Ep 24 Eindhoven Car Park Collapse

Engineering News – Hernando de Soto Bridge (3:00)

This week`s episode is about the Eindhoven Car Park Collapse (6:55). The car park used a BubbleDeck slab method (8:09) which has been used in Europe over the last few decades. But little consideration was made for the structural integrity of the slab when they rotated the BubbleDeck panels 90 degrees (14:42). Investigation of this failure provided an extensive list of recommendations to improve safety and the responsibility of risk within the Netherlands construction sector (16:25).

Intended Orientation of Panels

Installed Orientation of Panels

Rotated 90 Degrees


Eindhoven Car Park Collapse





Hi and welcome to Failurology; a podcast about engineering failures. I’m your host, Nicole, and I’m from Calgary, Alberta.

I mentioned at the start of the last episode that I rebuilt the Failurology website. Which was a very exciting and necessary change, at least I think so. What I didn’t realize was that when I switched the domain over to the new site, the podcast RSS feed was impacted. RSS feed, which stands for really simple syndication, is how podcast apps like Spotify and Apple Podcasts access the audio files of any given podcast. When I changed the website, I inadvertently changed the RSS feed link. Some podcast apps updated to the new link automatically, and some did not. Unfortunately, I didn’t realize this right away, but I have fixed it now. Sorry, I am still learning. There is a lot that goes into making a podcast and this is a one-woman show. If you were not able to listen to Ep 23 on time, I am sorry about that. But I definitely learned some things and am better prepared to prevent things like this in the future. Again, I sorry and I hope you will forgive me. Oh geez, I think my Canadian is showing.

I also wanted to give you some information on my other podcast project centered around women in engineering. Me and a team of Canadian female engineers have been working on this project since March. Our intent is to showcase the achievements of others and inspire women to pursue and continue careers in engineering. Our first season will be all about engineering disciplines. How can we give back and grow our communities within our chosen fields? And what is possible within various disciplines? The podcast is called Glorious Ladies of Engineering or GLOE for short, and our first episode drops June 23rd which is also International Women in Engineering Day. I am probably a little early in sharing this with you, but I am so excited and I couldn’t help myself.

Also, in the same vein, Engineers Canada has a free, multi-day virtual 30 by 30 conference in June. Their 30 by 30 initiative hopes to increase the percentage of newly licensed female engineers to 30% by 2030. The current number has been 14-17% over the last few years. I talked quite a bit about the initiative in the March episodes of Failurology as part of National Engineering Month!

I am going to put a link to the Glorious Ladies of Engineering website and registration for Engineers Canada’s conference in the show notes.

This week’s engineering news is about the Hernando de Soto Bridge west of Memphis Tennessee. You may have heard about this already. If you follow me on linked in or twitter, I have posted about this bridge.

On Tuesday, May 11, the Hernando de Soto Bridge, was closed to road traffic and waterway traffic was not able to cross underneath. The bridge is part of the interstate 40 highway, connecting Memphis Tennessee to eastern Arkansas, across the Mississippi River. Sidebar, do you think it started out as Our Kansas and then over time morphed into Arkansas? The spelling related to pronunciation of Arkansas is suspect, but I digress.

Anywho, the bridge was a busy one, serving 50,000 vehicles every day. And the waterway closure held up more than 60 watercraft including over 1000 barges, but luckily the river re-opened on Friday, May 14th.

Why did they close the bridge you ask? A crack in a critical steel support beam. You can’t see me, but I say crack in air quotes. Not only is the crack huge, it is a complete fracture of the beam. The two sections are no longer connected. I don't normally include photos from engineering news on the website, but you guys have to see this. Head to failurology.ca and go to the page for this episode

While we don’t currently know how long the fracture has been there, it can been seen in a kayaker's photo from 2016, so at least 5 years. It was also shown in drone footage from 2019. Needless to say, the employee responsible for inspecting the bridge has been fired by the Arkansas Department of Transportation. While I do think this is warranted, and I certainly don’t have an intimate knowledge of infrastructure inspections in the US, or Canada for that matter, I do want to share a couple of things I learned researching the Malahide Viaduct in Episode 8. Again, there may or may not be similarities between Malahide and the Hernando de Soto, but in the case of the Malahide Viaduct, the inspectors had significantly more infrastructure to inspect than any one person can handle. And in the case of Malahide Viaduct, the inspectors were not well trained in what to look for when carrying out their inspections. That said, the Memphis bridge fracture is pretty obvious and hard to miss even with a cursory review of the bridge.

