Recreating an Ancient Pump (with no moving parts)

[Note that this article is a transcript of the video embedded above.]

On the hill above Granada, Spain, sits the Alhambra: a medieval palace and fortress complex of the historic Islamic world. Built and modified over centuries, the Alhambra is now a UNESCO World Heritage site and stands as one of the best-preserved palaces in the world.

Every city needs a reliable source of water, and that stood as a challenge for the Alhambra, perched high above the nearby rivers. Medieval engineers used a lot of creative solutions to divert natural sources of water and distribute it to the cisterns, baths, and fountains within the complex. Another YouTube channel, Primal Space, has an excellent video on all the ingenious ways they managed water, and one of the details in that video really caught my imagination.

Alcazaba is the stone fortress on the western tip of the Alhambra that sits higher than most of the palace city. Apparently, throughout the Renaissance (and maybe even starting in the medieval period), the fortress was supplied by water using a pump that had no moving parts. In 1764, a priest observed the device. He couldn’t understand how it worked, but he did his best to describe it anyway. More than a century later, a Spanish engineering professor, Cáceres, took it upon himself to try and recreate the device using the priest's description. By that time, remnants of the device were gone. Historians estimate it existed until the end of the eighteenth century, when a higher canal replaced it. Even so, the professor got it to work, presenting his results at a 1911 scientific congress in Granada.

Was it the actual pump design the priest described? We’ll never know for sure, but it seemed likely to that professor, and more recent historians have found it plausible. And that’s pretty fascinating to me. A pump with no moving parts, able to lift water above its source, quietly serving a hillside fortress centuries ago. It is clever, effective, and, all these years later, mostly unknown today. You can’t pick one up off the shelf at your local hardware store, at least not yet. So I decided to take after Professor Cáceres and try to build one myself. I’m Grady, and this is Practical Engineering.

There’s something really magical about taking advantage of flowing water to accomplish work. I don’t know exactly what it is. Seeing a natural force, like the flow of a river, interacting with human ingenuity to do something important - it’s really cool to me. And it’s especially cool when it’s purely mechanical. Don’t get me wrong; I love electronics, circuits, and sensors. But doing a job with water alone - you have to admit that there’s something special about it. I’ve covered a few devices like this before. I built a trompe, which is basically a water-powered compressor. I also built a ram pump, which is a water-powered pump that uses check valves to harness kinetic energy, converting it to pressure. But I have to admit that Primal Space’s video is the first time I had ever heard of what seems to be mostly referred to nowadays as a pulser pump. And there really isn’t much information out there about them, despite the fact that they’ve been around for centuries. The idea isn’t really that complicated, but the details are a little tricky, so I decided I would try to come up with a design that boils it down as simply as possible. And you know we have to break out the acrylic.

I just had to tap the holes… then glue everything together. Now, let’s turn on the water so we can see this in action.

Step one is this basin up top. Rather than connecting directly to the hose, I wanted a free surface of water at the top, just so it’s clear, from an energy perspective, that this is the starting point. This tank provides a simple, consistent, and obvious input for the pulser pump. It’s the equivalent of the end of a canal in an ancient palace, and the goal is to raise the water above this level.

From the basin, the water falls down this vertical pipe. But if you look carefully, you can see it’s not just water. The water flows into this tee fitting that acts like a vent, allowing the stream to kind of swirl around and draw in air. There are quite a few ways to intentionally mix air and water. The historical description of the pump at the Alhambra was pretty unclear when it comes to this part. The priest didn’t provide much detail about how the air was entrained in the downward flow. Professor Cáceres tried two methods and had the most success using a whirlpool to draw water and air downward. I don’t know if this is exactly what he tried, but it is dead simple, and it worked surprisingly well. You can see the water in the pipe is full of bubbles, and it’s moving fast enough to carry them into the next tank.

The goal in this area is to separate all the air from the water. You can see the bubble float upward while most of the water continues onward. The sloped top helps trap the bubbles, so the flow exiting on the right is just water.

So far, this is basically just a trompe. I mentioned I built one of these before in my backyard and made a video about it. It looks a little different from this one, but the concept is basically the same. Entrain bubbles of air in a stream of water, carry them downward, and then separate them out - now under pressure - so the air can be used for things like smelting, powering tools, or in my case, blowing some dry grass around. It was just a scale demonstration.

Trompes aren’t used much these days. It’s easier to buy a compressor than to build a piece of infrastructure. But it’s still a cool idea, and their use is being explored to aerate remote pools of mine waste to speed up the bacterial reactions that can help clean up contamination. There are probably quite a few edge cases where a source of pressurized air is more valuable than a source of moving water, and a trompe basically lets you make that trade with no moving parts or electricity.

The sloped top helps trap the bubbles, so the flow exiting on the right is just water. You can see in my model, there’s a riser on the right, just like with the trompe demo. The purpose of this is to create enough pressure to encourage the bubbles upward. You can imagine if there was no back pressure on the system and I just let the water out at the bottom of the separator, eventually it would just fill up with air. That’s not what we want. So the water has to flow up the riser and then out through this hose, keeping the bubbles under pressure so that they’ll flow out of this tube: the discharge line for the pump.

I tried all this in my garage first, but kept spraying the ceiling, so I eventually decided to do this outside. My discharge line runs up above the inlet tank. As bubbles move into the separator, they float upward and out of this pipe. But, because the pipe is pretty narrow, water gets kind of trapped between the bubbles. This is a little finnicky, but basically, the buoyancy of the air mixed with the water occasionally creates enough lift for the water to make it all the way to the top. And now you can see why they call this a pulser pump. You don’t get a very continuous flow. But look at that! The water is actually going a lot higher than where it started in the upper tank. We are moving water uphill with no moving parts.

Actually this part of the pump is a pretty ubiquitous design. It’s usually called an air lift pump. Basically, pump air bubbles to the bottom of a pipe, and let them carry water upward. These are often used in dirty situations where you don’t want sand, grit, or plant matter clogging up the impeller of a more traditional pump. They’re not very efficient, but useful in certain situations like wastewater plants and dredges. And, this is also how coffee percolators work. The steam bubbles carry the liquid water to the top where they can percolate downward through the grounds.

I’m recirculating the water in this demo just using a bucket and pump below the table, and that drives home a couple of key points here. For one, all the water running through the pump does not actually get pumped. In fact, in my little demo here, I didn’t actually measure it, but I’d guess the discharge flow rate is somewhere less than five percent of the total flow rate through the pump. You need a lot of water to move just a little bit upward. So for two, this is not a free energy device in the same way a hydropower turbine isn’t producing free energy. In a practical sense, the pulser pump is extracting energy from the flowing water to push water bubbles downward, temporarily storing the energy. Then it’s extracted again to push some of that water back up.

So it really is that simple. A pulser pump is basically a combination of two steps: a trompe to supply the bubbles, and an air lift pump that uses those bubbles to carry water upward. But in some ways, it’s not simple at all. Two phase flow, where air and water move together, is pretty complex. If you thought fluid dynamics was tricky with one fluid, just try using two! You can tell just by looking at my demo that there’s not a lot of stability here. At the down tube, sometimes you get a regular stream of small bubbles, and occasionally you get one big one. At the discharge, sometimes you get regular pulses; sometimes you get big bursts. Every step of the process is just so …gurgly. There are a lot of knobs to turn here, and they all affect the system in different ways.

Let’s say you have a fixed flow rate, and a fixed amount of height between your inlet and outlet. You still have to select the diameter of your down pipe, which will affect the fluid velocity, and so how much air you can draw in. There are probably many different ways to mix the water and air that are more or less efficient, depending on the configuration. And there’s the diameter of the discharge line. A bigger pipe can move more water, but too big and the bubbles don’t crowd up enough to carry water with them. There is quite a bit of engineering guidance out there for air lift pumps, since they’re pretty widely used. Not so much for trompes, although I did find an interesting paper in the Journal of Applied Thermal Engineering. The author called them “hydraulic air compressors” and that’s actually one of the tricky parts to finding more information on devices like this. Since they’re pretty obscure, there’s not much consistency in terminology. The most I could find on pulser pumps was a few old YouTube videos and college projects. And this recent paper on the hydraulic techniques for water supply at the Alhambra doesn’t even venture a name for the device used there.

So this is still kind of just trial-and-error engineering. I’m sure I could spend hours trying different configurations and improving this demonstration. If you’re a grad student looking for a thesis idea, I think pulser pumps would make a pretty interesting project, because I can see some applications here. In fact, I’m not the only one.

Hydraulic ram pumps are pretty popular around the internet and in rural areas that have abundant water but no electricity. They were well known by the time Professor Cáceres did his experiment in 1911. In his paper, he said about the pulser pump:

“This arrangement will always have, over the hydraulic ram, the advantage of eliminating valves entirely, since it contains no moving solid parts. Doing away with the ram strokes seems to remove any source of fatigue in the pipes and, of course, the very annoying noise that makes the ram inapplicable near living quarters.”

I can’t help but think back to him in his lab, seeing the water spurt out from the top of the discharge line for the first time. You can tell his excitement in the paper:

“Beyond its historical appeal, the idea has real value for modern engineering. In cases where efficiency is not critical, reviving it could solve practical problems, using a layout so simple that it is remarkable it has not become common knowledge after several centuries.”

I wonder if he would be a little disappointed that the idea never really did catch on, despite its novelty. But I still think it’s pretty cool. And maybe someone will see my demo working and try it for themselves, carrying the ancient idea forward for new applications. Thank you for watching, and let me know what you think!

Sawing a Dam in Half (on Purpose)

[Note that this article is a transcript of the video embedded above.]

Concrete is the second-most-consumed substance on our planet. Only water beats it, and actually water is a major ingredient of concrete anyway. Every year, humanity mines, mixes, and places roughly three metric tons for every person on Earth. It’s ubiquitous. Most of us hardly even think about all the concrete around us. We’ve all seen the grey lumpy mixture flowing down chutes into formwork to become a road, sidewalk, footing, pile, patio, or foundation. It’s easy to think of concrete as a single, uniform substance used around the world. But it’s not.

The only reason we are able to use so much concrete in construction is that it’s cheap. Of the four main ingredients - sand, gravel, cement, and water - two of them come directly from the ground with little need for processing or refinement. One is water. Cement is the only ingredient that requires a significant manufacturing process, but the raw materials for it are fairly widespread across the globe. Many building materials are constrained by geography. They only grow, occur mineralologically, or are manufactured in specific locations. Then they have to be transported, often at great cost, to where they’re needed. It’s not true for concrete. No matter where you are on earth, there’s a pretty decent chance that somewhere nearby exists a ready source for at least most of the raw ingredients you need to make it. That simple fact has significantly contributed to its widespread use, but it’s done something else too.

Take a look at any geologic map. If you’re like me, you do this in your spare time anyway. You realize pretty quickly that there is tremendous variability in the different kinds of materials that make up the surface of Earth’s crust. And the practical result of that, at least for the purposes of this discussion, is that every batch of concrete is just a little bit different depending on where you go. In a way, that’s kind of special, right? In most cases, the concrete you see around you represents a particular place on Earth. Its strength, durability, appearance, and essence are highly local characteristics. It’s literally made from materials that were sourced not too far away. But, in some cases, we’ve learned too late that local materials had some hidden problems when used in concrete, and the ways we’ve worked to fix those problems have created some of the most interesting stories. I’m Grady, and this is Practical Engineering.

This is Fontana Dam on the Little Tennessee River in North Carolina. At 150 meters (or nearly 500 feet) in height, it’s the tallest dam east of the Mississippi River. The north shore of Fontana Reservoir forms the border of the Great Smoky Mountains National Park. And if you’re through-hiking the Appalachian Trail, the famed 2200-mile path through the wildest parts of the eastern United States, you have to walk right over the top of it. Built by the Tennessee Valley Authority (or TVA), Fontana was completed in 1944, just in time to provide hydropower to the Alcoa aluminum smelting plant at the end of World War II. It’s a concrete gravity dam, meaning that it derives its stability to hold back Fontana Reservoir entirely from its own weight. And boy does it have a lot of weight. More than 2.1 million cubic meters of concrete went into the structure before it was finished. That’s well over half the volume of Hoover Dam, and if you watch the same kinds of videos I do, you know that putting all that concrete in Hoover Dam was a major challenge.

Concrete heats up as it hardens, which can negatively affect the curing process, but more importantly, it causes the concrete to expand. For a structure like a dam sandwiched between two rocky abutments, that expansion can lead compressive stress to build up in the concrete. Then, after curing, when the concrete starts to cool back down, it shrinks. That shrinking can lead to cracks, especially in mass concrete structures that heat up and cool down unevenly. And cracks are not ideal for dams. To mitigate this issue, pipes were installed within the concrete at Hoover Dam, and chilled water was continuously circulated during construction to pull heat out of the concrete. The same thing was done when they were building Fontana Dam. In fact, in addition to the cooling lines, the dam was built with deliberate expansion joints that would allow each separate concrete block to cool off and shrink. Once the concrete cured, those joints were grouted to add strength and make the dam watertight. It was a pretty robust and thoughtful plan to avoid the buildup of stress in the structure, or so they thought.

In 1972, engineers inspecting the drainage gallery, a tunnel through the concrete dam used to collect and redirect drainage, noticed unexpected cracks right where the dam curves. Later investigation revealed that the cracks extended through a large part of the structure. At this point, the dam was still less than 30 years old. It shouldn’t be deteriorating this quickly. But the cracks were serious enough that something needed to be done.

Engineers initially blamed the Tennessee sun. Fontana Dam runs almost perfectly east to west, with its broad downstream facing directly south. That means a huge area of concrete is exposed to sunlight for most of the day. The sun heats the concrete, causing it to expand, and over thousands of cycles, cracks are inevitable. The curved section of the dam was most vulnerable. Reaction forces from the abutments align with the axes of the dam. Instead of pure compressive stress, the expansion of the concrete created bending stress (a combination of expansion and contraction) at the corner. In addition to the cracks, the movement was also causing the spillway gates to bind up.

After instruments were installed on the dam, the scope of the problem became clear. Thermal movement is cyclical with the seasons. Concrete may expand in the summer, but it returns to its original size in the winter as temperatures cool. Fontana had some of that, but underneath the cyclical changes was a continuous one. The concrete was permanently growing.

TVA took some cores of the concrete to start planning a repair, and sent them out for testing. When the results came back, the reason for the unexpected growth was discovered. The laboratory that examined the concrete under the microscope noticed that some of the aggregates inside had dark rims around them. That is a classic sign of alkali-silica reaction, or ASR, sometimes known as concrete cancer.

The fundamental components of concrete are aggregates, large and small, bound together by a paste of cement and water. As the cement paste hydrates, potassium and sodium hydroxides dissolve into the water within the tiny pore spaces of the concrete, creating an alkaline solution. In some cases, this is a good thing. The alkaline environment is great for steel reinforcement, helping to prevent rust. But for some types of aggregates, it causes a serious problem. Specifically, if reactive forms of silica are present, they can more readily dissolve in the high-pH water, combining with the alkalis to form a kind of gel. As that gel absorbs moisture, it swells and expands, causing internal stress and cracking.

This is an extremely widespread problem that has caused structural damage in every state in the US and many countries around the world. You usually don’t have to search far for an example of a cracked up bridge, broken sidewalk, or ruined building foundation that resulted from an alkali-silica reaction in the concrete. Fortunately, the reaction requires three conditions, so there are quite a few ways to deal with it.

For one, an alkali-silica reaction requires the aggregates to actually contain silica, also known as silicon dioxide. Well, 90 percent of the Earth’s crust is made up of silicate minerals, so this might not seem possible to avoid. Luckily, only certain forms of silica are significantly reactive in concrete. We have tests we can perform ahead of time to identify quarries or sources of rock that react with cement, allowing us to just avoid the issue altogether. But like I mentioned before, the cost of concrete is really sensitive to transportation costs. The farther you have to go to get suitable aggregates, the higher the project’s costs rise, so avoiding local materials is not always ideal.

The second condition required for an alkali-silica reaction is highly alkaline cement. So, we have ways to control for that too. Cement can be manufactured to have lower alkali content, and we can use what are called “Supplementary Cementitious Materials,” like fly ash, to replace some of the cement in concrete. Those solutions only work if the concrete isn’t already in place, though.

The third factor of an alkali-silica reaction is excess moisture. You can just keep the concrete dry with waterproof coatings or membranes. Without moisture, the gel can’t expand, so the problem is solved. But there are some structures where waterproofing is a pretty big challenge. So TVA was in a bind, literally. They were facing the possibility of just having to perpetually repair cracks and equipment as Fontana continued to expand. Then they decided to get creative. Kristen Smith is the Senior Program Manager for Dam Safety at TVA, and she explained the thought process:

Kristen: “You know, the impacts on the spillway and powerhouse equipment. That led to major maintenance and repairs [...] Need to move from the reactive approach - that's not a long-term solution - to a more proactive approach.”

The proactive approach they landed on was a fourth option for dealing with ASR: Rather than trying to stop the reaction, TVA decided to just give the concrete more room to grow. The solid rock abutments at each end of the dam had no room to give, so that space would have to be found in the dam itself.

In 1976, they embarked on a fairly novel operation to cut a relief slot all the way through Fontana Dam and do it without draining the reservoir or causing any disruptions to the hydropower plant. The idea was pretty simple: instead of building up axial stress as the concrete expands, the dam can expand into the newly cut slots. Simple in theory; pretty challenging in practice. How do you saw a dam in half? Luckily, TVA has done this at two of its other dams in addition to Fontana, and shared some footage of that so you could see it happen.

These are big dams, so this isn’t sawing with blades you find at ahardware store. The tool used for cutting through the concrete looks more like a rope than a saw blade.

Kristen: “It is diamond wire, and it's really neat. It's, if you touch it, you know, it's 15 millimeters, which is a little over half an inch. It's abrasive. I mean, you know, it would rub your skin if you drug it across your skin, but you can touch it. You can run your hand along it and it's not going to cut you. It can cut through concrete. It can cut through steel. [...] It looks like a big necklace.”

That big diamond necklace runs along pulleys strategically installed on the dam to advance the slot downward. The saw pulls the wire in a loop, managing the slack and keeping constant tension against the bottom of the slot. There are a lot of advantages to this, in addition to the practically unlimited depth. It causes very little vibration or dust, and provides a clean cut without breaking the edges. But, there’s a pretty obvious challenge of cutting a slot in a dam: how do you deal with the water?

Turns out, it depends on the dam. At Fontana, crews installed a cofferdam on the upstream face of the dam to hold back the reservoir during the operations. It’s basically half of a steel pipe that seals against the concrete face on the sides and bottom, just big enough for access to adjust the pulleys. At Chickamauga Dam, the geometry made a cofferdam less feasible. So instead, they broke the process up into three sections separated by boreholes drilled downward into the structure. One section could be cut by the diamond wire while the other borehole was sealed, preventing water from moving through the slot. That’s easier said than done, but you can look to your feet for inspiration. The seals installed in the boreholes are long rubber tubes called sock seals.

Kristen: “Well, it's like a sock you put on your foot, but a half-inch thick rubber and a hundred feet long. And I've heard it described as kind of like an inside-out fire hose. Very strong and waterproof, but to some degree flexible.”

