lack of obstructions. In the past, they have used wind warning lights - amber meaning the bridge was closed to motorbikes and high vehicles, and red meaning it was closed to all vehicles. But the current practice is to reduce speed limit to 40kph when wind speeds get too high. The bridge was designed by Freeman Fox & Partners or FF&P. World Services and Construction (WSC) was the bridge erection contractor who started the bridge construction, but in 1970 John Holland Construction (JHC) took over erection of the bridge while WSC continued to carry out sub-assembly of the boxes.
The pre-stressed concrete bridge approaches on either side of the river are supported on slender concrete columns, spaced 67m apart, which lead up to the steel bridge. There are five spans of the steel bridge across the river between pier 10 on the west side and pier 15 on the east side. Of the 5 steel spans, the two outermost spans are 112m each in length, the two interior spans are 144m each in length, and the middle span is 336m long. The middle span and the spans on either side of it are partly supported by two 45m tall cable stayed towers on top of piers 12 and 13. The cables run in a vertical plane down the centre line of the bridge, between the two directions of traffic. The steel spans are made up of several steel trapezoid shaped boxes, called girders. Several boxes make up a span, and the box girders were pre-assembled on the ground, before being lifted into place. In cross section, the girders are 4m deep, 25m wide at the top, and 19m wide at the bottom. They have 3m cantilever brackets which extend the deck width to 37m.
The west gate bridge was constructed using a very unusual method of construction, which had not been attempted before, anywhere in the world, under similar conditions. There are three common ways to erect a bridge, and then there’s the West Gate way. Each way has it’s own advantages and disadvantages. I don't think there is really a one size fits all method; I think you have to look at the bridge design, site conditions, as well as the expertise of the construction personnel to determine the best method for each bridge.
So what are the common ways to erect a bridge? First, you can erect all of the boxes or half boxes and connect them together in the air. This method requires a significant amount of temporary supports, several of which, would be in the water.
Another option is to use a traveling crane which travels past the ends of the ongoing spans; this is referred to as the cantilever method. With this method, the installed segments have to be strong enough so self support the bridge deck until you get to the next pier. And you also have to tilt the span slightly upwards so the deflection of the cantilever is still at the right height at the next pier. Third, you can assemble the whole span on the ground and lift into place. This method sounds the simplest. You have more control over the assembly environment and less temporary supports. But you have a much more complicated and heavy lift to bring the entire span into place on top of the piers at one time.
And then there is the method chosen for the west gate bridge. They assembled each span in two halves on the ground. Half meaning, full length, half width; so the full span from one pier to the next, but only one traffic direction at a time. Then they lifted each half span by jacking straps to the top of the pier, landed them on a rolling beam and moved them across into position. The plan was to correct any difference in vertical deflection, which is referred to as camber, between the two spans in the air and pull them together horizontally. Boy, did that plan go horrible awry.
Compared to cantilever method, the west gate method allowed much more of the assembly to be done on the ground. Comparted to lifting full span in one lift, splitting the span in half resulted in lighter lifts, although there were twice as many. The half span lift method also had its disadvantages. It was critical to accurately assemble the spans on the ground to limit camber correction required once the spans were in the air. And more importantly, the half spans were barely strong enough to self support on the pier without the other half to stiffen and strengthen the structural members. Even with the temporary stiffening and bracing to prevent the flanges from buckling, it wasn’t enough. Having never built a bridge myself, I am certainly not an expert on which method is best, but the West Gate method seems like a poor choice and not very well thought out. I will say, hindsight is 20/20. And the weaknesses of this method were relatively unknown before the collapse. But from what I've read, there appears to be a lack of oversight on the structural engineer to check and run additional calculations to support this erection method. Had they realize the half spans were too weak, they could have directed the contractor to use a different method, or outlined the required stiffeners to strengthen the half span.
That said, and I will get into this more later on in the episode, the limitations of the box girder bridges was not just an unknown here, but around the world as well. 5 steel box girder bridges collapsed in a 4 year period between 1969 and 1973. And the west gate bridge was the most deadly collapse of those five. After learning of the first collapse, the engineers did make adjustments to the bridge design to add stiffening, but it just wasn't enough.
The bridge section that collapsed was between piers 10 & 11 on the west side of the bridge. But before I can tell you about that span, I need to give you some context. Because span 10-11 was the second span they erected. They started on the east side of the bridge with span 14-15. I should also mention that when I refer to span 10-11 or span 14-15, those numbers are the piers numbers and I am referring to the bridge span between those piers.