On Monday, May 17th the Tennessee Department of Transportation announced that it had selected a company to repair the bridge in two phases. Unfortunately, both phases have to be complete before the bridge can re-open to road traffic. At this time, no opening date has been scheduled.

Phase one of the repairs includes installing over 8,000 kg of steel plates on each side of the fracture to strengthen the damaged structure. And then phase two involves replacement of the damaged components; for which design is currently underway for the second phase. Phase two also includes a more detailed review of major structural components to ensure the structure is sound.

I will definitely be following this story and based on the press it's gotten, I hope the investigative report is made public so I can read it and maybe feature it on a future episode.

Fun fact about the Hernando de Soto bridge, it has been undergoing a seismic retrofitting project since 2000, to withstand a 7.7 magnitude earthquake. The bridge is within 160 km of the New Madrid Seismic Zone and serves as a major cargo route across the river. I had no idea there was a seismic zone in the middle of the US. I only knew about the west coast, which also impacts part of British Columbia.

If you want to read more about the Hernando de Soto Bridge, check out the engineering news link on the Failurology website or just google it, there’s a bunch of articles out there.

Now on to this week’s engineering failure; the Eindhoven Airport Car Park Collapse.

Eindhoven is a small city in the Netherlands, located about 90 minutes south of Amsterdam. It’s the fifth-largest city in the Netherlands with the population of the greater area being around 750,000 people. The car park that collapsed was built to service the Eindhoven airport. We call parking garages parkades in Canada. It’s interesting working with people from other countries on projects because when you say parkade, they ask what you are talking about. And googling to verify that parkade is in fact a Canadian word, I found myself absorbed in a list of other Canadian words like toque, pop, washroom, even pencil crayon.

One of the reasons I chose the Eindhoven Car Park collapse is that the floor slab was constructed using a structural method called BubbleDeck or voided biaxial slab. BubbleDeck is way more fun to say though. The second reason I chose this failure is that there are a lot of lessons to be learned from it. Some of which we have mostly gathered in Canada, but are still valuable reminders of the importance of safety and proper planning.

So what is BubbleDeck? Concrete is a heavy building material and it struggles to span long distances without intermediate supports. In fact, it typically only has a loading capacity of one-third it's own weight. There are two main ways to increase the length and strengthen concrete spans. The first is by installing rebar within the concrete to stiffen it. The depth, spacing, and the diameter of the rebar are all calculated by the structural engineer depending on the design requirements. This is an extremely common method in Canada. In fact, until learning about BubbleDeck, I thought rebar was the only method. The rebar is placed in a grid pattern on the floor deck and secured in place and to other bars using wires. Depending on the strength of the slab required, I have seen rebar installed in one or two layers, called mats. One layer is typically near the bottom and another near the top of the concrete span. Once all of the rebar is in place, the concrete is poured over top and allowed to set. Then they carry on with the next area. Not being a structural engineer, I am not a rebar expert by any means, but I have spent considerable time trying to adjust plumbing or duct penetrations around rebar or trying to avoid having to cut rebar in renovation projects.

But there is another method to make concrete span long distances, and that is by reducing the weight. This can be achieved by creating voids in the concrete. In a concrete floor slab, the top and bottom of the slab are doing most, if not all, of the heavy lifting. The upper section is in compression, the lower section is in tension and the middle is mostly neutral but adds a lot of weight. By creating air-filled voids in the concrete you can reduce the weight and volume of the concrete by 50%, while still maintaining similar strength, allowing it to span longer lengths and reducing the number of vertical supports required. This is referred to as BubbleDeck.

It was invented by Danish structural engineer Jorden Breuning in the 1990s. Traditional BubbleDeck uses spherical or elliptical balls, but boxes or donut-shaped voids have also been used. The void forms are typically made from high-density polyethylene or sometimes recycled materials and they are installed in the middle of the concrete depth in a grid-like arrangement. Waffle slabs are similar in concept, except the voids are at the bottom of the slab instead of in the middle. Fun fact, the void forms are known to float to the surface during the concrete pour so they have to be held in place. A series of reinforcement mesh, vertical lattice girders and other elements trap the voids in a “bubble cage” to make sure they stay in place and also to add some additional reinforcement to the panel.