The mess is another problem. The dust from the fresh cut concrete mixes with lubricating water to form a slurry that runs out of the slot. Concrete slurry isn’t good for the environment. It mucks up the water and changes the chemistry. So the slurry generated by the cutting process has to be captured and pumped to holding tanks. After the concrete particles have settled out, the water can be recirculated to control the dust and lubricate the wire as it cuts. And this whole process happens essentially non-stop. Time is of the essence so that the internal stress doesn’t close the slot while the wire is still inside it. Slot cutting is relatively low impact on the dam operations, but parts of the dam have to be shut down to avoid an accident like a broken wire being pulled into a hydro unit or spillway gate.

One of the reasons this is possible at all is that TVA’s concrete dams experiencing ASR are all gravity dams. In essence, that means that any vertical slice of the dam is theoretically stable on its own without lateral support. Cutting a slot in an arch dam wouldn’t work, because they depend on axial thrust forces for stability.

Before, during, and after the slot cutting operation, there’s an intensive monitoring program to keep an eye on how the dam is behaving and methodically measure the movement and strains to make sure the dam responds in the way the engineers predict.

Kristen: “We have hundreds and hundreds of instruments on the concrete portion of the dam. We measure the slot that we've cut. Is it closing? Is it opening? At what rate is it closing or opening? We measure our spillway piers. Are they moving? We measure expansion joints. Everything in every direction we measure.”

And those measurements are important because the slot cutting isn’t a one-time permanent solution. This doesn’t slow down the alkali-silica reaction in the concrete at all. It just mitigates the stress building up in the structure as the concrete expands, which is basically a non-stop process. Over time, the slots close. That means that TVA has to go through the operation regularly.

Kristen: “Every approximately five years, we update, we use finite element analysis models on our concrete growth projects. So they take all of those years of new information data from the instruments and they recalibrate and they rerun these models and they can tell us how effective the slot cut is. They can tell us when we need to do it again. Whatever we need to do to ensure that we are maintaining the integrity of our damms and the adjacent equipment, that's what we do.”

I was curious why they don’t just cut a big slot to get a longer period of relief before having to do it again. In hindsight, it was kind of a dumb question:

Kristen: “The simple answer is so we don’t leave a big hole in the dam. The slot cut at Chickamauga is approximately a half and inch. It's a lot easier to stop water from flowing through a half an inch slot in a dam than it would be maybe a six inch wide slot. In addition, slot cutting is expensive.”

In other words, TVA wants to disturb their structures as little as possible, while still mitigating the problems AAR causes. It’s a back-and-forth thing. You cut, observe, wait, and only cut again when it’s necessary. It’s good stewardship of the resources available to take care of the structures we’ve already built.

Alkali-silica reaction in concrete is a huge problem. It’s something engineers have to consider when designing basically any concrete structure, which means it’s something that quarries, batch plants, testing labs, and contractors have to think about as well. Since the 1970s, we’ve gotten pretty good at avoiding it in our structures. But since it’s often a slow-growing issue, we’re still figuring out how to deal with the problems it’s causing on the stuff we built before we really had a handle on it. On mass concrete structures, like TVA’s dams, it could have been a death blow, significantly shortening the lifespans of these massive projects. But they figured out a creative solution to live with it.

Kristen: “I mean, it's cool. And when you think about a dam, it's a water barrier. It is designed to hold back water. So the last thing you expect to do is to cut a piece out of it. But we do. We do.”

Hurricane vs. Tiny House

[Note that this article is a transcript of the video embedded above.]

By the end of this video, one of these buildings will be knocked down by the force of a simulated storm surge, because there’s a lot we still don’t understand about hurricanes and their effects on buildings.

In September 2022, Hurricane Ian tore across the Caribbean and southeastern U.S., leaving a trail of devastation from Cuba to the Carolinas. It was one of the strongest and deadliest storms in modern history. We often think of hurricanes in terms of wind and rain. But in coastal areas, it’s the surge of seawater driven inland by the storm that causes the most catastrophic damage. Homes and buildings didn’t just get wet. Many were obliterated, swept from their foundations entirely.

But unlike many storms of the past, Ian came with data, and lots of it. Today’s tools for collecting and analyzing information mean that even tragic disasters can lead to really important insights into how we can build safer and smarter in the future. After Hurricane Ian, FEMA analyzed more than a thousand flood claims, and what they found about building performance was remarkable.

To dig deeper, I’m here at O.H. Hinsdale Wave Research Labratory at Oregon State University. A team of engineers is running a one-of-a-kind experiment to simulate storm surge and study how buildings actually respond. They invited me here to see it firsthand and share what they're learning with you. I’m Grady, and this is Practical Engineering.

Everyone knows hurricanes are destructive, but storm surge often gets underestimated, not just by the public, but policymakers and planners too. The damage from high winds is visually dramatic. We see footage of roofs ripped off and trees snapping like twigs. But just a few feet of storm surge can cause even greater damage. And waves amplify the destruction.

If you’ve spent time in coastal areas, you’ve probably seen homes raised on stilts. Since the early 2000s, this has become one of the most common construction types in flood-prone coastal zones. The concept is straightforward: move the living space above the reach of storm surge. If a hurricane hits, the lower area used for parking, storage, or access might flood, but the critical parts of the building stay dry. All the devastating power of the waves flows through and around the stilts instead of slamming into walls and destroying the structure. It turns out this idea is remarkably effective.

After Hurricane Ian, FEMA found that flood insurance claims for elevated structures in Fort Myers averaged about one-third the cost of claims for non-elevated buildings. That’s a staggering difference in performance. But zoom in, and things get more complicated. On one hand, this is pretty obvious stuff. You don’t need a massive wave laboratory to figure out that elevated structures survive storm surge much better than buildings at grade. But if you look at footage from Hurricane Ian, it paints a more nuanced picture, because some elevated buildings didn’t fare well at all. They weren’t all high enough to avoid the surge. And that gets to one of the most difficult questions in the entire field of hurricane engineering: how tall is tall enough?

Needless to say, it is expensive to lose your home in a storm. The conundrum is that it’s also expensive to build your home in such a way that it can withstand one. If it were easy, every building in Fort Myers would be a hundred feet above sea level. But the reality is that elevating a structure adds significant upfront cost, and the higher you go, the higher that expense climbs. It’s not just a cost for homeowners but also something that’s passed down to renters. Shifting the actual housing upwards shifts the affordability of housing downward for everyone. And because major hurricanes are relatively rare events, the return on that investment comes with a lot of uncertainty, with benefits that are invisible most of the time.

That’s one of the biggest challenges for engineers and officials. In theory, you can design a structure that withstands anything. But in practice, no one’s building hurricane bunkers as homes. Codes and policies have to balance safety with economic viability and long-term risks with the upfront cost of resilience. Local governments want robust, resilient development, but they also need development to happen in the first place. Overly strict codes can scare off builders or price out developers. And while the National Flood Insurance Program might prefer fewer claims, stricter floodplain regulations also come with tradeoffs: reduced property tax revenue, limited housing supply, and the burden of compliance placed on individuals.

These decisions might seem kind of trivial at the scale of a single structure, but when you multiply them out along developed coastlines, the implications of each extra foot of elevation are monumental. So what you end up with is a delicate balancing act, shaped by competing priorities, enormous uncertainty, and billions of dollars on the line. Changing building codes or policies requires buy-in from a broad array of stakeholders, and that kind of consensus demands reliable data.

But there’s one more thing that makes this even more complicated. Of course, “stuff getting wet” is a problem with storm surge, but it’s more than just typical flood damage you’re dealing with when it comes to hurricanes. In a sense, the surge is a rise in sea level itself, and once your home is essentially IN the ocean, that brings wave action into play. Forces intensify. Structural systems are tested in ways that ordinary flood damage doesn’t account for.

You can see why this idea of elevating structures is one of those engineering concepts that seems obvious on the surface, but gets way more complicated when you start looking into the details. And that’s why we’re here. Computer models are limited in their capabilities. And you can’t just call up an actual hurricane to knock over a test structure (and even if you could, it would probably violate the ethics rules). So we go to the next best thing: the wave lab.

The OH Hinsdale Wave Research Laboratory is one of the largest facilities of its kind in the world. Since the 1970s, this lab has supported cutting-edge research into coastal engineering challenges like sediment movement, tsunami behavior, and wave-structure interactions. It actually has two major test beds. This is the Large Wave Flume. It’s used for all kinds of hydraulic experiments related to waves, coastal structures, and erosion. It’s basically a super-sized version of the flume I use in a lot of my garage demos. It can do a lot, but it has a limitation in that it’s inherently two-dimensional. Flow can really only move in the direction of the flume. That’s why the lab also has this: the Directional Wave Basin.

Think of it as a wave pool turned up to eleven. This enormous tank uses dozens of piston-driven paddles, each with independent control, to generate complex, multi-directional waves. You can create a single tsunami-like pulse or dial in irregular wave trains to match the chaotic sea states found in real hurricanes. This facility is utilized in large-scale research projects on wave hydrodynamics, floating structures, and devices that harness wave energy to generate electricity. But, of course, it can also test coastal structures, like these houses.

Dr. Dan Cox is a Coastal Engineer and Civil Engineering Professor at Oregon State. He explained to me why they chose the basin for this experiment.

“The nice thing about the basin is that you can look at kind of a full 3-D picture, rather than just a slice. And I think for this set of tests, we really wanted to do an entire house, not just a wall, you know, a bit of the foundation. And that’s why we chose the basin for this one.”

The research team has spent months building two incredibly detailed model homes, each one a near-perfect one-third scale replica of a real coastal house. Each foot is equivalent to three feet in real life. And the only difference (besides color) between them is elevation. The green model is a foot or 30 centimeters higher up than the orange one. That corresponds to 3 feet in the real world or roughly one meter. In every other way, both structures are identical. They’ve got interior walls, windows, framing details, everything. At this scale, that means I’m about the size of an 18-foot-tall civil engineer… which is actually something I’ve had dreams about.

One-third scale is still just a model. But this is not a toy experiment. The researchers have carefully accounted for all the physics involved. The wave periods and velocities have been adjusted to simulate full-scale conditions, and the structures have reduced stiffness to reflect the relative rigidity of real-world buildings. It’s all about maintaining dynamic similarity, a fancy term for making sure the test results actually mean something when translated back to full size. And that’s a tough thing to do:

“On the structure side, it’s a lot more difficult to scale the structural behavior. So, for example, when we’re doing computer simulations, the simulations are primarily at scale - trying to get that difference in shaking. The forces can generally be scaled up as well, so we kind of know what the forces are. But I think the mode of failure - like how this structure failed - I’m not sure so much as like a quantitative scaling. It’s a little bit more like qualitatively, this is what we would expect to happen under these conditions.”

The experimental design has the waves start small and build gradually, both in height and frequency, simulating the approach of a storm. The goal is to observe how both buildings respond as conditions get worse and worse.

It’s mesmerizing to watch: the wave generators churn, sending pulse after pulse across the basin. Within seconds, the models are surrounded by rolling water, with each wave slapping against walls, flowing around supports, and rebounding off the basin walls and shoreline.

Even now, researchers at the lab are measuring the behavior of the structures. If you look carefully, you’ll notice targets for highly specialized cameras and lidar to carefully monitor the behavior of each structure. Sensors placed throughout the experiment are recording everything—wave height, velocity, pressure on the structure, accelerations, and even internal motion. The goal is to build a detailed, physics-based understanding of how each building absorbs and transfers energy from the storm surge. And that data is incredibly valuable.

For one, this expensive and elaborate test is just two buildings. And there are a lot more types of houses in the world than that. So this data can be used to calibrate and validate computer models, making it easier for engineers to get reliable answers to questions without having to build scale buildings and put them through huge model tests like this.

And some of those questions are big ones. When you’re looking at options for large-scale flood infrastructure, a major part of the process is estimating the differences in damage and loss of life between alternatives. Again, we can’t build infrastructure, call down a hurricane, and test it out in real life, then revise accordingly. Even engineers shouldn’t have THAT kind of power. So we have to be able to make predictions about how any proposal will work out. It’s educated guessing, essentially. But the better we understand the connections between all the variables (wave height, surge level, building elevation, movement, and damage), the more educated those guesses become.

“I would say the physical model is closer to the real world. Numerical simulation is kind of the best we think we can do. But - And it always looks pretty, always looks really cool. But there’s really - you have to verify it. You really have to show that it’s correct, not just looks cool. And I think when we get to the laboratory, like we’re seeing during this test, like okay, it’s not as simple as we think. So there’s a lot more complexity, I think, inherent in a physical model.”

That’s why even though these tests seem pretty straightforward at first, they can have a profound impact on how we allocate public funds, regulate floodplains, and ultimately, keep people safe. You probably wouldn’t buy a car without giving it a test drive first; it’s too big a financial decision to take a risk. Imagine changing the building code or floodplain regulations without good data to back it up. We necessarily make high-stakes decisions about how to manage flooding in the face of equally enormous uncertainties. So, you can see why information like this would give more confidence to engineers and regulators to write building codes and improve floodplain regulations, knowing those decisions are grounded in truth.

But it’s not just about the data. You might have noticed that these houses aren’t just bare minimum structures. The team has added details like roofing, window frames, and colorful paint jobs to make them look like real buildings, even though they don’t really affect the final results. That’s because this test is also a communication tool. Most people aren’t going to read the academic papers that get published as a result of this study, but this footage tells a story.

You don’t need data to understand which of these two structures you’d want to live in when a hurricane comes. And the more people who take storm surge seriously, the better the outcomes we can expect when a big storm arrives.

Each set of waves is programmed into the machine to simulate the variability of a storm, with the upper limit of wave amplitude increasing from one set to the next. After four sets of waves (delivered in about an hour), they raise the level in the basin using this massive bathtub faucet and repeat the process. It was actually pretty surprising how well both models were holding up for a while there.

It’s hard to communicate in a video just how awe-inspiring it is when the directional wave basin starts really churning. And eventually, a particularly violent wave comes crashing into the lower house, and we see our first damage. You can see the wall underneath the window give way, and now waves start penetrating into the interior of the structure. In a real house, this would already be catastrophic damage.

But of course, they don’t stop at the first sign of damage, and the team keeps hammering the models with more intense waves. Over the course of the experiment, the sea conditions just keep getting worse and worse, and the damage to the orange house does too. More and more of the first story of the lower house is swept away. Waves flow through the structure and knock out portions of the wall on the beach side, and everybody in the room fills with eager anticipation of a total failure.

And then, something I didn’t quite expect happened. The model seemed to almost stabilize. The walls of the front and back of the structure were so totally obliterated that the first floor almost began to act like another level of stilts! Despite the first floor being utterly wrecked, the second story remained more or less fine for quite a while, even as the waves got stronger.

Dan told us about a test at half this scale (one sixth of real life scale) that had shown similar progressive damage, but that led to collapse much earlier on:

“In the previous study, we started to see the deterioration and then very quickly, rapidly, the entire building destroyed and I thought, okay, well we'll see that again at larger scale, but we didn't.”

That’s one of the cool things about moving up in scale and realism: you learn things that aren’t always expected. If we had cameras on every structure during Hurricane Ian, we likely would have seen similar results - damages from storms rarely follow a linear, progressive trend. It comes in fits and starts. For a while, it seemed like it might be the end of the experiment, since the stronger waves weren’t causing more damage.

“…It was a tough problem, and I thought I knew the answer, and it turns out I didn’t. Little bit tough to swallow, but it also kind of highlights to me, like, okay this is a challenge. This is a hard problem. So for me, you know, I’m trying to put a positive spin on it, but I feel like that’s a success right there. To say hey, this is more complicated than we thought.”

Of course, everyone watching (including me) and those participating in the experiment were hoping for that final blow that would knock the whole thing over so they could get the full range of data needed from safe to damaged to destroyed. And eventually the moment came. The waves finally won, and the lower house collapsed.

What’s probably more interesting than that is the condition of the other house. Take a look at that. Almost no damage whatsoever. This building sat in the exact same conditions as the other house and took almost no damage. And in a way, that’s kind of remarkable. Because there really wasn’t that big of a difference between the two. I said it’s expensive to elevate a structure, but the marginal cost between the green and orange models is almost negligible compared to the overall value of the structure.

“In talking to people about flood risk, you know, we talk about the 100-year, 500-year. And I think there’s a misperception that the 500-year is like 5 times bigger, 5 times worse, I have to elevate 5 times greater. And I think just trying to show people it doesn’t take much. Like, there was not much of a difference in elevation between those two buildings. The one on the right is toast. The one on the left had a little bit of damage, but hardly any, and that was only after we really tried to take the other one out.”

Researchers will be studying the data from this experiment for years to come. But the story's pretty clear. Same surge, same waves. A little difference in elevation can make a huge difference to a structure when it comes to surviving a hurricane.

You might be watching these buildings get knocked about and thinking: “We don’t need more resilient structures in the floodplain; we just need them to not be there in the first place.” And in many ways, you’d be totally right. Often, the most economical way to reduce flood damage is to avoid building in flood prone areas, or if development has already happened, simply to buy out property, tear it down, and leave the land empty as a buffer. But where’s the line between flood-prone and not, especially when it comes to rare events like hurricanes, where the probabilities of occurring in a year are in the range of 1-in-100 or 1-in-500? And if there’s not a bright line between at-risk of flooding and not, what’s appropriate for the fringe?

The truth is that there is no catch-all solution to flooding. We need options to accommodate the vast array of situations where development occurs, whether those areas are flood-prone, flood-free, or, most importantly, somewhere in the middle. And not just options, but also the data to determine which of them is truly the best path forward. Engineering is a balancing act; we need structures that are both strong and safe, but also affordable, easy to occupy, and maybe even architecturally pleasing. Using knowledge gained from tests like this helps us get a clearer definition of the edges of the problem we’re solving.

Huge thanks to Dr. Dan Cox and his team of researchers for inviting us to see this happen.

The Hidden Engineering Behind the Falkirk Wheel

[Note that this article is a transcript of the video embedded above.]

This is the Forth and Clyde Canal in central Scotland. Completed in 1790, it was the first canal to cross any part of the British Isles. There are a lot of geographical terms for coastal features where the sea indents into the land: sounds, inlets, fjords, lochs, coves, bays, and so on. They all have subtly different meanings that can vary by location, but in Scotland, a lot of them are called “firths,” and they’re pretty important when it comes to navigation. The Forth and Clyde Canal, as its name strongly suggests, connects the Firth of Forth to the Firth of Clyde. It also has a branch into the heart of Glasgow. When it was built, this canal dramatically shortened the transit times for goods in the region, and it also served as the testing waters for the very first steam-powered boats.

Not long after the Forth and Clyde opened, another important canal was completed in Scotland. The Union Canal connected the cities of Falkirk and Edinburgh, opening up a route for coal and other minerals from the mines and quarries around Lanarkshire to the capital. Along the way, it passes over some pretty impressive aqueducts, including the Avon Aqueduct near Linlithgow. A connection to the Forth and Clyde Canal in Falkirk would provide a direct waterway link between the two largest cities in the country (Edinburgh and Glasgow) without ships needing to navigate the hazardous Firth of Forth. But there is the challenge of elevation. Union Canal sits about 115 feet or 38 meters above the level of the Forth and Clyde.

Moving people and goods by boat has a lot of advantages: it’s cheap, it’s efficient, it usually takes less infrastructure, and it allows for connectivity across the globe. But there is a major disadvantage: the waterways that ships and boats traverse have to be pretty much level. Boats don’t climb hills like cars, trucks, and trains. This is the main purpose of a lock: raising or lowering a vessel to navigate elevation changes in waterways. In fact, locks are pretty much the only solution to this engineering challenge… except in a few rare cases.