The north half of span 14-15 was the first of eight half boxes to be lifted. Once they got it into the air, they noticed significant compression and buckling on the upper plates. As I mentioned, the half spans were not designed to be self supporting without the other half to stiffen. So when they lifted the span into place, the buckle was as much as 375mm. Bracing had been provided for handling and transport, but it was inadequate once the half span was in place in the air. I don’t understand how this wasn’t an obvious flaw in the plan to those erecting the bridge. They should have lowered the span onto supports to remove the buckles and install adequate bracing, but they tried to fix it in the air. The engineers were on site at this time, and due to some errors in assumptions and arithmetic, they incorrectly assessed the situation and didn’t stop the lift by the contractor. The long and the short of it is, they didn't have any idea of the stresses actually placed on the span.
So with the north half of span 14-15, the south half was next. They learned from their mistakes in the first lift, sort of. And they added diagonal braces to the south span and the buckling was much less than the north half. Once up in the air the north span had a 75mm greater camber then the south, that they had to remove before they could join the two halves together. Using hydraulic jacks they raised one end of the south span by 175mm, which allowed them to make some of the bolt connections, and then used localized jacks to join the rest. One section, known as box five, posed some problems. By removing some bolts in the adjacent boxes, they were eventually able to relax the buckle, but not without putting added stress on other components in doing so. Ultimately they were able to connect spans 14-15 together. And because of this, they were given a false, misplaced confidence that they could repeat this on the west side at spans 10-11. Boy were they wrong.
You'd think after the experience on the east end at spans 14-15, they would have changed course and bolted the north and south spans together on the ground and lifted the whole span at once. But noooooooo.... why would you do that?!? Get out of here with your logic!!
Some camber difference is ok, less than 25mm is within reason. It would have allowed them to line up the sections, without the use of jacks. When they constructed each span on the ground, they built the north and south spans side by side, in their final orientation, on support structures called trestles, in an effort match the camber of both spans. But these spans were huge and it was nearly impossible to control the environment with such detail. Fun fact, that I learned while researching this bridge. Bridges and large structures can expand or contract significantly from environmental conditions. For example a hot day that turns to cooler temperatures at night. Being a relatively small human, I always consider concrete and steel to be rigid objects, but they are malleable to an extent, under the right conditions. Rather than installing temporary bracing, the contractors at the West Gate bridge used a series of interconnected jacks on a common pressure line, at the underside of the boxes to distribute equal load. Now, they only intended to use this jack system when the temperature conditions of the steelwork were not uniform. Unfortunately, the jack system ended up being used a lot more than expected; meaning most of the time the box connections were being made. JHC took over erection of the half spans from WSC, during construction of spans 10-11. And they didn't fully understand or appreciate the intent of the jack system and attempted to use them constantly, regardless of the temperature differences. This placed a greater importance on the perfection of fabrication. It was already important, but now it was imperative. In the worst case scenario, the camber between the two spans could be up to 400mm out; and although this was an extremely small risk, 75mm was a reasonable outcome. The importance of getting two cambers the same was so vital that higher importance should have been placed on the assembly method. And it wasn't until both spans were lifted onto the piers that actual camber could be measured.
Despite some warnings, the errors at spans 10-11 were worse than on the east side. Once the north and south halves of span 10-11 were in place, a 114mm camber existed. This was too much for the jacks to handle, so they couldn't repeat what they’d done on the previous side of the bridge. Here's were this already bad plan went completely off the rails.
JHC proposed to use 10 kentledge concrete blocks to weigh or push down the north span relative to it’s south counterpart. Kentledge blocks remind me of jersey barriers, except kentledge blocks are cube shaped. The engineer was “reluctant” it would work, but didn’t object. And no calculations were done and no one offsite appears to have been contacted. I’m not making this up, and I see how looking back and having the full story, this flaw is obvious. But also, lets just talk about this for a second. They used an erection method that hadn't been done before, the half spans were already proving to be a struggle to install, and when the camber was worse on the second span, they tried to push down the higher section with 8 ton blocks. While it is a.... creative solution. It proved to be a really bad one.
By sat sept 5th, 1970, seven blocks were placed over box four and the west half of box 5 of the north half of span 10-11. On wed sept 9th one of the engineers wrote in his site log that he witnessed an “obvious overstress due to concrete kentledge”.
A buckle had formed at the joint between boxes 4&5 on span 10-11, which was a clear indication that partial failure of the structure had occurred. There were a few design factors that led to the buckle. A low margin of safety was used in the bridge design. And the details of the splice plates was poor. Also, the temporary diagonal brace might have also been a factor; either being inadequate or poorly placed. The buckle was so bad, it resulted in plastic yielding of the plates at the joint between boxes 4-5. Plastic yielding is where a material deforms so badly that it doesn't return to its initial state once the load is removed. To relax the buckling they had to loosen the bolts at joint 4-5 in groups of six or eight. Then the buckling and plates were inspected by the engineers before the next grouping of bolts could be loosened.