Because the balls or void forms are often made of recyclable material, and the amount of concrete is significantly reduced, there is opportunity to qualify for sustainability programs. These programs, like LEED, Net Zero or Built Green are generally a series of categories which include several steps that can be taken to improve the efficiency, sustainability and reduction of energy during the design and construction of a project. Each step is worth points, some steps are mandatory and some steps are optional. Each project needs to achieve a set number of points in each category to achieve different levels of the program. This is a general and oversimplified explanation, but I hope you get the gist. A project using BubbleDeck might get points in one category for the reduction of concrete, points in another category for recycled content, and points in a third category for design innovation. Because BubbleDeck required less vertical supports or columns, you might even get points for better floor plans, use of space or occupant comfort. I should also mention that BubbleDeck slabs are often slimmer than traditional concrete rebar slabs, and can reduce the overall height of a building. This can offer both cost savings and potential for more sustainability program points. Mechanically speaking, we get points for reduced water usage, increased equipment efficiency, reduced heating and cooling loads due to better building envelopes (read less glass) and recovering heat from building exhaust.

While BubbleDeck can be constructed on-site, in situ, it is often precast in a mould off-site and delivered in pre-determined sizes to suit the specific structure. Which is what happened at Eindhoven. The concrete floor slab was constructed off-site in panels and then once installed in place on-site, a topping layer of concrete was installed over several panels to create a uniform finish. Because this method has been in use in Europe for over 15 years, I have to assume the precast panels are more readily available and more cost-effective. Now you don’t just go to the hardware store and buy these panels, they need to be sized and laid out to suit the project, but there are likely more BubbleDeck manufacturer’s in Europe than in Canada. That said, I did find an Atlantic manufacturer out of Dartmouth, Nova Scotia. They worked on the Balsillie School of International Affairs building in Waterloo Ontario. On that project specifically, the BubbleDeck was able to span 12m between columns creating lots of open space for classrooms. As well as cost and time saving due to a reduction in the overall building height by using the slimmer structure. They also tested out a different cooling method of installing cooling piping within the precast floor panels to create in-floor cooling. I have worked on several in-floor heating projects, but never in-floor cooling. I think it would be entirely dependent on the usage of the space. For example, in a school where everyone keeps their shoes on and there are larger areas of space to allow simpler zoning, in-floor cooling is a viable solution. In a residential complex where people are barefoot or sitting on the floor, managing occupant comfort with an in-floor cooling system also becomes an issue. As well, you wouldn't be able to run the water through the floor very cold without risk of creating condensation or occupant discomfort. If we thinking about how heat rises and cold sinks, introducing the cold at the floor level would likely not result in the perception of cooling up where the people are 1-2m off of the floor. All that said, it's not a bad method, just a different one. Pretty much every method has advantages and disadvantages, it's just a matter of what your goals are and what you have to work with.

But back to Eindhoven. The intent was for the BubbleDeck panels to span between columns and beams. So if the columns and beams were in-line running north-south, the panels would span east-west and bridge from one column to the next. However, for reasons that are unclear to me, and probably others as well, the panels were rotated 90 degrees, meaning they now spanned north-south. This placed the greatest deflection of the floor at the seams between the panels, with no ability to transfer the loads back to the columns. When the decision was made to rotate the panels, no changes or consideration was given to the seam connections to accommodate for any changes in loads or deflection.

On Saturday, May 27, 2017, one month away from completion, with an outdoor temperature of 33C, the panels on the upper floor partially buckled under their own weight and collapsed, pancaking onto the floors below until the entire section of the building had collapsed. While there had been workers on this floor a few hours before, there was no one on it or under it when it collapsed, and no one was hurt.

I am not familiar with the construction process in the Netherlands, how contracts are divided and how engineering site reviews are completed. But, even my mechanical engineering self can see that rotating the panels was a bad idea. And I’m unclear how this change passed through the precast panel manufacturer, general contractor or engineering team without any red flags. During construction, there were even cracks and puddles forming throughout the slab, which indicated the potential for structural issues, but no one called them into question.

An investigation opened up immediately and some other nearby construction sites and buildings which had used the same flooring system were temporarily closed.

The Eindhoven airport and the general contractor operating out of the Netherlands, Royal BAM Group, conducted investigations into the cause of the collapse. Their claim was that the failure was caused by a poor bond between the concrete panels and the floor topping. Since this investigation was not a third party, I’m going to call it biased. Also, I don’t see how the topping bond would cause such a failure; at least not a topping I am used to seeing which offers more of an aesthetic, cohesive finish, rather than a structural one. Sure, a poor topping bond would result in flaking and cracking on the top surface, but the panels that make up the structure would still remain intact below.