You’ve seen the title. You know where I’m headed with this. But the story of the Falkirk Wheel - the only rotary boat lift in the world - is fascinating, not just because of the mechanisms, but also how it came to be in the first place. It’s not easy to accomplish projects like this. The Falkirk Wheel is not the passion project of some lone eccentric billionaire. This is public infrastructure, which means a vast array of stakeholders had to come together and agree that this bizarre structure was worth the resources that went into building it. It’s got some very clever engineering under the hood, and a lot of lessons in its creation that, I think, apply to other challenges we face today. I’m Grady, and this is Practical Engineering.

Of course, the original connection between the Forth and Clyde and Union canals did use locks. A lot of locks. This map from 1898 shows the flight of 11 locks required to get boats up and down between the two. You can imagine the time, resources, and effort involved in navigating this staircase. The process took the better part of a day, and not only that, it used a lot of water from the Union Canal. Even though boats can move through locks in both directions, water only moves through in one. Each structure always fills from the upper canal, and always drains to the lower one. That’s just gravity. But, it’s important to realize that even though most locks don’t use pumps, the energy required to raise and lower boats through isn’t free. Each passage through costs roughly one “lock-full” of water from the upper canal.

In addition to the inconvenience and water usage, other factors eventually drove these canal systems in Scotland into disrepair and abandonment. The canals were small, and as ships got larger, the narrow and shallow passages became less useful for transporting materials and goods. The railroads also started competing with the canals, offering faster connections between major cities. By 1930, the canals were barely used, and by 1960, they were choked with vegetation and debris. Motorway construction disconnected several segments, and authorities decided to close them for good. That could have been the end of the story, and honestly, it wouldn’t be too surprising. It’s been the fate of many of the world's great canals and inland waterways as transportation technologies and overland shipping have passed them by. But then the year 2000 happened.

Maybe you remember this. It was a weird time to be alive. There was this strange tension between excitement about the new century and fear that all our computer systems would crash into an apocalypse. The programmers and IT professionals took good care of us on the computer side, but there were people working hard on the celebrations, too. One of those organizations was the United Kingdom’s Millennium Commission. The idea was simple: take some of the money from the National Lottery and direct it toward interesting and impactful projects that would help mark the turn of the century.

In Scotland, a large consortium of organizations - public, private, and volunteers - got together and applied for a grant from the Millennium Commission. In 1997, funding was awarded to cover approximately half of the cost of the Millennium Link: a massive undertaking to revitalize and reopen the canals that once connected Scotland from coast to coast, restore locks, build hike-and-bike trails, and rehabilitate bridges. The work included The Kelpies, a sculpture of two huge horse heads that serve as the gateway to the Forth and Clyde canal. That was a pretty fascinating civil engineering project in its own right. But of course, the Millennium Link’s flagship project was reconnecting the Forth and Clyde to the Union Canal.

But rather than do it with locks, the group wanted a 21st-century landmark, or I guess, more of a watermark. A fast, efficient connection that would serve as a capstone to the canal revitalization, draw tourists from around the world, and serve as a symbol of the region that was once a hub of transportation and commerce in Scotland. And, in fact, a hub is a good metaphor for what they came up with. The Falkirk Wheel opened for traffic in May 2002, and now, more than two decades later, it’s pretty clear that they nailed the idea. Here’s how it works:

Boats bound for the Union Canal enter a circular turning basin at the bottom. The Wheel has two opposed arms, each with water-filled gondolas (or caissons) spanning between them. Those gondolas are mounted on bearings that ride on circular rails. When one goes up, the other comes down, so traffic can move both ways. The wheel is driven at its center using hydraulic motors that keep the motion smooth and slow. Idler pinions mesh between two identical ring gears: one fixed and centered on the shaft; the other surrounding each gondola. This arrangement enables the gondolas to counter-rotate as the wheel moves, maintaining their perfect upright position throughout the full range of motion.

The elegance of the Falkirk Wheel hides some fairly complicated systems that make it function. At the top and bottom, each gondola has to be able to open and close to let boats in and out. And the aqueduct at the top needs the same capability so water doesn’t just flow off the edge when the wheel is moving. The docking and undocking procedure is a delicate dance. When a gondola reaches the top position, stow pins extend to lock it in place. Then an extendable lance connects it to a hydraulic power unit. A U-shaped seal extends to bridge the gap between the two structures, and pipes fill the gap between the gates with water, balancing the pressure. Finally, hydraulic rams open the gates on both sides, allowing boats to enter or leave. The whole process happens in reverse, and then the wheel is free to move again.

Part of the engineering genius of the Falkirk Wheel is that it’s always balanced, whether there are boats inside the gondolas or not. This is one of those confusing things about buoyancy: a floating vessel always displaces its own weight in water. Theoretically, as long as the water level stays the same, when a boat floats over an aqueduct, there is no change in forces on the columns. The displaced water flows away, balancing the new force of the boat. Same thing for the gondolas. When a boat floats in, its weight in water flows out, maintaining a balance between the two sides. As a result, the Falkirk Wheel doesn’t really require a lot of power to operate. It’s about one-and-a-half kilowatt-hours for a half turn of the wheel, often compared to the power required to boil eight kettles of water. Where I live, that’s less than 25 cents in electricity. And unlike the day-long climb of the industrial-revolution-era locks, the Wheel moves boats between levels in about five minutes.

The Scottish Canals see almost no commercial shipping these days. They’re still too small, and the road and rail networks are still faster. But the canals do see a lot of traffic. There’s a whole class of vessels specifically designed for navigating the unique and historic canals of the UK. Similar to RV culture in the US, narrowboats ply the inland waters across England, Wales, Scotland, and beyond, used for holidays, touring, and even as long-term homes. During the early Industrial Revolution, boats like this were pulled along canals by horses or donkeys from towpaths that ran alongside them. Modern narrowboats are self-propelled and often equipped with domestic comforts, including bathrooms, kitchens, heating, and internet. The number of boats has been steadily increasing over the past decade, offering the freedom and lower cost of a nomadic lifestyle on the canals.

Even for those not living on narrowboats, cruises and tours along the canals offer something unique. It’s a totally different way to experience the landscape in some of Great Britain’s most beautiful areas, and it offers insights into the history of the region that you can’t get anywhere else. And of course, you also get to see the fascinating infrastructure, including a boat lift that you won’t find anywhere else in the world.

But it doesn’t go all the way up. When a segment of the canal was relocated as part of the Millennium Link, it needed to cross the Antonine Wall, a Roman-era defensive perimeter and UNESCO World Heritage Site. Rather than disturb it, the new canal was built into a tunnel below. From the aqueduct at the top of the Wheel, two new locks raise boats the remaining distance once they pass underneath the wall to the top of the Union Canal.

The Antonine Wall marked the far northern border of the Roman Empire. On another edge, just a handful of decades before it was built, Mount Vesuvius erupted, burying the city of Pompeii in ash. But there’s a twist to that story. My friends at the podcast, RadioLab, are just about to premiere a fantastic video about the survivors of Pompeii and how we discovered that some people actually escaped. This simultaneous release is part of a collaboration with the Independent Media Initiative to highlight some of the best educational and artistic creators on the internet. I’m really thankful for the award the channel won this year for my Practical Construction special, and I’m so excited to hand off to one of my favorite shows on the internet, RadioLab, for the next video in this collaboration. Go check it out after this!

When I was a kid, my dad used to tell me, “If the only reason you want something is because it’s cool, you probably don’t need it.” You can look at me and probably tell I took that advice to heart. But there are situations where it’s worth doing something just because it’s going to be impressive. The Falkirk Wheel is a perfect example. Locks are a perfectly functional solution to get boats up and down to different elevations. There are thousands of them around the world diligently serving our inland waterways. Scotland wanted something special, something that would spark a resurgence in their canal system and revitalize the sense of pride in the communities along them. It took guts to try something completely different, and it paid off. Millions of people have visited to watch it turn or travel through it. The Falkirk Wheel didn’t just reconnect two canals. It reconnected people with the idea that infrastructure can be both useful and pretty cool.

Concrete’s Greatest Weakness

[Note that this article is a transcript of the video embedded above.]

On March 2, 1973, the Skyline Plaza tower was under construction in a suburb of Washington, DC. Crews had just placed a portion of the floor slab for the 24th story, just two floors short of the project’s final height. Shortly after lunch, workers noticed that the new slab was deflecting. Suddenly, a portion of the building collapsed, killing 14 and injuring many more. The collapse left a gap in the building 18 meters or 60 feet wide, essentially slicing it in two.

Investigators later found that workers had removed the formwork and shoring for the lower floors too early. Because of cold weather, the already-placed concrete in those lower floors hadn’t gained strength as quickly as they expected. Without the shoring transferring loads into the structure below, the under-cured concrete was forced to bear the weight. And it just wasn’t strong enough.

Concrete is an incredible material. I’ve covered a lot of concrete topics in previous videos. There are good reasons why we use so much of it in the built environment. But, and this is hard for me to say, it’s not without its flaws. Even putting aside the environmental issues, as a building material, concrete creates challenges that are unique and, in many cases, not that well-understood.

Most building materials, after they're fastened or put in place, are immediately ready to use. That’s not true for concrete, and even if it seems kind of obvious, it creates some really interesting challenges for engineers, architects, and contractors. So I’ve cast some concrete cylinders in the garage, and we’re going to break them to understand this weird property of concrete and some of the ways we work around it. I’m Grady, and this is Practical Engineering.

As soon as water meets the cement in concrete mix, the clock starts ticking, and there’s basically no stopping it. The working life of concrete consists of two key phases, and they demand almost opposite properties. Phase one has to be workable and easy to shape. Concrete placement and finishing is a ton of work with a lot of steps that each have to happen at the right time. Of course, the second phase is strength; no matter how beautifully formed concrete is, it’s useless unless it can handle its designed load.

The process begins even before the concrete arrives on site. Most large jobs rely on ready-mix batch plants, where ingredients are measured and blended according to project specifications, then loaded into rotating drum trucks for delivery. Concrete is relatively cheap by weight compared to other building materials. At its most basic, it’s just sand, gravel, cement, and water. But placing it is labor-intensive, time-sensitive, and expensive, plus many projects use a lot of it. So it’s important that the right stuff makes it to the job. Engineers often put strict specifications not only the the ingredients themselves, but how the concrete is handled on the way to the job site. Some even put limits on the number of drum revolutions allowed before the concrete is dispensed, helping to prevent ingredient breakdown and loss of entrained air.

Once on site, the first task is getting the concrete into the forms. At this stage, workability is everything. It doesn’t need to flow like water, but it should move easily enough to be placed quickly and completely. You want some flow, especially for complex shapes or when you have a lot of reinforcement. Next is consolidation - usually with vibration or agitation - to get rid of excess trapped air. For slabs, workers screed the surface to level it, then use floats to push down coarse aggregates and prepare for the final finish. This is physically demanding work, and every step has to be done before the mix becomes too stiff to work with.

We do have some tools to manage this process. Admixtures can adjust the set time and improve workability without adding extra water, which would otherwise weaken the final product. But the water in concrete isn’t a solvent that dries out. Concrete cures through a chemical reaction called hydration. The water becomes a part of the concrete. And that hydration process can be affected by jobsite conditions like temperature, wind, or delays at the batch plant, which are out of your control. That unpredictability can make a big concrete pour extremely stressful. You don’t get do-overs.

Depending on conditions, concrete typically reaches its initial set in about 2 to 4 hours. That’s when the mix is firm enough that you can’t easily press a finger into it. At this point, it’s ready for finishing, whether that’s troweling for a smooth floor, brooming for a textured sidewalk, or stamping for decorative work. Each technique has to happen during a short window between the initial and final set, when the concrete is firm enough to support workers but still soft enough to shape.

On big projects, timing is critical. Standardized tests are often used to measure set times and guide trial batches so that each task can be scheduled precisely. After final set, the next phase begins: waiting. I cast a bunch of concrete cylinders to show you exactly what I mean.

It’s 24 hours later, so let’s get these on the hydraulic press. I’ve got Brady in the shop supervising the process. And my scale isn’t calibrated, so we’ll do all the comparisons in arbitrary units of force. Some people suggested kilogradys last time I used this, so let’s go with that. Even without looking at the scale, you can tell these samples aren’t very strong. Under the press, they kind of crumble more than break apart, and this is pretty typical. After a day, concrete’s strong enough to walk on. And, depending on the structure, this could be a good time to strip off the formwork, but you’re not going to get away with much more than that. I broke 3 cylinders, and we’ll plot them on the graph like this. Let’s fast forward to 7 days.

For large projects, the concrete specifications often require a test at this point. It’s the same idea as what I’m doing here, just with more sophisticated equipment. Samples collected on site are put in cylindrical or cubic molds, taken to a lab, and cured in controlled conditions. Then they’re put into a press much more complicated than this, and the force required to break them is measured. The idea behind a 7-day test is that, if the concrete isn’t going to reach its required strength, you want to know as early as possible.

Let’s put these test results in on our graph. The average was 9300 kilogradys so, a 3X increase from the 1-day breaks. Strength gain usually follows a predictable curve, so early results can be extrapolated with reasonable confidence. If something’s wrong, you can often tell early and start planning accordingly, even if that means tearing out a pour and resetting the schedule. As costly as it sounds, it’s nothing compared to the consequences of trusting concrete that isn’t as strong as the engineer assumed in design.

This highlights one of the biggest challenges with concrete: you can’t fully test quality until after installation. Most building materials go through inspection before arriving on site. With concrete, you can test the raw ingredients and even make trial batches, but the real test is whether the mix you placed in the formwork meets strength requirements after it cures. That uncertainty adds risk. To hedge against it, suppliers often design mixes with extra strength margin to make sure that, even with some random variation, strength will never come in too low. Sometimes, waiting longer can help a borderline mix catch up. But in some cases, a failed strength test really does mean tearing everything out and starting over.

Another complication is where samples are cured. Standard lab specimens are kept in tightly controlled environments. This helps verify that the supplier met the required mix specifications. But it doesn’t always reflect conditions in the actual structure, where temperature, humidity, and weather can vary wildly. That’s why many projects also include testing of field-cured samples, which gives a more realistic picture of the in-place strength. If this had been done at Skyline Plaza, the cold-weather delays in curing might have been caught, preventing a costly and deadly failure when shoring was removed too early.

On a well-run job, a good 7-day result gives confidence that everything is on track. Even though the concrete hasn’t reached its target strength yet, you have a solid indication that it will.

I also broke some 14-day samples, not typically required on jobs, but useful for seeing the big picture. [Pause VO and show breaks if needed]. The graph shows that strength continues to rise, though the rate is already slowing. Let’s jump ahead two more weeks.

28 days is a fairly arbitrary, but widely used benchmark for when the rate of hydration flattens out. Usually, when we talk about the compressive strength of concrete - 4000 psi or 28 MPa, 10,000 kilogradys per square smoot, or whatever it might be - we’re talking about the minimum 28-day strength. A significant amount of concrete engineering is based on this strength. The goal is that 28 days after placement, you can feel confident that the structure will perform up to the maximum loads as it was designed. My 28-day samples broke at an average force of about 11,000 kilogradys, about 20 percent stronger than the 7-day ones. Pretty close to the rule of thumb that concrete reaches around 75% of its final strength after one week.

But you see the problem here. A month is a long time, and time is money in the world of construction. There are some things you can do in the interim - maybe install anchors or apply light loads. For a sidewalk or driveway that rarely sees heavy vehicles, concrete might be strong enough at 7 days. But for applications where the margin between expected loads and material strength are tighter, you just have to wait. And this can be a real problem in some cases. Think about concrete roadways. How long are you willing to wait to keep a lane closed after a repair? Tall buildings have a similar problem. If you wait 28 days for every floor to cure, it’s going to be a long and slow project. You can see how concrete cure time turns into a serious bottleneck and can often become the critical path on a construction schedule.

Luckily, there are a few ways to speed things up. One is just to use a stronger mix. The logic here is simple. Say you need a 4000 psi concrete, but you don’t want to wait 28 days. If you use a 5000 psi mix design, theoretically, you’ll hit 4000 psi after just over a week. This adds material cost, but the time savings can make it worthwhile. Other strategies include using “high early strength” cement that’s ground more finely to speed up hydration, or altering the mix ratio by adding more cement or reducing water. Heating the mix water or curing under blankets can also help.

Chemical accelerators are another tool. Calcium chloride is a popular choice because it’s cheap, but it has drawbacks. Chloride ions can speed up corrosion of steel reinforcement, so lots of engineers won’t allow calcium chloride in concrete in their projects. Non-chloride accelerators (or NCAs) have gotten better over the years and may be a safer alternative, but they still pose challenges. The curing of concrete is an exothermic reaction, so faster hydration generates more heat, which can lead to cracking as the concrete cools. And, of course, it shortens the working time for placing and finishing.

I hope you can see the complexity in all this. There is a lot we ask concrete to do, and because it hardens relatively slowly, there’s a lot riding on how and when concrete gains strength. It’s not just about stripping forms or removing shoring. In many construction projects, the strength gain of the concrete governs every downstream operation. It determines when floors can support framing, when roads can open, and when a project can move forward.

And there’s nothing magical about 28 days. It’s just four weeks. It’s a number of convenience that makes it easy to talk about concrete strength and compare properties. In fact, most concrete will continue to gain strength for months or even years after that first four weeks, depending on the mix design and steps taken during curing. And many projects require that it does. Compressive strength isn’t everything when it comes to concrete. There are time- or exposure-dependent failure modes like shrinkage, creep, and long-term degradation from freeze-thaw that play an important role in design. So some projects like dams and bridges often have 90-day requirements to ensure that the concrete eventually reaches a strength to resist them, even if it doesn’t need to happen right away.

But that 28-day convention gives a hint about concrete’s greatest weakness: time. Really, no other structural material requires you to wait weeks before knowing whether it will actually perform as expected. While most materials arrive on site ready to use, concrete requires a leap of faith. And then, a long pause.

Concrete is strong, durable, and incredibly versatile. There’s nothing like it! It’s a building material worth celebrating in many ways, but only on its own terms. You can place it quickly. You can shape it into nearly anything. But you can’t rush what happens next. That’s the challenge and the art of concrete construction: it’s a balancing act between acting fast and waiting long enough. It’s a material that embodies both a sprint and a marathon.

The Hidden Engineering of Niagara Falls

[Note that this article is a transcript of the video embedded above.]

Niagara Falls is one of the most spectacular waterfalls in the world. With a vertical drop of more than 50 meters or 164 feet and a flow rate that often exceeds 2800 cubic meters per second or 100,000 cubic feet per second, it’s one of North America’s crown jewels. Roughly ten million people visit the falls every year just to catch a glimpse of the curtains of water pouring over the edge and the constant clouds of mist at the bottom. But Niagara Falls isn’t just a tourist attraction. The special geology and hydrology of this region, situated between Lake Erie and Lake Ontario, have resulted in some fascinating feats of infrastructure, from shipping to electricity to water control. It’s basically a microcosm of all the things I love. The falls themselves have required quite a bit of engineering over the years, and they’ve even been shut off for maintenance. Let’s take a little tour of the Niagara Peninsula (even though it’s really an isthmus), and I’ll show you some of the things that aren’t usually listed in a guidebook. I’m Grady, and this is Practical Engineering.

Let’s get oriented first. This is a map of the isthmus. We’ve got Lake Erie to the south, Lake Ontario to the north, Buffalo and western New York to the East, and Ontario, Canada, to the west. The Niagara River runs northward, connecting the two great lakes. And right in the middle, it plunges off the Niagara Escarpment, creating the famous falls. On the US side, there are the American Falls and the smaller Bridal Veil Falls. And on the Canadian side is the Horseshoe Falls where a majority of the river flows. It’s pretty impressive to see in person, but it’s actually not entirely a benefit. Because these falls pose a major problem for shipping.