At 830am on oct 15th, after 16 bolts were removed, there was significant slipping of the plates and the remaining bolts were tightly jammed. The engineers recommended that the bolts be tightened with an air gun until they broke. The shock reaction of the bolts failing in tension dislodged the broken pieces and cleared the holes. This is actually a really interesting solution and I’m going to have to keep it in mind in case I come across any real life applications. After 37 bolts were removed, the camber had reduced from 114mm to 28mm. But the sliding movement was said to be so great, that some of the bolt holes were blind. Meaning that when looking through the hole in the front plate, you couldn't see the hole in the back plate. Although examination of the wreck shows this to be an exaggeration.
At this point, the buckle spread and a gentle settlement of the north half span was felt. The north half span no longer had adequate strength to sustain its own weight and only its connection to the south half allowed it to survive. But the margin of safety on south half was not great enough to support both. Even though it wasn't obvious on the surface, the engineers and contractors knew there was a problem and re-bolting began with urgency. But it was too late.
At 11:50am, span 10-11 collapsed. It was a 2,000 tonne mass and the collapse was heard over 20km away. 35 construction workers, including a few of the engineers who were on site to inspect the buckling, were killed and 18 injured. To make matters worse, there were some job shacks under the bridge and they were crushed by the falling span.
The Royal Commission Investigation began almost immediately and was completed on July 14, 1971. They issued a 143 page report that is available online, and there’s a link to it in the show notes. This report was a great resource for my research, so I have already covered their findings.
There is a west gate bridge memorial, which includes 6 twisted fragments of the bridge and is located at the gardens in the engineering facility of Monash University Clayton Campus. Commemorations are held every Oct 15th since the collapse. A West Gate Bridge Memorial Park opened on the 34th anniversary in 2004.
As I mentioned earlier in the show, not understanding the structural mechanics behind the design of box girder bridges proved to be a major problem that led to 5 bridge collapses between 1969-1973. I want to touch on them quickly, I think they are an important part of the story here and offer some context into what happened at the West Gate Bridge.
In November 1969 in Vienna Austria, The Reichsbrücke buckled in three places overnight and was left hanging in the air. The bridge was constructed using the free standing cantilever method. The day before the collapse was a hot day, and the bridge expanded to be longer than it would be in its final position. Overnight as it cooled down, the top chord was put further into tension and the bottom chord further into compression, until the bridge failed and buckled in three places
On June 2, 1970, a 61m cantilever section of the Milford Haven Bridge in the UK collapsed, due to a lack of diaphragm stiffeners in the top flange of the span. It was also constructed using the free standing cantilever method. There was a misunderstanding of the complexity of the diaphragm forces at the time which ultimately led to a rewrite of the design codes around the world
On Oct 15, 1970, the West Gate bridge collapsed which you just heard about. This was the most fatal collapse of this grouping of bridges.
On Nov 10, 1971, the South Bridge in west Germany collapsed. It too was constructed using the free standing cantilever method. The cause was determined to be a lack of stiffener plates.
And lastly, there was one final bridge collapse in this grouping. This one in east Germany in 1973. It was covered up at the time, and not not much was known about the collapse until the Berlin Wall came down. Again this collapsed was caused by a lack of stiffeners.
So there you have it, the west gate bridge collapse. A combination of a bad construction methods and insufficient engineering oversight led to the bridge collapse and the loss of 35 lives. As tragic as it was, it revolutionized the construction industry in Melbourne and shifted more focus to workers rights. The collapse also led to significant changes in how box girder bridges were designed.
Engineering is the application of scientific principals in real life. I think its often overlooked, how creative engineering can be; there are often multiple ways to solve a problem. And while I definitely don’t want to stifle creativity, it is imperative that we have the calculations to ensure our designs and solutions are safe, for both those building and those using the project. As an industry, we must ensure we have a good understanding of the engineering fundamentals of our designs, and that we apply lessons from those before us. This is my intent with this podcast, to create a conversation about how and why things fail so we can one, learn from them and hopefully prevent history from repeating itself, and two I am hoping we can all get a little more comfortable being in the uncomfortable space of failure. We are all human, we all make mistakes. Learning how and why things fail is an integral part of engineering.
For transcripts, photos and sources from this week’s episode, check out the show notes or head to Failurology.ca. If you’re enjoying what you’re hearing, please rate, review and subscribe to failurology, so more people can find it. If you want to chat with me, my twitter handle is @failurology, you can email me at firstname.lastname@example.org, or you can connect with me on Linked In. Check out the show notes for links to all of these.
Thanks everyone for listening. And tune in to the next episode of Failurology where I will cover the Station Square collapse in Burnaby British Columbia. Minutes after the grand opening, the roof top parking lot collapsed into the grocery store, earning it the nickname, Cave-On-Foods. But more on that next time. Bye everyone, talk soon!