Luckily for us, the Dutch Safety Board, an independent party that is free to decide which incidents they investigate, and operates on a primary focus on personal safety, completed their own investigation. Firstly, they disagreed with the Eindhoven airport and Royal BAM Group and criticized their findings. And secondly, they confirmed that it was the rotation of the BubbleDeck panels 90 degrees that caused the failure. In fact, the Dutch Safety Board even confirmed that the poor adhesion between the topping slab and the panels was a consequence of rotating the panels and causing greater deflection at the seams. It’s a wonder the panels stayed up as long as they did. They could barely support themselves and a slight load increased caused by temperature was enough to cause the house of cards to come crashing down.

In addition to these findings, the Dutch Safety Board, along with the Dutch government, recommended and implemented some fairly significant improvements. None of the recommendations were new concepts, in fact, they had been discussed each time a major construction accident occurred, but no changes had been made. It’s like when you tell someone the stove is hot 5 times and then they touch it and they're all shocked Pikachu face. Except the Dutch construction industry kept touching the hot stove, even after they got burnt. Even findings and lessons from prior investigations were not instigating substantial changes to the industry. They just keep on keeping on as if it won’t happen again. I want to say, in Calgary, in my experience, all parties, owners, consultants and contractors are pretty risk-averse. Sure stuff still happens, but most contractors are not afraid to tell you when they think something will be a problem. And you should listen to them. One because they could be right and you don’t want egg on your face. And two, because if you just blow them off, they might not come to you next time. I prefer to maintain an open dialogue with contractors. We are supposed to work together. If we both work to prevent and solve conflicts that arise on-site, we are all better for it at the end of the day. But in the Netherlands, it was more blame culture than anything. There was more focus on deflecting blame to someone else than improving safety.

The Dutch Minister of the Interior and Kingdom Relations released an informational document called Assessing the Safety of Wide Slab Floors in Existing Buildings, which included a step-by-step procedure to assess structural safety. I believe the intent of this is to ensure formal investigations are carried out as accurately and without bias as possible.

When there was a risk, it wasn’t clear who was responsible for it. As projects get larger and more companies are involved, the more division of labour and the more likelihood of mistakes. But no one was looking after the bigger picture. It’s like the project is a bus, and all parties are on it, but no one is driving. Bus drivers are pretty important in my opinion. A single party needs to be appointed to take responsibility during the construction process. This person or group would be referred to as the Process Manager and they would manage continuous and systemic risk from concept to construction by collecting and recording anything that could impact safety. And the organizing professional discussions to resolve said risk. The intent would not be to release any party from risk but to hold everyone accountable for their part in creating and mitigating risk.

As municipalities conduct less supervision over construction sites, which is happening in many jurisdictions, the construction industry has to step up to fill that role, and that wasn't happening. In fact, there was a bill in process to shift this responsibility from municipalities to the industry. When the bill was first proposed, municipalities started pulling back, but industry didn't step up because the bill hadn’t passed and they weren't legally required to. Luckily, the bill was formally adopted in 2019.

Recommendations were also made to improve the Safety in Construction Governance Code to make it more approachable and more likely to be followed. When you make a code that doesn't seem to make sense, people ignore it.

And lastly, competition in construction is nothing new. I think it pretty much universal actually. But competition that is based solely on price without consideration for safety or quality just doesn't work out. Unless all tender bids are pretty close in price, or the contract requires the lowest bidder, I don’t recommend taking the lowest. When the lowest bidder is significantly lower than the second-lowest bidder, usually they either missed something or they intend to recoup their costs elsewhere. And you know what that means. Change orders! I assume most consultants will agree that change orders take up considerable time on a project. The odd one is understandable and almost unavoidable, but when every little thing becomes a change order, it's a real pain, for everyone.

So there you have it, a simple mistake, rotating pre-cast BubbleDeck slabs brought down an entire section of the building on a hot day. Luckily the building wasn't quite complete and no one was hurt. Had this been opened to the public, the consequences could have been much more severe. I’ve said this before, and I’ll say it again, failure is an integral part of engineering. But if we are just going to ignore the mistakes and failures of those before us, how are we ever going to improve upon our work. As tragic as they are, each failure is like a little treasure of information on what not to do and how to prevent the same thing from happening again. It is paramount that this information does not fall on deaf ears. We have to do better. We can do better.

For photos, sources and transcripts from this week’s episode 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 thefailurologypodcast@gmail.com, 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 Deepwater Horizon. A deep-water offshore drilling rig that sank off the coast of Texas in 2010. They even made a movie about it, starring Marky Mark, which you might have seen. But more on that next time. Bye everyone, talk soon!