The Great Lakes form the largest inland freshwater transportation system in the world. Since the 19th century, they’ve served as the backbone for moving iron ore, coal, grain, and manufactured goods between the American heartland and the Atlantic Ocean. Ore from Minnesota and grain from the Midwest can travel by ship all the way to steel mills or export terminals on the East Coast. Barges and freighters are efficient at moving bulk cargo in a way rail and trucks can’t match. For a time, the Niagara Escarpment was a natural bottleneck between Lake Erie and Lake Ontario, preventing goods from moving directly between the upper lakes and the Atlantic. Freight had to be offloaded and portaged around the falls before it could continue its journey. The Erie Canal solved the problem somewhat, starting in 1825, bypassing Lake Ontario. But it could only accommodate smaller vessels, and even before the Canal opened, another solution was being planned.

The Welland Canal runs through the peninsula west of the Niagara River, connecting two massive areas by shipping traffic for the first time in 1829. The canal fueled the early growth of cities along the Great Lakes and St. Lawrence River - including Cleveland, Detroit, Milwaukee, Chicago, Toronto, Montreal, and Quebec City - and it’s been rebuilt and moved several times over its life. The Welland Canal is really a titanic engineering achievement and, were it not positioned next to one of the natural wonders of the world, it would probably be famous in its own right. Because of the huge difference in elevation between the two lakes created by the escarpment, eight separate locks are required to allow ships to traverse between them. And all different kinds do - from personal leisure craft to the lakers that stay in fresh water to the salties that travel between the lakes and the ocean through the St. Lawrence Seaway.

Starting on the upstream, Lake Erie side of the canal, the first lock isn’t really for lifting or lowering ships so much as for control. The level of Lake Erie actually fluctuates throughout the year, and there are longer-term trends as well. Wind storms also raise the level locally similar to the way storm surge works during hurricanes. The control lock does just that: it controls the level in the downstream canal. It prevents excess water from rushing down the canal when the lake is high, kind of like an airlock on a spaceship keeps air from rushing out when astronauts step outside for a spacewalk.

Downstream of the control lock, the canal splits in two. The original pathway of the canal flows through the eponymous town of Welland, while the larger and newer section of canal, the Welland Bypass… well, it bypasses Welland to the east. If you look carefully, you’ll also notice a small river, the Welland River, which passes underneath both the original and bypass canals. On the way downstream from Lake Erie to Lake Ontario, shipping traffic passes over aqueducts that pass over a natural river. A hydrological wonderland!

Continuing downstream from the aqueducts, the remaining seven locks are lift locks, more like what you think of when you imagine a lock. Notice how they’re clustered tightly around the terrain and not distributed evenly along the length of the canal. That’s the Niagara escarpment, the same geological feature that the water cascades down at the falls. This is the elevation diagram of the entire Great Lakes and St. Lawrence Seaway system from Lake Superior to the Atlantic Ocean, and you can see that this drop is the biggest one of the whole thing. And that’s pretty important for another part of the infrastructure on the peninsula.

The power available from a moving fluid is directly proportional to the flow rate multiplied by the height of the drop. In most hydropower applications, that height is created artificially by a dam. There aren’t that many places in the world where you have both a large volume of flowing water and a significant natural drop in elevation. But that combination made Niagara Falls the birthplace of large-scale electric power in North America. In 1895, the Niagara Power Company opened the Edward Dean Adams Power Plant, built with Westinghouse AC generators based on the ideas and patents of Nikola Tesla. The plant served as the basis for the modern electrical grids we have today, and many of the fundamental concepts are basically unchanged.

But the power infrastructure at Niagara Falls definitely has changed. Where the Adams Power Plant put out about 40 megawatts of power in 1895, now the combined capacity from the region is in the neighborhood of 5 gigawatts. But in both cases, it wasn’t as simple as putting a turbine at the base of the falls. While it might be technically possible to generate power by placing a water wheel directly in the stream of a waterfall like a kid’s bath toy, it’s not the most efficient way (plus it would take away from the beauty). The water used to power the hydroelectric plants on both the US and Canadian sides of the Niagara River is water that never actually flows over the falls. Instead, it’s diverted into five massive tunnels - two on the US side and three on the Canadian side.

Like most tunnels, you can’t really see the extent of the hydro tunnels at Niagara Falls. There are a few conspicuous clues though, like these gigantic buildings. These interesting protrusions from the landscape house enormous steel doors, nearly 60 feet tall, that can drop down into the tunnels and close off the flow for inspections and maintenance. Both the Ontario and New York sides of the river feature similar structures.

From the tunnels, water flows into major hydropower plants on both sides of the border: the twin Adam Beck stations on the Canadian side and Robert Moses station on the US side. Then it’s released into the the lower part of the river below the falls. When you add them up, that’s 39 turbines with a combined capacity of more than 4000 megawatts. It’s a tremendous amount of power generation in one place. But actually, that’s not all of it.

These tunnels divert 50-75% of the flow of the Niagara River. That wide range in percentage of diversion isn’t because we don’t know how much is diverted, but because we actually control how much water is diverted, depending on the tourist requirements agreed upon in a treaty by both nations. During the day in peak tourist season, more water is allowed to flow over the falls to ensure the grandeur of the falls is on full display for the huge crowds of tourists that visit every year. At night and during the winter, more of the flow is diverted to generate power. That’s all managed by this structure upstream of the falls: the international control dam.

I’ve always thought this is an interesting dam, since it doesn’t even go all the way across the river. But it doesn’t need to. This structure’s not meant to create a reservoir; it just subtly adjusts the level in the river to control how much water flows over the falls versus into the hydropower intakes. The US side of the Niagara River is pretty shallow, so that side acts kind of like an uncontrolled spillway. Then, the gates on the Canadian side can be adjusted to balance the competing demands on water between tourism and power.

But there’s one big problem with those competing needs: they both have the same timing. We want thunderous cascades of water over the falls during the day when tourists are visiting, but daytime is also when the demand for electricity is highest. It’s like if solar panels only worked at night. To accommodate this, both the US and Canada have pumped storage plants. At night, excess electricity is used to pump diverted water into reservoirs, essentially storing both the power and the extra water that’s available during off-peak hours. Then, during the day, the water is released back into the forebay of the power plants. You get a little extra power from that drop out of the reservoir into the forebays, so both sides have small hydropower facilities to capture that. But more importantly, you get a lot more water during the day than would otherwise be available to run through the big plants, making more power when it’s needed most. And there’s just something funny to me that the infrastructure is duplicated on both sides of the river, like neither country was willing to be one-upped by the other.

All of this diversion noticeably reduces the flow of water over the falls. Even when they are at ‘full blast’ during the day in the tourist season, only 50% of the flow of the Niagara River makes it over the falls. You can imagine how powerful the falls would be if 100% of the flow were to cascade over. It might seem like this diversion detracts from the majesty of the falls, but in another sense, it actually preserves it.

All waterfalls undergo some degree of erosion as the water and sediment suspended in it scours away the rocks and soil underneath. Without any diversion, Niagara Falls would be receding towards Lake Erie at a rate of about 3 feet every year. At the end of the last ice age, the falls were right at the edge of the Niagara Escarpment, but thousands of years of erosion have caused them to work their way upstream. You can actually see how far it’s already progressed by looking at this elevation map. Over the last 12,000 years or so, the falls have migrated by erosion to their current location. By diverting a significant portion of the flow, the power plants have actually slowed the rate of erosion to approximately one foot per year, which will help preserve the falls for a longer period.

While flow on the falls is downregulated by diversion for hydropower, the falls are never ‘turned off’...except for the one time in the 1960s. The smaller American Falls (and nearby Bridal Veil Falls) have a pile of loose rocks and boulders, called talus, at their base. This pile of rocky debris actually extends a good fraction of the way up the falls, and officials worried that the falls might ultimately transition into a series of rapids cascading down the slope of talus rather than remaining a majestic waterfall. So, in 1969, the Army Corps of Engineers built a temporary cofferdam between the New York shoreline and Goat Island, diverting the water over the Canadian Horseshoe Falls and leaving the American Falls dry(ish)!

After the engineers got a chance to inspect the situation, they determined that the best course of action was just to leave the majority of the talus in place, since it seemed to be stabilizing the cliff face. Sometimes, doing mostly nothing is a decision you make as an engineer, even if you have to do a monumental amount of work to come to that conclusion. So the cofferdam was taken out, and water has flowed continuously over all the falls since then.

It really highlights the complexity of Niagara Falls. On the one hand, you have one of the natural wonders of the world, an absolutely enormous set of waterfalls that inspire awe and wonder in the countless travelers who are lucky enough to take in the view. The same thing that makes it impressive for tourists (the big drop) makes it valuable for power and a major challenge for shipping. And out of that comes all kinds of fascinating infrastructure, not only to facilitate the tourism but the other stuff too: a major canal with locks and aqueducts, the international dam control gates, pumped storage reservoirs, epic tunnels, towering gates, massive hydropower plants, and so much more. It’s really a pretty remarkable place for engineering.

The Bizarre Bases of Antenna Towers

[Note that this article is a transcript of the video embedded above.]

In 1974, a new world record was set for the tallest structure on Earth. Soaring to 646 meters or 2,120 feet, the Warsaw Radio Mast was built to broadcast radio programs to Polish-speaking audiences across Europe. If the atmospheric conditions were just right, those signals could be picked up from nearly anywhere in the world. But like all big infrastructure projects, building it was only half the battle. Maintaining a structure that tall—and that slender—was incredibly expensive. Over time, the guy wires that held the tower upright began to wear out. By 1991, many of them were frayed and overdue for replacement, a job that wasn’t just costly, but also fairly complex.

To replace a guy wire, two temporary guys needed to be attached to the mast first. Then the old guy could be removed and swapped out for a new one. But on August 8, 1991, the sequence got mixed up. Reports vary, but it seems that one of the main cables was disconnected before the temporary ones were fully installed. A gust of wind twisted the tower, pulling the temporary cables away, and the unsupported mast collapsed. Incredibly, no one was injured in the failure, but it was a catastrophic loss nonetheless. Usually, the tallest structures in the world lose their position because something else is built taller. In this case, a tower in North Dakota regained the lead by default.

It’s actually not an unusual story. This particular type of structure, called a guyed mast, has some seemingly bizarre structural characteristics that make it possible, including the sometimes unusual bases that seem to defy logic. But they come with risks, too. At least nine guyed masts taller than 600 meters have collapsed, mostly in the US, and hundreds of similar shorter structures around the world as well. They’re pretty interesting structures: cool to look at, incredibly tall, just rare enough that seeing one is kind of special. So this video is an ode to guyed masts, and of course, I built a little demo in the garage to help explain how they work. I’m Grady, and this is Practical Engineering.

Radio communication is a remarkable technology that enables a huge variety of wireless devices, from garage door openers to cell phones. If humans could perceive the full spectrum of electromagnetic radiation, even just the human-made stuff, we would be completely overwhelmed by the volume and variety of information moving through the airwaves. Many of the frequencies used for communication, especially those broadcast by radio and television stations, require a clear line of sight; the path between the transmitter and receiver has to be relatively unobstructed, at least by objects that are opaque to radio waves, like the earth. That’s why many antennas are mounted at the tops of hills, mountains, or (lacking those) gigantic towers. The higher they are, the further their signals can extend.

Antenna towers are some of the tallest human-made structures in the world, with many topping out above 600 meters (roughly 2,000 feet). At that height, the distance to the horizon is more than 50 miles (or 80 kilometers). To achieve that has required some very clever structural engineering. Let me show you what I mean.

This is my model antenna tower. Pretty basic; just a steel welding rod stuck in a plate. This isn’t going to match the structural behavior of an actual mast, but it’s close enough for a garage demo. The main load on a tower like this, besides its own weight, is wind. So let’s apply some wind and see what happens. [Beat]. The tower’s still standing - it didn’t collapse. But structural engineering isn’t all about strength. A structure can “not fall down” but still fail. We also have to address the concept of serviceability: does the structure actually do what it’s meant to? And in this case, hopefully it’s clear that the answer is no. Many antennas are designed to be directional. It takes a lot of power to radiate signals, so you don’t want to waste it sending them where they’re not needed. This varies a lot depending on the end use. Radio and TV broadcasts are less sensitive to movement than microwave communications, but in general, we can’t have antenna towers wobbling around like floppy wet noodles in the sky.

You can imagine that to adequately stiffen this tower, it would have to be a lot wider at the base. And that’s just what we do with so-called self-supporting towers. They’re designed to be freestanding and stable against the wind entirely on their own. Self-supporting towers don't take up much space, so they are ideal in urban areas where land comes at a premium. But, they are expensive to build because of all the extra material required for stiffness and stability against lateral wind loads. In fact, their cost goes up roughly proportional to the height squared. For guyed masts, it's roughly height to the power of 1.5. You need more land for a guyed tower since the guys extend so far out, so there is more cost there, but above a certain height (that depends on those land costs), it becomes the most economical option. And for really tall towers, it’s really the only technically feasible one. They are just so structurally efficient, it's almost unbelievable. To give you an example, at 324 meters tall (or 1,060 feet) the Eiffel Tower weighs around 7000 tons. A guyed tower of the same height would weigh roughly five percent of that.

So let me add some guys to my tower and we’ll see how it works. Of course, you can’t add just one. Wind can come from any direction, and don’t forget one of the most important adages of civil engineering: you can’t push a rope. So it takes at least three guys to get some tension in every direction. Some towers use four lanes, but most stick with three. This seems like a more stable situation, but now we’ve got a new problem. Watch what happens when I apply a lateral load. It's still just not that stiff, and actually, the tower buckles. And here’s why:

The guys can’t pull horizontally on the tower to resist lateral loads directly. They have to be anchored to the ground, which means they meet the tower at an angle. Any tension in the cable is going to necessarily put the tower in compression as well. And what happens with skinny compression members? They buckle.

Steel can take a lot of compression. Theoretically, this rod is strong enough to hold my entire weight without a material failure. If it were short, it’d be more than capable of bearing a full Grady, but when it’s tall and skinny like this, it can barely hold its own weight. When the tower takes a lateral load, the guy wires transfer that into compressive force. And unless the structure is stiff enough, it buckles. If I move the guys out so they’re at a shallower angle, you can see it takes a lot more wind load to buckle the structure. Less cable tension is needed for an equivalent horizontal force. And this is one of the many structural tradeoffs with guyed towers. You have to balance the land cost of extending anchors outward against the cost of a stiffer tower that can withstand steeply angled guys.

But you can see we’re not quite out of the woods here. Some shorter guyed towers can get away with one level of supports, but mine is still pretty flimsy in the middle. Lateral forces can still deflect it quite a bit, and it’s still prone to buckling under compressive loads, like, for example, the weight of an antenna mounted to the top. And now this is kind of like a bridge on its side.

We’ve got supports on both ends and loads trying to bend the structure in the center. So we can do what the bridge engineers do: either stiffen the structure or add more intermediate supports. It’s a little more complicated than that though, since every guy adds additional compressive load on the tower, in addition to providing lateral support to reduce the unbraced height. You’re kind of adding to both sides of the equation. Luckily, the lower you go on the tower, the shallower the angle of the cable. Just as a little demonstration of this, let’s compare the loads my little tower can support as we add more guys. With just one level, it’s right around 50 grams. This can barely support its own weight, let alone any extra on top. With a second level halfway up, it’s quite a bit stiffer. I could get 100 grams on top with no failure. Adding two more levels, now this thing feels rock solid. I’m not sure if it comes across on camera, but the change in stiffness is dramatic. It passes the wind test with flying colors. It couldn’t quite hold a kilogram, but Brady could sit on it just fine, even if it made him a bit uneasy (since his hard hat is still damaged from the last demo).

One of the other tradeoffs with this is the pre-tension of the cables. These guys sag along their length; they’re not perfectly straight. Under high wind, they tighten up and add stiffness. But in calm conditions, that slack can cause the tower to wobble. The obvious solution is to pre-tension the guys to take the sag out, but again, that pretension puts extra compression on the tower, requiring stronger members or more guys. So this is a balancing act as well.

And then there’s the base. You have essentially two choices here. We’re used to seeing large columns with a rigid attachment to the foundation. I did a whole video on base plates diving into this topic deeper if you want to learn more. You can see in my model that, with a fixed connection, my tower holds itself up just fine without loading. Obviously, this rod is solid steel - not a thin latticework of individual members - so the behavior is a little different. But remember that buckling is a function of the end connections of the column. With the bottom fixed, it takes about 140 grams to buckle the rod. When it’s free to rotate at the bottom, it buckles at around half that. The problem in this case is that fixing such a tall tower rigidly to the foundation makes the design a lot more complicated.

If you want rigid restraint, you have to have a way to transfer the loads into the ground. So the foundation has to be designed to resist rotation and pullout forces, and for not a lot of structural benefit. So the other option is to use a spherical bearing or pin support. And if you keep your eye out, you’ll see that a lot of these masts have these sorts of unusual bases where they taper down to a narrow point. In this way, you can just rely on the guys to handle almost all the restraint. The foundation only has to resist the vertical force, and maybe a touch of shear. This allows some movement or settlement of the foundation without inducing stress into the structure. And it just makes the design process easier. Removing the restraint simplifies the structural response and makes the tower more predictable, so you don’t have to be super conservative or spend tons of engineering effort and use sophisticated modeling software in the design. Finally, some towers aren’t used to mount antennas; they are the antennas themselves. For lower frequency transmissions like AM radio, you need a big antenna, so the tower itself is energized. In those cases, the base needs to be electrically insulated from the ground, which is much easier to do at a single point. If you look closely at some towers, you’ll see they’re actually standing on a ceramic disc.

Beyond structural design, these masts come with a lot of other engineering challenges. Of course, there’s the hazard to aircraft. Aviation regulations often require them to be painted in alternating orange and white bands and equipped with warning lights, whose color and flash rate are carefully prescribed, and can even be synchronized with nearby towers to avoid dazzling pilots at night.

Ice is another big one. These towers stretch into colder, wetter layers of air where ice can build up on the mast and guys. That adds weight, but it also adds surface area, sometimes dramatically increasing wind loads. When it melts, it can fall and damage anything below, so often you’ll see protective structures over the radio transmission lines.

Lightning is another threat. For most towers, it’s not a question of IF, but rather HOW OFTEN they’ll be struck. Towers are often equipped with lightning rods or other protection devices and robust grounding systems to keep stray voltage out of the transmission lines and sensitive equipment on the ground. Obviously, those mast radiators I mentioned earlier, where the entire tower services as the antenna, can’t be grounded for lightning protection. So most use some type of spark gap to keep the tower insulated. If lightning strikes, the air in the gap ionizes, allowing the surge to safely reach the ground.

Like all infrastructure, antenna towers need maintenance - painting, changing light bulbs, and servicing antenna equipment. Technicians with specialized training for heights and electrical hazards have to do the work. Some tall towers are even equipped with elevators to provide access, but most require some manual climbing. Although the frequencies used for radio communication are non-ionizing (meaning the waves can’t break apart molecules), that doesn’t mean they aren’t dangerous. Electromagnetic radiation can generate heat; it’s the fundamental principle of a microwave oven. And if the tower itself is energized, a person can become part of the circuit.

With so much of our telecommunication happening through the internet these days, it’s easy to forget the importance of large-scale radio broadcasting and communications. The cells for cellular communications are small, so we’re used to seeing those antennas relatively close to the ground. But you have to look way up to remember how critical the other wireless systems are, especially in emergency situations where radio and television signals can be an essential link to information. So next time you pass one of these towers by, take a closer look, and I hope you’ll appreciate some of the thoughtful engineering that goes into them.

An Engineer's Perspective on the Texas Floods

[Note that this article is a transcript of the video embedded above.]

This is an animation of the weather radar in central Texas starting at noon on July 3, 2025. You can see there was torrential rain across the state throughout the afternoon from remnants of Tropical Storm Barry. But focus on this area northwest of San Antonio. Around midnight on July 4, a severe storm gets stuck in this area and just stays in place for several hours. When you put it in context with the rest of the system, it looks kind of insignificant, but that little storm dropped enough rain to raise the Guadalupe River higher than ever in recorded history, at least in the upper part of the basin. The water quickly rushed through summer camps, RV parks, and rural communities in the middle of the night. And the result was one of the deadliest inland flooding events in the past 50 years.

I live not too far from some of the worst-hit areas, and although my family wasn’t directly affected by the weather, it’s been a tough situation for me to wrestle with, personally. I spent the better part of my career as an engineer thinking about flooding and designing projects to cope with it. I’ve worked on and played in the Guadalupe River. And I have kids who are getting close to summer camp age. As a dad, it’s almost impossible to comprehend a tragedy like this. As an engineer, I’ve dedicated a large part of my professional career to understanding events exactly like it. So, as I’ve ruminated about this flood over the past few months, I’ve collected some thoughts that might be worth putting into the world. Let’s take a look at this event through an engineering lens, talk a little bit about how technical and regulatory decisions play out in the aftermath of tragedy, and see if any lessons become apparent. I’m Grady, and this is Practical Engineering.

One of the fundamental problems we face in engineering, and really life in general, is that we can’t predict the future. That sounds like a ridiculous thing to say, but out of that uncertainty comes the framework for how we think about so many things. Because, we have to make all kinds of decisions - many of them with extremely high stakes - in the face of the unknown. In civil engineering, a lot of the loads we account for come from the most classically volatile and unpredictable aspect of the earth: the weather. Wind, ice, snow, waves, and rain - you cannot look ahead 50 or 100 years and know what forces a structure will be subjected to. You just have to guess.

And that’s a pretty hard thing to do, especially because you tend to have two opposing forces pushing your guess around. On the one hand, caution dictates overestimating forces to leave a wide margin of safety, but on the other hand, costs and budget constraints tend to push the estimate the other way. I can make this dam taller or this bridge higher, but it’s going to cost me a lot more money, and maybe it’s not necessary. So how do you draw the line? The same way we try to predict the future in so many other parts of life: we look to the past.

Surely past performance is an indicator of future results, right? I know that’s a stock line, but what else do we have? Over the years, we have gone to considerable lengths to apply historical data to predictions of future floods. Of course, this gets pretty complicated. One of the resources widely used in the United States for decades is Technical Paper 40, published in 1961. It represents a monumental effort to compile rainfall data across the contiguous United States, find probability distributions that fit the data, and map the results. It’s divided up by duration and recurrence interval, so you get this big group of separate maps. But what is a recurrence interval?

I’ve talked about the so-called 100-year flood in a few of my videos, but it’s a concept so widely misunderstood that it’s worth explaining again, especially because it’s so relevant to the Guadalupe River flood in July. We can’t really use historical data to determine when a flood might happen in the future, but we can make an estimation about how probable one might be. The bigger the flood, the lower the probability that it might occur. So there’s a relationship between probability and magnitude. In hydrology, we often express the probability as a quote-unquote “return period,” which means, on average, how many years you would expect to pass before you see that magnitude equalled or exceeded again. But that “on average” is doing some heavy lifting in the definition.

This terminology is debated endlessly in the hydrologic community because saying something like the 100-year flood has an underlying implication that storms are cyclical; that somehow if a particular magnitude of storm was to occur, we might have a period of security before it happened again, or the flipside: that if a flood hadn’t occurred in some time, we might be more “due” for it. And that’s just not how it works. Floods are statistically independent events. Every year, the atmosphere rolls the metaphorical dice to see what the biggest one is going to be. The odds of rolling a two or snake eyes in craps are 1 in 36, but if you go 35 rolls without a snake eyes, the odds of rolling it on the next one haven’t changed. The dice don’t remember what happened before. No one calls snake eyes the 36-roll throw because we understand it’s possible to do it twice in a row, and it’s possible to go a lot more than 36 rolls without getting one. So why do we call it the 100-year flood? Probably because the only good alternative is the storm with a 1% annual exceedance probability. Just doesn’t roll off the tongue. But it is the technically correct definition: the 100-year rainfall is the depth of precipitation (over a given duration) that has a one percent probability of being equalled or exceeded in a given year. It’s a tough concept to wrap your head around, but it’s fundamental to engineering hydrology.

If you take a look at these maps, you can see that the 100-year rainfall over a 24-hour duration in Kerr County, Texas is around 9.5 inches (or about 240 millimeters). But again, this is from 1961. And it’s based entirely on historical data. So there are decades of rainfall not included in this analysis, not to mention limitations in the statistical methodology and data processing methods of the time. TP 40 wasn’t the only resource for precipitation frequency data in the US, but it was probably the most widely used until Atlas 14 came along, or is coming along (it’s still a work in progress). NOAA has been working to update this information with the entire historical record and more rigorous statistical methods. For most of the US, this is easy to navigate online. Just mark a spot on the map and you get this table of values and confidence intervals for a range of durations and return periods. And you can see that the 100-year, 24-hour precipitation in Kerr County is 11.5 inches (or nearly 300 millimeters). That’s a pretty big jump from the 1961 estimate - an increase of about 20 percent. What was the 100-year rainfall in 1961 is now just the 50-year storm AND look at those confidence intervals! 8 to 16 inches.

I know this is kind of long-winded, but the whole point I’m trying to make here is the tremendous uncertainty we have when it comes to hydrology. In some ways, this rainfall data is extremely rigorous, and I couldn’t even begin to explain some of the statistical methods used to develop it. It serves a really important purpose in the world of engineering, planning, and emergency management. But in another sense, it’s almost meaningless. And I can show you a few of the reasons through the lens of the Guadalupe River Flood.

Here’s an hourly map of the rainfall that hit central Texas on July 4, 2025. That yellow area is the watershed for the upper Guadalupe River. When I loop through it again, you can see that cell right there caused the majority of the flooding you probably read about on the news. It was there and gone in four hours. More rain came in later that morning and the next few days, but this was a classic flash flood: A relatively short burst of heavy rainfall on a small, steep, rocky basin, where most of it runs off into a river within minutes or hours. Here’s the thing: hourly rainfall records weren’t very common until the 1940s. I counted about 100 rain gauges used by Atlas 14 within a 50 mile radius of Hunt, Texas, where most of the fatalities occurred. None had hourly records before 1940, and of the group that did collect hourly data, only four had a record longer than 70 years. That might seem like enough data to understand flooding in the area, but let me show you why it’s not.

Here’s that loop of rainfall again. What do you see on this map? Because I’ll tell you what I see: enormous spatial variability. If you were to pick four random pixels on this map, how good a picture do you think it would give you of what really happened? That’s essentially what we’re doing with rainfall frequency analysis. Compared to modern data collection methods, like the radar rainfall I showed, our historical records are extremely sparse, especially for data that varies so significantly across space. Imagine trying to recreate the Mona Lisa from scratch with just a dozen random pixels. Most of the rain gauges we use to estimate flood probabilities have never even seen an event of the magnitude we’re trying to use them to predict. There’s a whole lot of extrapolation going on.

To hammer this point home: This is the 24-hour rainfall totals for the flood, and you can see that even within this single watershed, some areas saw extreme precipitation, while others just got an inch or 25 millimeters of rain. And actually, I mapped the percentage of the 100-year rainfall that this storm amounted to, and you can see, at least in the Upper Guadalupe Basin, only a small area got close to the 100-year rainfall. For most of the watershed, this was more like a 2- or a 5-year storm.

And here’s what makes this even tougher: When we’re talking about flooding, we don’t actually care too much about rainfall. We care about the outcome of rainfall, specifically the rise in a river or stream. Here’s the graph of a stream gage upstream of Hunt during the flood. You can see that, starting around 2:00 on the morning of the 4th, the river rose by 20 feet or 6 meters in three-and-a-half hours. A little further downstream, similar story. Starting at 2 AM, the river went up 35 feet or nearly 11 meters in 3 hours before the gage broke. That is a staggeringly fast increase. In a hydrologic sense, it’s practically a wall of water. And the results were devastating. In Kerr County, there just wasn’t enough time to coordinate an evacuation. More than 100 people were killed, many of them children. So a rain gauge here, or here, or here would have completely missed the fact that the watershed it was within was experiencing the flood of record.

That’s the value of measuring the thing you actually care about. Just like precipitation, you can take historical stream gage data, fit it to a probability distribution, and get a sense of the likelihood of major floods in the future. But these gages are even more sparse in coverage than rain gauges, their records often don’t go back as far, they’re a lot more expensive to install and maintain, and, as we saw in one graph, they can go offline, ironically as a result of flooding, completely missing the peak. Engineers or hydrologists actually often visit the affected area and map the high water line after a flood to validate and confirm the data from stream gages (or to fill in the gaps if one breaks). So, although they serve an extremely important role, most of the time when engineers are trying to predict flooding or its effect on infrastructure and the built world, instead of using stream gages, they’re using hydrologic models to convert rainfall into runoff and flooding, a process that introduces a whole new set of uncertainties into the mix.

And there’s one more thing. Everything we’ve been talking about so far is predicated on a crucial underlying assumption: temporal stationarity, basically, the idea that the distribution of extreme events doesn’t change over time - or put another way - that future precipitation can be represented by past observations. But, even though those past observations are relatively sparse, in a lot of cases, we can already see that it’s probably not a great assumption. I understand this is a point of pretty strong contention in the public discourse. But within the professional community of hydrologists, engineers, and climate scientists, it’s not really a question of “is the climate changing” but more a question of how much, how quickly, and where the effects of that are most pronounced. For example, in the Texas Volume of Atlas 14, the team tested for long-term trends in the data. They found some scattered weather stations that did show an increase in extreme rainfall over time; most of them didn’t. Other studies have found more pronounced increases by looking at only the past few decades. So there are no broad statements that capture the complexity of the situation as we understand it, and importantly, this is a tough thing to figure out.

Say you have 100 years of historical data. How many 100-year floods happened within that time? Could be a few. Could be none. So, especially for very extreme events on the 1-in-a-century scale, there’s a lot of uncertainty when it comes to teasing out any trends. That said, there is a strong consensus among the various climate models and recorded data that a warming atmosphere has already resulted in an overall increase in the intensity and frequency of rainfall, a trend that will likely continue. And you can see why that poses a problem. Particularly for infrastructure with a design life of 50 to 100 years, we need to design not just for the storms of today but those decades in the future, and our current methods of doing that is, on average, systematically underestimating them if we assume a stationary climate.

Just to be clear, I’m not trying to blame a flood on climate change. Although attribution studies can estimate the contribution of extra energy in the climate system, there’s no way to ascribe any particular weather event to global warming deterministically. For many places, it might not even be a major source of uncertainty compared to all the other factors I’ve mentioned when it comes to predicting the magnitude of future floods. My point is that it’s just one more confounding aspect of estimating flood risks. And it gets to the heart of the entire issue. Because why does any of this even matter?

There‘s been a lot of discourse about what should have happened before the storm and what should be done in its wake. But before you can take any action to mitigate flood impacts, you have to know what the actual risks are. On the upper Guadalupe, we’ve seen it with our eyes, but how many similar watersheds just got lucky that night, or really, any night? I think you’ll agree with me that this is complicated stuff. And humans are notoriously bad at using probabilities and risks to make decisions. Almost nothing in our biology is optimized for long-term, rational decision-making about rare and extreme events. Almost every day of everyone’s lives, there’s not a flood. That makes it really tough to consider it as a priority and devote resources toward preparations. And I think part of the problem is that we rarely talk about the uncertainties.

Even within the field of engineering, where we should know better, we have a strong tendency to treat everything deterministically. It sure makes things a lot simpler. Take the bold number in the table, plug it into your equations and computer models, and just forget that those uncertainty bands even exist. In some ways, it makes sense. Ultimately, you do have to choose a number: how high to build a bridge or how large a culvert to install, or how wide to make a spillway. But, in a lot of cases, those decisions get translated into a sort of confidence that doesn’t actually exist. The concept of the floodplain is a perfect example.

In the US, a lot of the framework for how we think about and prepare for floods comes out of the National Flood Insurance Program. And to participate in this program, communities are required to regulate what happens in the floodplain, or more specifically, what and how things get built there. And so, a fundamental part of regulating the floodplain is deciding where it actually is and isn’t. We’re not going to dive into that process, but billions of dollars have been invested in making these maps and keeping them up to date in the US.

If you take a look at one, it’s a lot to parse depending on the location. There are quite a few different hazard areas with different meanings. The simplest for riverine locations is the base flood, essentially the 100-year flood. Some maps show the 500-year flood as well. Many maps show the floodway, which is kind of the main part of the channel needed to pass floods, so it’s usually regulated more strictly. But there’s something I notice when I look at floodplain maps. All of these zones are bordered with nice crisp lines. You’re inside the floodplain here, and you’re outside of it here. And property owners often go to great lengths to refine these maps; to shift the line just slightly and reduce their regulatory responsibilities. But consider everything we’ve talked about with estimating flood risk and ask yourself, what’s the difference in the risk profile between here and here? Is it enough to have a sharp line between them? And if not - if the true situation is more nebulous - is the map doing a good job of communicating flood risk to the public?

Because, just to be clear, that is one of the stated purposes of floodplain maps. Of course you need to delineate zones clearly to be able to regulate where permits are required and where buildings can be built and so on. But, to me at least, it sends a complicated message to have this binary definition of inside the floodplain or outside of it as a way to explain to individuals, homeowners, renters, and the general public about the risks that they’re actually exposed to.

You look at these maps and there is absolutely no indication about uncertainty, despite the fact that almost every step of the process that goes into creating them has huge margins of error. And then, when we get more historical data, or land uses change, or our understanding of the floodplain evolves, and we try to change the map, that immediately sows distrust. You hear it all the time (at least if you run in similar circles as I do): “We’ve had two hundred-year floods in the past 5 years. These engineers don’t know what they’re talking about…” Part of that, of course, is just a misunderstanding about what the hundred-year flood actually means, but part of it is that we don’t do a good job communicating risk and uncertainty well. The meteorologists get the same thing. People get salty when forecasts are wrong without any acknowledgement at all that the job is essentially predicting the future. You know, it’s wizard stuff. Weather is really complicated, and I think we have a lot of room to grow in how we discuss and disseminate the things we don’t know for sure.

Because flooding is capricious. If you look back at the maps from July 4, you can see a lot of places where rainfall was more intense than in Kerr County and the Guadalupe River. Many areas of central Texas received more than the 100-year, 24-hour precipitation from Atlas 14. And there were severe storms and flooding across the region in the days that followed as well. But nearly all the fatalities happened in this one place. I don’t have a good answer for why. Maybe some combination of timing, warning systems, the rural location, differences in floodplain regulations, and plain bad luck. I think scientists, engineers, and emergency planners can probably learn a lot by simply comparing the flooding between Kerr County and some of the other areas in central Texas hit by this storm system, and why the outcomes were so drastically different.

My heart goes out to the victims and their families who were affected by this flood. I’ve been thinking so much about it in the weeks since, and why these kinds of risks can go so underappreciated that we wouldn’t bat an eye at having such a large population of people sleeping in the floodplain of a flashy watershed.

I think there are a lot of lessons to learn here, but the one that keeps coming back to me is about communication. People can’t act to reduce their risk unless they can internalize what it actually is. Professionals think about these issues every day; they have technical training, knowledge, and experience to make informed decisions about infrastructure, land use, and zoning. But most people don’t have the same cognizance of the hazards. You can’t blame them. It’s a crazy world we live in, and even individuals who live, work, and play in areas at risk of flooding might not come face-to-face with the danger in their entire lives. Like I said, weather is complicated, and we don’t all have the headspace to try and understand spatial variability, annual exceedance probabilities, climate stationarity, and so on.

So I think the professional community has a responsibility to improve how we communicate flood risks to the public, not only for accessibility but honesty. We need to have language that anyone can grasp, but we also need to be better about acknowledging uncertainty. It sounds counterintuitive, but I think facing the limitations of our understanding head-on actually instills more trust than pretending like we have all the answers. And when people understand those uncertainties, they get a deeper appreciation for how flood hazards vary across the landscape, giving them more insight, not less, to prepare for what’s ahead. Thanks for watching, and let me know what you think.

The Weirdest Tool in Underwater Construction

[Note that this article is a transcript of the video embedded above.]

In 1989, the Loma Prieta earthquake shook the central coast of California, collapsing buildings and damaging infrastructure across the Bay Area. Bridges, in particular, suffered extensive damage. In one case, a major section of the eastern span of the Bay Bridge's deck collapsed, falling onto the lower deck like a trapdoor. Sadly, one person died driving off the upper deck. Crews had the bridge repaired within a month, but Caltrans knew that the next earthquake could be worse and started making plans to replace the structure.

Knowing that the replacement project would require heavy-duty piles, Caltrans developed a testing program to identify risks and challenges during design and minimize the chance of unanticipated problems cropping up during construction. And they found a pretty big one. In October of 2000, the barge began the pile driving operation, dropping a large hammer to drive the 8-foot (or 2.4 meter) diameter steel pipe deep into the seafloor. Almost immediately, fish began dying in the surrounding area. Biologists involved in the project collected fish and documented injuries to their organs and swim bladders. They weren’t being directly hurt by the hammer itself; it was above the water anyway. The damage was coming from the intense sound.

That massive steel pipe rang like a humongous bell on every hammer blow, radiating sound pressure through the San Francisco Bay. It even had serious impacts on aquatic wildlife up to a kilometer away, which was a pretty big deal. Because San Francisco Bay is home to quite a few threatened or endangered species of fish. The problem was that the replacement bridge would need more than 250 of these piles. Caltrans had to figure out how to install them without affecting the wildlife in the process, and the way they did it, I think, is pretty cool. And I even built a model in the garage to show you it works. I’m Grady, and this is Practical Engineering.

If you want to know the answer right away, it's bubbles. But I think the most interesting part is why it works in the first place. And this matters. Pile driving isn’t the only thing that creates excessive noise underwater. We do a lot of construction in waterways, oceans, rivers, and bays. We also occasionally have to blow stuff up underwater, like for demolition of structures or safe disposal of old munitions and mines. Any loud work underwater has the potential to disrupt, injure, or even kill aquatic wildlife. The phenomenon we know as sound is just fluctuations in pressure within a medium, whether it’s air or water (or even concrete). We sense those fluctuations mainly through our ears, but pressure fluctuations can do a lot more than just vibrate the thin membranes, tiny bones, and hairs. Barotrauma is the term used to describe the damaging effects of compression and decompression on wildlife. And it really has only been in the past few decades that we’ve really started to apply the science of hydroacoustics to our own activities and try to mitigate the impacts.

You’ve probably heard of sound pressure expressed in decibels. It’s really just a logarithmic scale of convenience thing because meaningful pressures can range across many orders of magnitude. So the decibel system just makes the numbers easier to compare. The equation for a decibel is just 20 times the base 10 logarithmic function of the sound pressure divided by a reference pressure. Sounds complicated, but it just means a 1-decibel increase corresponds to an increase in sound pressure of about 26 percent. The amount of time over which sound pressure is measured also matters. Look at a waveform and you can see there are peaks (both in compression and rarefaction). But that’s only for a split second. So a lot of measurements use a root mean square of the sound pressure over a given time to provide a better estimate. We don’t have to go into the math of that, just think of it as a fancy kind of average.

It’s important to point out that, in air, we use 20 micropascals as the reference pressure, which is approximately the limit of human hearing. So that’s 0 decibels. Underwater, we use a reference pressure of 1 micropascal, mainly just for standardization purposes, so just keep in mind that underwater decibels aren’t really equivalent to sound pressures you might have as references in your head like the 75-decibel vacuum cleaner or the 140-decibel jet engine. And really, what you think of as sound has less meaning underwater because our ears and brains are calibrated for the physics of sound in air. The underwater version of “loudness” doesn’t translate well to human perception. But it matters a lot to fish and marine mammals.

Sound behaves a lot differently in water than air. Of course, water is denser, and sound moves through it at roughly 4 times the speed it does in air. Sound also carries a lot further in water, and importantly, the acoustic impedance of water is way different than air. Impedance is basically a measure of opposition to sound flow, kind of like resistance in an electrical circuit. It’s a function of the medium’s density and the speed of sound through it. And at a boundary between two media, there are two things that can happen to sound. It can transmit into the new medium or it can reflect back, and the difference in impedance between the two determines how much of each will occur. If impedances match, more sound will transmit through the boundary. If they’re way off, like water and air, most of the sound is reflected. The practical effect of that is a transmission loss between air and water of about 30 decibels. It’s why stuff happening underwater is quiet above the surface, and we can take advantage of impedance mismatch in underwater construction.

I built a new acrylic tank for this demo, and I’ve got a new helper in the shop. This is Brady. I figured since half the internet calls me that anyway, we might as well get a Brady in here. He can wave and nod, and he can probably do a lot of other stuff too, but that took me several hours, so he’s just going to bravely hold the hydrophone for now. And on the other end of the tank, I have this bluetooth speaker. It claims it’s underwater rated, so we’ll see if it works out.

And here’s the setup; pretty simple. I found a few recordings of hammering and pile driving sounds to play on the speaker. And this is how they come across on the hydrophone, which is connected to a sound recorder. I also did a frequency sweep so we can do a little more scientific comparison. Now let’s add some air.

At this point, one of my glue joints on this tank catastrophically failed and flooded my garage with water. I didn’t catch it on camera, but Brady took the brunt of the fall. Thankfully, he was wearing his hard hat. I got it all fixed up, and now let’s see if we can soften these construction sounds.

I have four air stones made for aquariums hooked up to an air pump. When I flip these on, we get a nice curtain of small bubbles between the speaker and the hydrophone. And I’ll record those same sounds again. Here’s a look at the waveforms from the hydrophone with and without the air. Although it’s not a dramatic difference, you can definitely see a difference, especially for the higher pitched hammering sounds toward the end. And here’s a look at the waveforms without and with the air for the frequency sweeps. Even though the sweep should have had a constant sound pressure across the full range of frequencies, the water and demo itself cause pretty serious distortions. You can see a lot of resonance at low frequencies, and a lot of attenuation at high frequencies. That makes it a little hard to gauge the effectiveness of the bubbles. It’s similar to the hammering sounds - not much difference at the lower frequencies, but a pretty substantial reduction at higher frequencies.

This is not an ideal setup for one reason: even though there’s a big mismatch in acoustic impedance between air and water, there’s not that much difference between acrylic and water. So, it’s pretty easy for pressure waves to propagate into the acrylic, travel past my bubble curtain, and back into the water on the other side. So I’m not getting the kind of sound reduction, what the pros call attenuation, that you might expect in the real world, for example, by surrounding a pile with a circular ring of air pipes. Thankfully, the researchers studying solutions like this have put a lot more resources into figuring out the right way to do it. The measurements at the Bay Bridge compared fairly well with mine. Attenuation was highest as the higher frequencies. But this is not as simple as just blasting air out of a pipe.

These bubble curtain systems require a lot of logistics. Massive compressors or blowers feed air sometimes deep below the surface into complex plumbing assemblies. They usually have filters to remove oil from the air to make sure the water isn’t being contaminated. The system has to sit flush with the bottom to make sure sound can’t travel underneath the bubble curtain. But also, there are currents. Any movement of the water is going to move the bubbles too, potentially creating gaps in the curtain or dispersing it altogether. So it’s often necessary to have multiple levels of plumbing to keep a continuous screen all the way to the surface. If that’s not enough, there are ways to confine the bubbles around a pile or construction activity using an outer casing or even a flexible membrane. But how do you know it actually works?

Maybe the most comprehensive engineering guidance on this topic is put out by Caltrans in their manual on the Hydroacoustic Effects of Pile Driving on Fish. Appendix 1 in the report is a nearly 300-page compendium of pile driving sound data. You might not have known this, but we’ve been measuring a lot of pile-driving sounds! If you’re an engineer or environmental scientist trying to get a permit to build something underwater and sound is going to be an issue, this is kind of your bible. It’s got quite a few ways to minimize impacts, including timing work when important species aren’t present, changing designs to reduce underwater work, using vibratory hammers instead of conventional equipment, and bubble curtains that reduce the propagation of underwater sound pressure. Based on all the testing and real-world case studies so far, they suggest you can get about 5 decibels of attenuation this way.

Just like my demo, sounds don’t only travel through the water. They also move through the sea floor and even through the barge on the surface, bypassing the bubbles. 5 decibels doesn’t sound like a big reduction, but you have to remember that it’s a logarithmic scale. A 5 decibel reduction means the actual sound pressure is nearly cut in half. You also have to remember that what we care about most is area. For any loud construction or demolition activity, there’s an invisible ring some distance away that marks the injury threshold level. Since sound pressure decreases with distance, eventually you’re far enough away from the sound that it doesn’t result in injury. So every foot or meter that you can pull that ring back toward the activity through attenuation reduces the impact area proportional to the distance squared, dramatically reducing the area in which fish may sustain injuries. That’s why bubble curtains are used in so many underwater construction projects these days, but that’s not all they’re used for.

What’s that old saying? If your only tool is a bubble curtain generation system, every problem starts to look like a loud underwater sound. Something like that. It turns out that bubbles can do a lot more than create an impedance mismatch for sound pressure propagation. For one, they aerate water, which can be useful to prevent algae and other issues with stagnant pools. For two, they create vertical water currents. That can help keep things separated, like trash. You can see it’s a lot harder for me to move this little boat across the barrier created by the bubbles. Of course, a net or boom or rack can do this too, but those don’t allow boats to pass through. And this doesn’t just work for trash. Bubble curtains have been used to contain oil spills, and they’re often used in underwater construction not just to control sound but turbidity. We really don’t want disturbed sediments clouding up our waterways, again, primarily for environmental reasons, so these can be an important tool when booms aren’t practical. They’ve also been used to control saltwater and keep it from migrating up rivers in tidal areas. And they’ve even been employed to confine herbicides for invasive plants, allowing for fewer chemicals and less non-target damage to nearby flora.

I’ll definitely be in trouble with the biology folks if I don’t point out that it’s not just people who use bubbles as a tool. Humpback whales cooperate to create bubble curtains that corral fish to a central point. Then they lunge into the center to gulp them down, a behavior called bubble net feeding. And we use bubbles this way on occasion as well, not for fishing but to keep fish out of certain areas, usually to prevent the spread of invasive species.

By 2005, the pile driving operation on the east span replacement of the Bay Bridge was complete, and Caltrans and its consultant were awarded the Environmental Excellence Award by the Federal Highway Administration for all the work they did on minimizing underwater noise impacts on endangered fish species. And the lessons from that project have been applied across the world in the two decades since.

You know I love heavy construction. The bigger and louder the machinery, the better. But I think that anything we can do to limit the effect we have on the other things we share this world with is a win, especially when it’s something as clever and creative as blowing bubbles.

California’s Tallest Bridge Has Nothing Underneath

[Note that this article is a transcript of the video embedded above.]

Foresthill Bridge soars across the valley of the North Fork of the American River just outside Auburn, California. At more than 700 feet or 200 meters above the canyon floor, it’s the fourth-tallest bridge in the United States. When it opened in 1973, crowds cheered for the impressive new structure. But if you take a closer look, it doesn’t really make any sense.

This isn’t an interstate highway or even a major thoroughfare. The road sees only a few thousand vehicles a day, connecting Auburn, an exurb of Sacramento with a population just shy of 14,000, to scattered rural communities and recreation areas in the western foothills of the Sierra Nevadas. And while the American River does occasionally flood, it doesn’t flood 700 feet. Before this, the crossing was basically a low-water bridge.

A structure of this magnitude just looks out of place. But it wasn’t just a boondoggle, at least not at the outset. It was built that way for a reason, and the story behind it is not only pretty wild, but it also sits at the hinge point of a major chapter in American infrastructure. I’m Grady, and this is Practical Engineering.

California’s Central Valley is one of the world’s great agricultural regions: over 400 miles long, more than 50 miles wide, this remarkably fertile area is nearly half the size of England. The city of Sacramento sits near its center, right where the Sacramento and American Rivers meet.

To manage and distribute water across this enormous landscape, the federal government launched the Central Valley Project in 1933, a sweeping effort by the U.S. Bureau of Reclamation to store water in the wetter northern part of the valley and distribute it to the drier south. In the process, the system would also generate hydropower and reduce flood risk for growing urban centers. I’m glossing over a lot here. The history of California is steeped in water issues, and even just the Central Valley Project is nearly a century of details. But, critically, Folsom Dam was one of the first big components of the plan.

Built in 1955 on the American River, the concrete gravity dam provided significant flood protection to the City of Sacramento. However, it was constructed relatively early in our understanding of basin-scale hydrology and the uncertainty surrounding the frequency and magnitude of flooding over long periods of time. It became clear pretty quickly that Folsom Dam didn’t quite offer as much flood protection as was originally promised. Plus, because Folsom had to keep its flood pool empty to handle potential inflows, its ability to store water for irrigation or municipal supply purposes was somewhat limited.

The answer to these problems, at least according to the federal government, was Auburn Dam, authorized by Congress in 1968. The new structure would sit upstream of Folsom and control the variable flows of the North and Middle Forks of the American River. It would be the tallest dam in California and one of the tallest in the country. And work began in earnest in the early 1970s.

One of the first steps in the process was rerouting the American River. Crews built a large cofferdam and carved a diversion tunnel through the canyon wall. With the water redirected, they could begin drying out the bend in the river where the huge new dam would eventually sit.

Once the site was dried out, crews began exploring the underlying geology more thoroughly. They drilled boreholes, excavated tunnels and shafts, and surveyed the rock that would serve as the dam’s foundation. The site’s geology turned out to be more complex than expected. Some zones of rock were more compressible than others, which could lead to dangerous stress concentrations in the dam. And, there were a lot of joints and fissures in the rock mass, making it more challenging to predict how they would behave under extreme loads, in addition to creating paths for water. So the next phase of the project was a major foundation treatment program starting in 1974. This mainly involved pressure grouting fractures to reinforce weak zones against the enormous weight of the structure and to make the geology more watertight, preventing seepage from flowing under the dam.

With major construction works underway, anticipation for the reservoir was growing. Around the future rim, land values soared, and developers rushed to stake claims. Lakefront homes were planned. Entire communities emerged, built on the promise of a shining new shoreline. Then, in August 1975, a magnitude 5.9 earthquake struck near Oroville Dam, only about 50 miles or 80 kilometers away from the site.

The quake only caused minor damage to structures in the area, but it rattled confidence in the Auburn project. The geology of the western Sierra Nevadas had long been considered stable. But the Oroville earthquake introduced a troubling possibility: that the loading and filling of large reservoirs could trigger seismic events in the area. This phenomenon, known as reservoir-induced seismicity, is still not well understood even to this day. The pressure of water infiltrating bedrock and the weight of a reservoir can change the balance of forces along faults, potentially triggering movement. You know, when Oroville is full, that’s roughly 10 billion pounds of force or 4 billion kilograms of mass. It’s a staggering amount. You can imagine how that might affect the underlying geology.

The Auburn Dam, as a thin concrete arch, in contrast to the concrete gravity dam at Folsom or the earthfill embankment at Oroville, would be especially vulnerable to earthquakes. Thin-arch dams rely on the canyon walls to resist the thrust of the structure. In fact, I’ve made a video all about the topic you can check out after this! If one side shifts even a little during a quake, the results could be catastrophic. In April 1976, a report by the Association of Engineering Geologists concluded that an earthquake like the one at Oroville could cause the proposed Auburn Dam to catastrophically fail. It was back to the drawing board for the project, even as the foundation grouting program continued. And then the project was shaken again.

That same year, the newly completed Teton Dam in Idaho collapsed during its first filling, killing 11 people and causing billions in damage. It had been built by the same agency, the Bureau of Reclamation. Concern continued to mount about the safety of Auburn Dam, which would have catastrophic consequences for the thousands of Californians downstream if it were to fail. It was all enough to bring Auburn’s momentum to a halt.

While dam construction paused, one aspect of the project had already been finished: Foresthill Bridge. With a cofferdam on the river and the diversion tunnel only sized for smaller floods, there was a risk of overtopping the existing bridge, cutting off access between Auburn and the Sierra foothills. So, the Bureau of Reclamation decided to get a head start on a project that would eventually be inevitable: a new bridge, permanent and high enough to span the reservoir once it filled. If they were going to build a new bridge, they figured they might as well build it right the first time.

The result was a striking steel cantilever bridge with two slender concrete piers soaring skyward from the canyon floor. [Actually, there was another bridge planned over the Middle Fork of the American River - the Ruck-a-Chucky Bridge. It was a wild idea: a curved cable-stayed bridge where all the cables are anchored in the hillsides rather than tall towers. But while that project was shelved, Foresthill made it all the way through design and construction.] At the time of its opening in 1973, it was the second-highest bridge in the United States. But as time went on, it became increasingly clear they had jumped the gun.

By 1980, engineers floated two new dam designs that could withstand potential earthquakes. Both would be shifted slightly downstream from the original site. But by then, the tide of public and government support for the dam had turned.

Construction costs had ballooned, and Auburn Dam was looking less feasible every day. As originally proposed, the structure would be even larger than the Hoover Dam size, but store less than 10% of Lake Mead’s volume. Meanwhile, upgrades to Folsom Dam and improved levees around Sacramento offered far cheaper ways to reduce the flood risk that was the major impetus for the dam in the first place. New hydrologic data also suggested that earlier flow estimates had been overly optimistic, reducing its value for conservation. The benefits of Auburn Dam were shrinking as the costs grew. It was turning into an incredibly expensive solution in search of a problem.

At the same time, environmental and advocacy groups were gaining momentum. The project would flood canyons used for whitewater rafting and kayaking. It would drown ecosystems, inundate archaeological sites, and destroy long segments of the wild and scenic forks of the American River. It became clearer and clearer that the ends simply couldn’t justify the means. And yet, the idea never fully went away.

In 1986, a massive flood hit the area. Water backed up at the diversion tunnel at Auburn, overtopped the cofferdam, and caused it to fail. Downstream levees were breached, and much of Sacramento flooded. For a moment, the momentum behind Auburn Dam and its promise of flood protection returned. But, it later became clear that the flood wasn’t entirely a natural disaster. The Bureau hadn’t followed the operating guidelines at Folsom Dam, worsening conditions downstream. And by then, grassroots opposition, cost concerns, and shifting priorities had all but put the Auburn Dam project to bed. Various proposals resurfaced over the years, including the idea of a “dry dam” that would only hold water during floods, but none gained much traction. With its many iterations and proposals, the project became known as the dam that wouldn’t die. But in 2008, the state of California revoked the Bureau’s water rights permit for the project, maybe not sealing its fate completely, but at least burying it several feet deeper.

This story really gets to the heart of the challenge with large-scale public works projects. No matter how you configure them, there are big losers and big winners. There’s no doubt that a dam across the American River upstream of Folsom could provide significant benefits to the public: flood control, water supply, hydropower, recreational opportunities, or some combination of them all. But those benefits have to be weighed against real costs: environmental damage, staggering capital investment, long-term maintenance, the inherent risk of catastrophic failure, and the social toll of displacement and disruption.

The mid-20th century was the heyday of American dam building, an era driven by ambition and optimism, but also by uncertainty. We didn’t have enough historical data to fully understand river systems. We couldn’t yet grasp the long-term consequences of altering them. And we couldn’t see into the future to know what the true impacts of these structures would be or what the cost of keeping them in good shape might amount to.

Since then, we have a lot more experience with huge multi-purpose reservoirs. And it seems, in general, that the more we learn, the more the answer to whether they’re worth it seems to be: maybe not. And that maybe turns into a probably when you consider that all the best sites are already taken.

New Melones Dam, completed by the Bureau of Reclamation in 1979, not too far from Auburn, faced a lot of similar controversy and pushback. Although the project was eventually completed, the fight was bitter, and its legacy so far is mixed. The project is widely considered to be the last great American dam. At least, great in size, if not in public sentiment. No other reservoir of that scale has been built in the U.S. since. And with the Auburn Dam project mostly dead, it seems doubtful there ever will be.

The American River continued flowing through the diversion tunnel until 2007, when a new pump station and restoration project returned the river to its original channel. Kayakers can now navigate downstream, and even have some new features at the pump station to choose from: the artificial rapids on the left or the screen channel on the right. After more than three decades, the river was back in its place, tying a bow on a dam that was never built.

And yet, just a few miles upstream, the Foresthill Bridge still stands, dramatic, overbuilt, and strangely out of sync with its surroundings. And we’re still kind of stuck taking care of this bridge, whose scale is so out of proportion with its purpose. In the 2010s, the bridge underwent a major seismic retrofit to improve its safety and make future inspections easier. Most recently, it was part of a nationwide program inspecting bridges built with T-1 steel, an alloy that, in some cases, has shown concerning cracking at welds. The I-40 bridge crack in Memphis, which I covered in an earlier video, triggered the effort. And there have been quite a few defects found in bridges since then, so here’s hoping that Foresthill doesn’t make the list.

It’s a cool structure in its own right. But it stands for more than just an engineering achievement. Auburn Dam left a lot of scars, both on the physical landscape and the political one. But it also left this bridge that became more than just an out-of-place oddity. In a sense, it’s become a monument to the end of an era in US major public works projects, and, hopefully, a tribute to the caution and care that will shape the next one.

Why Are There No Short Arch Dams?

[Note that this article is a transcript of the video embedded above.]

Flaming Gorge Dam rises from the Green River in northern Utah like a concrete wedge driven into the canyon, anchored against the sheer rock walls that flank it. It’s quintessential, in a way. It’s what we picture when we think about dams: a hulking, but also somehow graceful, wall of concrete stretching across a narrow rocky valley. But to dam engineers, there’s nothing quintessential about it. So-called arch dams are actually pretty rare. For reference, the US has about 92,000 dams listed in the national inventory. I couldn’t find an exact number, but based on a little bit of research, I estimate that we have maybe around 50 arch dams - it’s less than a tenth of a percent.

The only reason we think of arch dams as archetypal is because they’re so huge. I counted 11 in the US that have their own visitor center. There just aren’t that many works of infrastructure that double as tourist destinations, and the reason for it is, I think, kind of interesting. Because an arch dam isn’t just an engineering solution to holding back water, and it’s not just a solution to holding back a lot of water. It’s all about height, and I built a little demo to show you what I mean. I’m Grady, and this is Practical Engineering.

Engineers love categories, and dams are no exception. You can group them in a lot of ways, but mostly, we care about how they handle the incredible force of water they hold back. Embankment dams do it with earth or rock, relying on friction between the individual particles that make up the structure. Gravity dams do it with weight. Let me show you an example.

I have my tried and trusted acrylic flume with a small plastic dam. Once this is all set up, I can start filling up the reservoir. This little dam is a little narrower than the flume. It doesn’t touch the sides, so it leaks a bit. The reason for that will be clear in a moment. And hopefully you can see what’s about to happen. This gravity dam doesn’t have much gravity in it, so it doesn’t take much water at all before you get a failure. I’m counting failure as the first sign of movement, by the way. That’s when the stabilizing forces are overcome by the destabilizing ones. And the little dam by itself could hold until my reservoir was about a quarter of the way to the top.

Gravity dams get their stability against sliding from… you guessed it… friction. Bet you thought I was going to say gravity. And actually, it kind of is gravity, since frictional resistance is a function of just two variables: the normal force (in other words, the weight of the structure) and a coefficient that depends on the two materials touching. Engineers analyze the stability of gravity dams in cross-section, essentially taking a small slice of the structure. You want every slice to be able to support itself. That’s why I didn’t want the demo touching the sides of the flume; it would add resistance that doesn’t actually exist in a cross-section. The destabilizing force is hydrostatic pressure from the reservoir, which increases with depth. And the stabilizing force is friction. There are some complexities to this that we’ll get into, but very generally, as long as you have more friction than pressure, you’re good; you have a stable structure.

So let’s add some normal force to the demo and see what happens. [Beat] You can see my little reservoir gets a little higher before the dam fails, about halfway to the top. And we can try it again with more weight. But the result gets a little more interesting… the dam didn’t actually slide this time, but it still failed.

Turns out gravity dams have two major failure modes: sliding and overturning. Resistance to sliding comes from friction, which really doesn’t depend on how the weight of the dam is distributed. That’s not true for overturning failures. Let’s look back at our cross-section. For a unit width of dam, the hydrostatic pressure from the reservoir looks like this. Pressure increases with depth. And the area under this line is the total force pushing the dam downstream. We can simplify that distribution and treat it like it’s a single force, and it turns out when you do that, the force acts a third of the way up the total depth of water. Most dams want to rotate about the downstream toe, so you have a destabilizing force offset from the point of rotation. In other words, you have a torque, also called a moment. The dam has to create an opposite moment around that point to remain stable. Moment or torque is calculated as the force multiplied by its perpendicular distance from the point of rotation. So, the further the center of mass is from the downstream toe, the more stable the structure is, and the demo shows it too.

Here’s where we left the weights the last time, and let’s see it happen again. The reservoir makes it about two-thirds of the way up the walls before the dam overturns. Let’s make a simple shift. Just move the weights further upstream and try again. It’s not a big difference. The reservoir reaches about three-quarters the way up before we see a sliding failure, but shifting the weights did increase the stability. And this is why a lot of gravity dams have a fairly consistent shape, with most of the weight concentrated on the upstream side, and usually a sloped or stepped downstream face.

Interestingly, you can use the force of water against itself in a way. Watch what happens when I turn my little model around. Now the hydrostatic pressure applies both a destabilizing and stabilizing force, so you get more resistance for a given depth. A lot of deployable temporary storm barriers and cofferdam systems take advantage of this kind of configuration. You can imagine if I extended the base even further, I could create a structure that was self-stable just from its geometry alone. The weight of the water on the footing would overcome the lateral pressure. But there’s a catch to this. This is fully stable now, but watch what happens when I give the dam just a bit of a tilt. All of a sudden, it’s no longer stable.

This might seem kind of intuitive, but I think it’s important to explain what’s actually going on. Hydrostatic pressure from the reservoir doesn’t only act on the face of a dam. With smooth plastic on smooth plastic, you get a pretty nice seal, but as soon as even a tiny gap opens, water gets underneath. Now there’s upward pressure on the bottom of the dam as well. If you’re depending on the downward force of a dam from its weight for stability, it’s easy to see why an upward force is a bad thing. And it’s so dramatic in the example with the upstream footing specifically. In that case, the downward pressure of the reservoir is acting as a stabilizing force, but if water can get underneath that footing, it basically cancels out. The pressure on the bottom is the same as the pressure on the top. But this isn’t only an issue in that case.

The ground isn’t waterproof. In fact, I’ve done a video all about the topic. Soil and rock works more like a sponge than a solid material, and water can flow through them. That’s how we get aquifers and wells and springs and such. But it’s a problem for gravity dams, because water can seep below the structure and apply pressure to the bottom, essentially counteracting its weight. We call it uplift.

Looking back at the cross-section, we can estimate this. Of course, you have the triangular pressure distribution along the upstream face. But at this point you have the full hydrostatic pressure also pushing upward. And at the downstream toe, you have no pressure (it’s exposed to the atmosphere). So, now you have a pressure distribution below the dam that looks like this. Of course, this part can get a lot more complicated since most dams don’t sit flush with the ground, and many are equipped with drains and cutoff walls, so definitely go check that other video out if you want to learn more.

But let me show you the issue this causes with some recreational math on our cross-sectional slice of the dam. The taller the dam, the greater the uplift force. That happens linearly. In other words, the force is proportional to the depth of the reservoir. But look at the lateral force. Again, remember it’s the area under this triangle. Maybe you remember that formula: one-half times base times height. Well, the height is the depth of the water. And the base is also a function of the depth. More specifically, it’s the unit weight of water times depth. Multiply it together, and you see the challenge: the force increases as a function of the depth squared. So for every unit of additional height you want out of a gravity dam, you need significantly more weight to resist the forces, which means more material and thus a lot more cost.

Hopefully all this exposition is starting to reveal a solution to this rapid divergence of stability and loads as a reservoir increases in height. Dams don’t actually float in space like my demonstration and graphics show. You know, by necessity, they extend across the entire valley and usually key into the abutments on either side. Naturally, that connection at the sides is going to offer some resistance to the forces dams need to withstand. And if you can count on that resistance, you can significantly lower the mass, and thus the cost, of the structure. But, again, this gets complicated. Let’s go back to the demo.

Now I’m going to replace my gravity dam with something much simpler. Just a sheet of aluminum flashing, and, to simulate that resistance provided by socketing the structure into the earth, I’ve taped it to the bottom and sides… with some difficulty, actually. When I fill up the reservoir with water, it holds just fine. There’s a little leaking past my subpar tape job, but this is a fully stable structure. And I think the comparison here is pretty stark. When you can develop resistance from the sides you can get away with a lot less dam. But it’s harder than you might think to do that.

For one, the natural soil or rock at a dam site might not be all that strong. The banks of rivers aren’t generally known for their stability, so the prospect of transferring enormous amounts of force into them rarely makes a lot of engineering sense. But the other challenge is in the dam itself. Take a look back at this demo. See how my dam is bending behind the force of the water. It’s holding there, but, you know, we don’t actually build dams out of aluminum flashing. Resisting loads in this way basically treats the dam like a beam, like a sideways bridge girder. Except, unlike girder bridges that usually only span up to a few hundred feet, dams are often much longer. Even the stiffest modern materials, like prestressed concrete boxes, would just deflect too much under load to transfer all the hydrostatic pressure across a valley into the abutments. Plus we usually don’t like to rely on steel too much in dams because of issues with corrosion and longevity. So where a typical beam experiences both tensile and compressive stress on opposite sides, we really need to transfer all that load, creating only compressive stress in the material. I’m sure you see where I’m going with this.

How have we been building bridges for ages from materials like masonry where tensile stress isn’t an option? It’s arches! The arch is a special shape in engineering because you can transfer loads by putting the material in compression only, allowing for simpler, cheaper, and longer-lasting materials like masonry and concrete. You basically co-opt the geology for support, reducing the need for a massive structure. For completeness’s sake, let me show you how it works in the demo. I’ve formed a little arch from my thin sheet of aluminum. Now when I fill up the reservoir, there’s no deflection like the previous example. And again, side by side, it’s easy to see the benefits here. You get a lot more efficiency out of your materials than you do with an earthen embankment dam or a gravity structure.

Of course, there are some drawbacks here. For one, arches create horizontal forces at the supports called thrusts that have to be resisted. Sites that use this design really require strong, competent rock in the abutments to withstand the enormous loads. And just like with bridges, the span matters. The wider the valley, the bigger the arch needs to be, so these dams generally only make sense in deep gorges and steep, narrow canyons. The engineering is a lot more complicated, too. You can’t use a simple 2D cross-section to demonstrate stability. The structural behavior is inherently three-dimensional, which is tougher to characterize, especially when you consider unusual conditions like earthquakes and temperature effects. And since they’re lighter, arch dams don’t resist uplift forces very well, making foundation drainage systems more critical. All this means that it’s really only a solution that makes economic sense in a narrow range of circumstances, one of the most important being height.

For smaller dams, the additional complexity and expense of designing and building an arch aren’t justified by the structural efficiency. Gravity and embankment dams are much more adaptable to a wider range of site conditions. And there are other types of dams, too, that blend these ideas. Multiple-arch dams use a series of smaller arches supported by buttresses, dividing the span into more manageable components. Even what is perhaps the most famous arch dam in the world - Hoover Dam - isn’t a pure arch structure. Technically, it’s a gravity-arch dam, meaning it resists part of the water load through mass while also distributing the forces into the canyon through arch action. The proportions are carefully balanced to take advantage of the unique site conditions and relatively wider canyon than most arch dams are built in.

And so, when you look at the tallest dams on Earth, one structural form dominates. By my estimation, around 40 percent of the tallest 200 dams in the world incorporate an arch into their design. There aren’t that many places where it makes sense, but when you compare what it takes to hold a reservoir back in a narrow canyon valley, I think the case for arches is pretty clear.

The Hidden Engineering of Floating Bridges

[Note that this article is a transcript of the video embedded above.]

In the early 1900s, Seattle was a growing city hemmed in by geography. To the west was Puget Sound, a vital link to the Pacific Ocean. To the east, Lake Washington stood between the city and the farmland and logging towns of the Cascades. As the population grew, pressure mounted for a reliable east–west transportation route. But Lake Washington wasn’t easy to cross.

Carved by glaciers, the lake is deceptively deep, over 200 feet or 60 meters in some places. And under that deep water sits an even deeper problem: a hundred-foot layer of soft clay and mud. Building bridge piers all the way to solid ground would have required staggeringly sized supports. The cost and complexity made it infeasible to even consider.

But in 1921, an engineer named Homer Hadley proposed something radical: a bridge that didn’t rest on the bottom at all. Instead, it would float on massive hollow concrete pontoons, riding on the surface like a ship. It took nearly two decades for his idea to gain traction, but with the New Deal and Public Works Administration, new possibilities for transportation routes across the country began to open up. Federal funds flowed, and construction finally began on what would become the Lacey V. Murrow Bridge.

When it opened in 1940, it was the first floating concrete highway of its kind, a marvel of engineering and a symbol of ingenuity under constraint. But floating bridges, by their nature, carry some unique vulnerabilities. And fifty years later, this span would be swallowed by the very lake it crossed.

Between that time and since, the Seattle area has kind of become the floating concrete highway capital of the world. That’s not an official designation, at least not yet, but there aren’t that many of these structures around the globe. And four of the five longest ones on Earth are clustered in one small area of Washington state. You have Hood Canal, Evergreen Point, Lacey V Murrow, and its neighbor, the Homer M. Hadley Memorial Bridge, named for the engineer who floated the idea in the first place.

Washington has had some high-profile failures, but also some remarkable successes, including a test for light rail transit over a floating bridge just last month in June 2025. It's a niche branch of engineering, full of creative solutions and unexpected stories. So I want to take you on a little tour of the hidden engineering behind them. I’m Grady, and this is Practical Engineering.

Floating bridges are basically as old as recorded history. It’s not a complicated idea: place pontoons across a body of water, then span them with a deck. For thousands of years, this straightforward solution has provided a fast and efficient way to cross rivers and lakes, particularly in cases where permanent bridges were impractical or when the need for a crossing was urgent. In fact, floating bridges have been most widely used in military applications, going all the way back to Xerxes crossing the Dardanelles in 480 BCE. They can be made portable, quick to erect, flexible to a wide variety of situations, and they generally don’t require a lot of heavy equipment. There are countless designs that have been used worldwide in various military engagements.

But most floating bridges, both ancient and modern, weren’t meant to last. They’re quick to put up, but also quick to take out, either on purpose or by Mother Nature. They provide the means to get in, get across, and get out. So they aren’t usually designed for extreme conditions. Transitioning from temporary military crossings to permanent infrastructure was a massive leap, and it brought with it a host of engineering challenges.

An obvious one is navigation. A bridge that floats on the surface of the water is, by default, a barrier to boats. So, permanent floating bridges need to make room for maritime traffic. Designers have solved this in several ways, and Washington State offers a few good case studies.

The Evergreen Point Floating Bridge includes elevated approach spans on either end, allowing ships to pass beneath before the road descends to water level. The original Lacey V. Murrow Bridge took a different approach. Near its center, a retractable span could be pulled into a pocket formed by adjacent pontoons, opening a navigable channel. But, not only did the movable span create interruptions to vehicle traffic on this busy highway, it also created awkward roadway curves that caused frequent accidents. The mechanism was eventually removed after the East Channel Bridge was replaced to increase its vertical clearance, providing boats with an alternative route between the two sides of Lake Washington.

Further west, the Hood Canal Bridge incorporates truss spans for smaller craft. And it has hydraulic lift sections for larger ships. The US Naval Base Kitsap is not far away, so sometimes the bridge even has to open for Navy submarines. These movable spans can raise vertically above the pontoons, while adjacent bridge segments slide back underneath. The system is flexible: one side can be opened for tall but narrow vessels, or both for wider ships.

But floating bridges don’t just have to make room for boats. In a sense, they are boats. Many historical spans literally floated on boats lashed together. And that comes with its own complications. Unlike fixed structures, floating bridges are constantly interacting with water: waves, currents, and sometimes even tides and ice. They’re easiest to implement on calm lakes or rivers with minimal flooding, but water is water, and it’s a totally different type of engineering when you’re not counting on firm ground to keep things in place.

We don’t just stretch floating bridges across the banks and hope for the best. They’re actually moored in place, usually by long cables and anchors, to keep materials from overstressing and to prevent movements that would make the roadway uncomfortable or dangerous. Some anchors use massive concrete slabs placed on the lakebed. Others are tied to piles driven deep into the ground. In particularly deep water or soft soil, anchors are lowered to the bottom with water hoses that jet soil away, allowing the anchor to sink deep into the mud.

These anchoring systems do double duty, providing both structural integrity and day-to-day safety for drivers, but even with them, floating bridges have some unique challenges. They naturally sit low to the water, which means that in high winds, waves can crash directly onto the roadway, obscuring the visibility and creating serious risks to road users. Motion from waves and wind can also cause the bridge to flex and shift beneath vehicles, especially unnerving for drivers unused to the sensation. In Washington State, all the major floating bridges have been closed at various times due to weather. The DOT enforces wind thresholds for each bridge; if the wind exceeds the threshold, the bridge is closed to traffic. Even if the bridge is structurally sound, these closures reflect the reality that in extreme weather, the bridge itself becomes part of the storm.

But we still haven’t addressed the floating elephant in the pool here: the concrete pontoons themselves. Floating bridges have traditionally been made of wood or inflatable rubber, which makes sense if you’re trying to stay light and portable. But permanent infrastructure demands something more durable. It might seem counterintuitive to build a buoyant structure out of concrete, but it’s not as crazy as it sounds. In fact, civil engineering students compete every year in concrete canoe races hosted by the American Society of Civil Engineers.

Actually, I was doing a little recreational math to find a way to make this intuitive, and I stumbled upon a fun little fact. If you want to build a neutrally buoyant, hollow concrete cube, there’s a neat rule of thumb you can use. Just take the wall thickness in inches, and that’s your outer dimension in feet. Want 12-inch-thick concrete walls? You’ll need a roughly 12-foot cube. This is only fun because of the imperial system, obviously. It’s less exciting to say that the two dimensions have a roughly linear relationship with a factor of 12. And I guess it’s not really that useful except that it helps to visualize just how feasible it is to make concrete float.

Of course, real pontoons have to do more than just barely float themselves. They have to carry the weight of a deck and whatever crosses it with an acceptable margin of safety. That means they’re built much larger than a neutrally buoyant box. But mass isn’t the only issue. Concrete is a reliable material and if you’ve watched the channel for a while, you know that there are a few things you can count on concrete to do, and one of them is to crack. Usually not a big deal for a lot of structures, but that’s a pretty big problem if you’re trying to keep water out of a pontoon.

Designers put enormous effort into preventing leaks. Modern pontoons are subdivided into sealed chambers. Watertight doors are installed between the chambers so they can still be accessed and inspected. Leak detection systems provide early warnings if anything goes wrong. And piping is pre-installed with pumps on standby, so if a leak develops, the chambers can be pumped dry before disaster strikes. The concrete recipe itself gets extra attention. Specialized mixes reduce shrinkage, improve water resistance, and resist abrasion. Even temperature control during curing matters. For the replacement of the Evergreen Point Bridge, contractors embedded heating pipes in the base slabs of the pontoons, allowing them to match the temperature of the walls as they were cast. This enabled the entire structure to cool down at a uniform rate, reducing thermal stresses that could lead to cracking.

There were also errors during construction, though. A flaw in the post-tensioning system led to millions of dollars in change orders halfway through construction and delayed the project significantly while they worked out a repair. But there’s a good reason why they were so careful to get the designs right on that project. Of the four floating bridges in Washington state, two of them have sunk.

In February 1979, a severe storm caused the western half of the Hood Canal Bridge to lose its buoyancy. Investigations revealed that open hatches allowed rain and waves to blow in, slowly filling the pontoons and ultimately leading to the western half of the bridge sinking. The DOT had to establish a temporary ferry service across the canal for nearly four years while the western span was rebuilt.

Then, in 1990, it happened again. This time, the failure occurred during rehabilitation work on the Lacey V. Murrow Bridge while it was closed. Contractors were using hydrodemolition, high-pressure water jets, to remove old concrete from the road deck. Because the water was considered contaminated, it had to be stored rather than released into Lake Washington. Engineers calculated that the pontoon chambers could hold the runoff safely. To accommodate that, they removed the watertight doors that normally separated the internal compartments. But, when a storm hit over Thanksgiving weekend, water flooded into the open chambers. The bridge partially sank, severing cables on the adjacent Hadley Bridge and delaying the project by more than a year - a potent reminder that even small design or operational oversights can have major consequences on this type of structure.

And we still have a lot to learn. Recently, Sound Transit began testing light rail trains on the Homer Hadley Bridge, introducing a whole new set of engineering puzzles.

One is electricity. With power running through the rails, there was concern about stray currents damaging the bridge. To prevent this, the track is mounted on insulated blocks, with drip caps to prevent water from creating a conductive path.

And then there’s the bridge movement. Unlike typical bridges, a floating bridge can roll, pitch, and yaw with weather, lake level, and traffic loads. The joints between the fixed shoreline and the bridge have to be able to accommodate movement. It’s usually not an issue for cars, trucks, bikes, or pedestrians, but trains require very precise track alignment. Engineers had to develop an innovative “track bridge” system. It uses specialized bearings to distribute every kind of movement over a longer distance, keeping tracks aligned even as the floating structure shifts beneath it. Testing in June went well, but there’s more to be done before you can ride the Link light rail across a floating highway.

If floating bridges are the present, floating tunnels might be the future. I talked about immersed tube tunnels in a previous video. They’re used around the world, made by lowering precast sections to the seafloor and connecting them underwater. But what if, instead of resting on the bottom, those tunnels floated in the water column? It should be possible to suspend a tunnel with negative buoyancy using surface pontoons or even tether one with positive buoyancy to the bottom using anchors. In deep water, this could dramatically shorten tunnel lengths, reduce excavation costs, and minimize environmental impacts.

Norway has actually proposed such a tunnel across a fjord on its western coast, a project that, if realized, would be the first of its kind. Like floating bridges before it, this tunnel will face a long list of unknowns. But that’s the essence of engineering: meeting each challenge with solutions tailored to a specific place and need.

There aren’t many locations where floating infrastructure makes sense. The conditions have to be just right - calm waters, minimal ice, manageable tides. But where the conditions do allow, floating bridges and their hopefully future descendants open up new possibilities for connection, mobility, and engineering.

The Hidden Engineering of Liquid Dampers in Skyscrapers

[Note that this article is a transcript of the video embedded above.]

There’s a new trend in high-rise building design. Maybe you’ve seen this in your city. The best lots are all taken, so developers are stretching the limits to make use of space that isn’t always ideal for skyscrapers. They’re not necessarily taller than buildings of the past, but they are a lot more slender. “Pencil tower” is the term generally used to describe buildings that have a slenderness ratio of more than around 10 to 1, height to width. A lot of popular discussion around skyscrapers is about how tall we can build them. Eventually, you can get so tall that there are no materials strong enough to support the weight. But, pencil towers are the perfect case study in why strength isn’t the only design criterion used in structural engineering.

Of course, we don’t want our buildings to fall down, but there’s other stuff we don’t want them to do, too, including flex and sway in the wind. In engineering, this concept is called the serviceability limit state, and it’s an entirely separate consideration from strength. Even if moderate loads don’t cause a structure to fail, the movement they cause can lead to windows breaking, tiles cracking, accelerated fatigue of the structure, and, of course, people on the top floors losing their lunch from disorientation and discomfort. So, limiting wind-induced motions is a major part of high-rise design and, in fact, can be such a driving factor in the engineering of the building that strength is a secondary consideration.

Making a building stiffer is the obvious solution. But adding stiffness requires larger columns and beams, and those subtract valuable space within the building itself. Another option is to augment a building’s aerodynamic performance, reducing the loads that winds impose. But that too can compromise the expensive floorspace within. So many engineers are relying on another creative way to limit the vibrations of tall buildings. And of course, I built a model in the garage to show you how this works. I’m Grady, and this is Practical Engineering.

One of the very first topics I ever covered on this channel was tuned mass dampers. These are mechanisms that use a large, solid mass to counteract motion in all kinds of structures, dissipating the energy through friction or hydraulics, like the shock absorbers in vehicles. Probably the most famous of these is in the Taipei 101 building. At the top of the tower is a massive steel pendulum, and instead of hiding it away in a mechanical floor, they opened it to visitors, even giving the damper its own mascot. But, mass dampers have a major limitation because of those mechanical parts. The complex springs, dampers, and bearings need regular maintenance, and they are custom-built. That gets pretty expensive. So, what if we could simplify the device?

This is my garage-built high-rise. It’s not going to hold many conference room meetings, but it does do a good job swaying from side to side, just like an actual skyscraper. And I built a little tank to go on top here. The technical name for this tank is a tuned liquid column damper, and I can show you how it works. Let’s try it with no water first. Using my digitally calibrated finger, I push the tower over by a prescribed distance, and you can see this would not be a very fun ride. There is some natural damping, but the oscillation goes on for quite a while before the motion stops. Now, let’s put some water in the tank. With the power of movie magic, I can put these side by side so you can really get a sense of the difference.

By the way, nearly all of the parts for this demonstration were provided by my friends at Send-Cut-Send. I don’t have a milling machine or laser cutter, so this is a really nice option for getting customized parts made from basically any material - aluminum, steel, acrylic - that are ready to assemble.

Instead of complex mechanical devices, liquid column dampers dissipate energy through the movement of water. The liquid in the tank is both the mass and the damper. This works like a pendulum where the fluid oscillates between two columns. Normally, there’s an orifice between the two columns that creates the damping through friction loss as water flows from one side to the other. To make this demo a little simpler, I just put lids on the columns with small holes. I actually bought a fancy air valve to make this adjustable, but it didn’t allow quite enough airflow. So instead, I simplified with a piece of tape. Very technical. Energy transferred to the water through the building is dissipated by the friction of the air as it moves in and out of the columns. And you can even hear this as it happens.

Any supplemental damping system starts with a design criterion. This varies around the world, but in the US, this is probability-based. We generally require that peak accelerations with a 1-in-10 chance of being exceeded in a given year be limited to 15-18 milli-gs in residential buildings and 20-25 milli-gs in offices. For reference, the lateral acceleration for highway curve design is usually capped at 100 milli-gs, so the design criteria for buildings is between a fourth and a sixth of that. I think that makes intuitive sense. You don’t want to feel like you’re navigating a highway curve while you sit at your desk at work.

It’s helpful to think of these systems in a simplified way. This is the most basic representation: a spring, a damper, and mass on a cart. We know the mass of the building. We can estimate its stiffness. And the building itself has some intrinsic damping, but usually not much. If we add the damping system onto the cart, it’s basically just the same thing at a smaller scale, and the design process is really just choosing the mass and damping systems for the remaining pieces of this puzzle to achieve the design goal. The mass of liquid dampers is usually somewhere between half a percent to two percent of the building’s total weight. The damping is related to the water’s ability to dissipate energy. And the spring needs to be tuned to the building.

All buildings vibrate at a natural frequency related to their height and stiffness. Think of it like a big tuning fork full of offices or condos. I can estimate my model’s natural frequency by timing the number of oscillations in a given time interval. It’s about 1.3 hertz or cycles per second. In an ideal tuned damper, the oscillation of the damping system matches that of the building. So tuning the frequency of the damper is an important piece of the puzzle. For a tuned liquid column damper, the tuning mostly comes from the length of the liquid flow path. A longer path results in a lower frequency. The compression of the air above the column in my demo affects this too, and some types of dampers actually take advantage of that phenomenon. I got the best tuning when the liquid level was about halfway up the columns. The orifice has less of an effect on frequency and is used mostly to balance the amount of damping versus the volume of liquid that flows through each cycle. In my model, with one of the holes completely closed off, you can see the water doesn’t move, and you get minimal damping. With the tape mostly covering the hole, you get the most frictional loss, but not all the fluid flows from one side to the other each cycle. When I covered about half of one hole, I got the full fluid flow and the best damping performance.

The benefit of a tuned column damper is that it doesn’t take up a lot of space. And because the fluid movement is confined, they’re fairly predictable in behavior. So, these are used in quite a few skyscrapers, including the Random House Tower in Manhattan, One Wall Center in Vancouver (which actually has many walls), and Comcast Center in Philadelphia. But, tuned column liquid dampers have a few downsides. One is that they really only work for flexible structures, like my demo. Just like in a pendulum, the longer the flow path in a column damper, the lower the frequency of the oscillation. For stiffer buildings with higher natural frequencies, tuning requires a very short liquid column, which limits the mass and damping capability to a point where you don’t get much benefit. The other thing is that this is still kind of a complex device with intricate shapes and a custom orifice between the two columns. So, we can get even simpler.

This is my model tuned sloshing damper, and it’s about as simple as a damper can get. I put a weight inside the empty tank to make a fair comparison, and we can put it side by side with water in the tank to see how it works. As you can see, sloshing dampers dissipate energy by… sloshing. Again, the water is both the mass and the damper. If you tune it just right, the sloshing happens perfectly out of phase of the motion of the building, reducing the magnitude of the movement and acceleration. And you can see why this might be a little cheaper to build - it’s basically just a swimming pool - four concrete walls, a floor, and some water. There’s just not that much to it. But the simplicity of construction hides the complexity of design.

Like a column damper, the frequency of a sloshing damper can be tuned, first by the length of the tank. Just like fretting a guitar string further down the neck makes the note lower, a tank works the same way. As the tank gets longer, its sloshing frequency goes down. That makes sense - it takes longer for the wave to get from one side to the other. But you can also adjust the depth. Waves move slower in shallower water and faster in deeper water. Watch what happens when I overfill the tank.

The initial wave starts on the left as the building goes right. It reaches the right side just as the building starts moving left. That’s what we want; it’s counteracting the motion. But then it makes it back to the left before the building starts moving right. It’s actually kind of amplifying the motion, like pushing a kid on a swing. Pretty soon after that, the wave and the building start moving in phase, so there’s pretty much no damping at all. Compare it to the more properly tuned example where most of the wave motion is counteracting the building motion as it sways back and forth.

You can see in my demo that a lot of the energy dissipation comes from the breaking waves as they crash against the sides of the tank. That is a pretty complicated phenomenon to predict, and it’s highly dependent on how big the waves are. And even with the level pretty well tuned to the frequency of the building, you can see there’s a lot of complexity in the motion with multiple modes of waves, and not all of them acting against the motion of the building. So, instead of relying on breaking waves, most sloshing dampers use flow obstructions like screens, columns, or baffles. I got a few different options cut out of acrylic so we can try this out. These baffles add drag, increasing the energy dissipation with the water, usually without changing the sloshing frequency.

Here’s a side-by-side comparison of the performance without a baffle and with one. You can see that the improvement is pretty dramatic. The motion is more controlled and the behavior is more linear, making this much simpler to predict during the design phase. It’s kind of the best of both worlds since you get damping from the sloshing and the drag of the water passing through the screen. Almost all the motion is stopped in this demo after only three oscillations. I was pretty impressed with this. Here’s all three of the baffle runs side by side. Actually, the one with the smallest holes worked the best in my demo, but deciding the configuration of these baffles is a big challenge in the engineering of these systems because you can’t really just test out a bunch of options at full scale.

Devices like this are in service in quite a few high-rise buildings, including Princess Tower in Dubai, and the Museum Tower in Dallas. With no moving parts and very little maintenance except occasionally topping it off to keep the water at the correct level, you can see how it would be easy to choose a sloshing damper for a new high-rise project. But there are some disadvantages. One is volumetric efficiency. You can see that not all the water in the tank is mobilized, especially for smaller movements, which means not all the water is contributing to the damping. The other is non-linearity. The amount of damping changes depending on the magnitude of the movement since drag is related to velocity squared. And even the frequency of the damper isn’t constant; it can change with the wave amplitude as well because of the breaking waves. So you might get good performance at the design level, but not so much for slower winds.

Dampers aren’t just used in buildings. Bridges also take advantage of these clever devices, especially on the decks of pedestrian bridges and the towers of long-span bridges. This also happens at a grand scale between the Earth and moon. Tidal bulges in the oceans created by the moon’s tug on Earth dissipate energy through friction and turbulence, which is a big part of why our planet’s rotation is slowing over time. Days used to be a lot shorter when the Earth was young, but we have a planet-scale liquid damper constantly dissipating our rotational energy.

But whether it’s bridges or buildings, these dampers usually don’t work perfectly right at the start. Vibrations are complicated. They’re very hard to predict, even with modern tools like simulation software and scale physical models. So, all dampers have to go through a commissioning process. Usually this involves installing accelerometers once construction is nearing completion to measure the structure’s actual natural frequency. The tuning of tuned dampers doesn’t just happen during the design phase; you want some adjustability after construction to make sure they match the structure’s natural frequency exactly so you get the most damping possible. For liquid dampers, that means adjusting the levels in the tanks. And in many cases, buildings might use multiple dampers tuned to slightly different frequencies to improve the performance over a range of conditions. Even in these two basic categories, there is a huge amount of variability and a lot of ongoing research to minimize the tradeoffs these systems come with.

The truth is that, relatively speaking, there aren’t that many of these systems in use around the world. Each one is highly customized, and even putting them into categories can get a little tricky. There are even actively controlled liquid dampers. My tuning for the column damper works best for a single magnitude of motion, but you can see that once the swaying gets smaller, the damper isn’t doing a lot to curb it. You can imagine if I constantly adjusted the size of the orifice, I could get better performance over a broader range of unwanted motion. You can do this electronically by having sensors feed into a control system that adjusts a valve position in real-time. Active systems and just the flexibility to tune a damper in general also help deal with changes over time. If a building’s use changes, if new skyscrapers nearby change the wind conditions, or if it gets retrofits that change its natural frequency, the damping system can easily accommodate those changes.

In the end, a lot of engineering decisions come down to economics. In most cases, damping is less about safety and more about comfort, which is often harder to pin down. Engineers and building owners face a balancing act between the cost of supplemental damping and the value of the space those systems take up. Tuned mass dampers are kind of household names when it comes to damping. A few buildings like Shanghai Center and Taipei 101 have made them famous. They’re usually the most space-efficient (since steel and concrete are more dense than water). But they’re often more costly to install and maintain. Liquid dampers are the unsung heroes. They take up more space, but they’re simple and cost-effective, especially if the fire codes already require you to have a big tank of water at the top of your building anyway. Maybe someday, an architect will build one out of glass or acrylic, add some blue dye and mica powder, and put it on display as a public showcase. Until then, we’ll just have to know it’s there by feel.

How Sewage Recycling Works

[Note that this article is a transcript of the video embedded above.]

Wichita Falls, Texas, went through the worst drought in its history in 2011 and 2012. For two years in a row, the area saw its average annual rainfall roughly cut in half, decimating the levels in the three reservoirs used for the city’s water supply. Looking ahead, the city realized that if the hot, dry weather continued, they would be completely out of water by 2015. Three years sounds like a long runway, but when it comes to major public infrastructure projects, it might as well be overnight. Between permitting, funding, design, and construction, three years barely gets you to the starting line. So the city started looking for other options. And they realized there was one source of water nearby that was just being wasted - millions of gallons per day just being flushed down the Wichita River. I’m sure you can guess where I’m going with this. It was the effluent from their sewage treatment plant.

The city asked the state regulators if they could try something that had never been done before at such a scale: take the discharge pipe from the wastewater treatment plant and run it directly into the purification plant that produces most of the city’s drinking water. And the state said no. So they did some more research and testing and asked again. By then, the situation had become an emergency. This time, the state said yes. And what happened next would completely change the way cities think about water. I’m Grady and this is Practical Engineering.

You know what they say, wastewater happens. It wasn’t that long ago that raw sewage was simply routed into rivers, streams, or the ocean to be carried away. Thankfully, environmental regulations put a stop to that, or at least significantly curbed the amount of wastewater being set loose without treatment. Wastewater plants across the world do a pretty good job of removing pollutants these days. In fact, I have a series of videos that go through some of the major processes if you want to dive deeper after this. In most places, the permits that allow these plants to discharge set strict limits on contaminants like organics, suspended solids, nutrients, and bacteria. And in most cases, they’re individualized. The permit limits are based on where the effluent will go, how that water body is used, and how well it can tolerate added nutrients or pollutants. And here’s where you start to see the issue with reusing that water: “clean enough” is a sliding scale.

Depending on how water is going to be used or what or who it’s going to interact with, our standards for cleanliness vary. If you have a dog, you probably know this. They should drink clean water, but a few sips of a mud puddle in a dirty street, and they’re usually just fine. For you, that might be a trip to the hospital. Natural systems can tolerate a pretty wide range of water quality, but when it comes to drinking water for humans, it should be VERY clean. So the easiest way to recycle treated wastewater is to use it in ways that don’t involve people. That idea’s been around for a while.

A lot of wastewater treatment plants apply effluent to land as a disposal method, avoiding the need for discharge to a natural water body. Water soaks into the ground, kind of like a giant septic system. But that comes with some challenges. It only works if you’ve got a lot of land with no public access, and a way to keep the spray from drifting into neighboring properties. Easy at a small scale, but for larger plants, it just isn’t practical engineering. Plus, the only benefits a utility gets from the effluent are some groundwater recharge and maybe a few hay harvests per season. So, why not send the effluent to someone else who can actually put it to beneficial use?

If only it were that simple. As soon as a utility starts supplying water to someone else, things get complicated because you lose a lot of control over how the effluent is used. Once it's out of your hands, so to speak, it’s a lot harder to make sure it doesn’t end up somewhere it shouldn’t, like someone’s mouth. So, naturally, the permitting requirements become stricter. Treatment processes get more complicated and expensive. You need regular monitoring, sampling, and laboratory testing. In many places in the world, reclaimed water runs in purple pipes so that someone doesn’t inadvertently connect to the lines thinking they’re potable water. In many cases, you need an agreement in place with the end user, making sure they’re putting up signs, fences, and other means of keeping people from drinking the water. And then you need to plan for emergencies - what to do if a pipe breaks, if the effluent quality falls below the standards, or if a cross-connection is made accidentally.

It’s a lot of work - time, effort, and cost - to do it safely and follow the rules. And those costs have to be weighed against the savings that reusing water creates. In places that get a lot of rain or snow, it’s usually not worth it. But in many US states, particularly those in the southwest, this is a major strategy to reduce the demand on fresh water supplies. Think about all the things we use water for where its cleanliness isn’t that important. Irrigation is a big one - crops, pastures, parks, highway landscaping, cemeteries - but that’s not all. Power plants use huge amounts of water for cooling. Street sweeping, dust control. In nearly the entire developed world, we use drinking-quality water to flush toilets!

You can see where there might be cases where it makes good sense to reclaim wastewater, and despite all the extra challenges, its use is fairly widespread. One of the first plants was built in 1926 at Grand Canyon Village which supplied reclaimed water to a power plant and for use in steam locomotives. Today, these systems can be massive, with miles and miles of purple pipes run entirely separate from the freshwater piping. I’ve talked about this a bit on the channel before. I used to live near a pair of water towers in San Antonio that were at two different heights above ground. That just didn’t make any sense until I realized they weren’t connected; one of them was for the reclaimed water system that didn’t need as much pressure in the lines. Places like Phoenix, Austin, San Antonio, Orange County, Irvine, and Tampa all have major water reclamation programs. And it’s not just a US thing. Abu Dhabi, Beijing, and Tel Aviv all have infrastructure to make beneficial use of treated municipal wastewater, just to name a few.

Because of the extra treatment and requirements, many places put reclaimed water in categories based on how it gets used. The higher the risk of human contact, the tighter the pollutant limits get. For example, if a utility is just selling effluent to farmers, ranchers, or for use in construction, exposure to the public is minimal. Disinfecting the effluent with UV or chlorine may be enough to meet requirements. And often that’s something that can be added pretty simply to an existing plant. But many reclaimed water users are things like golf courses, schoolyards, sports fields, and industrial cooling towers, where people are more likely to be exposed. In those cases, you often need a sewage plant specifically designed for the purpose or at least major upgrades to include what the pros call tertiary treatment processes - ways to target pollutants we usually don’t worry about and improve the removal rates of the ones we do. These can include filters to remove suspended solids, chemicals that bind to nutrients, and stronger disinfection to more effectively kill pathogens.

This creates a conundrum, though. In many cases, we treat wastewater effluent to higher standards than we normally would in order to reclaim it, but only for nonpotable uses, with strict regulations about human contact. But if it’s not being reclaimed, the quality standards are lower, and we send it downstream. If you know how rivers work, you probably see the inconsistency here. Because in many places, down the river, is the next city with its water purification plant whose intakes, in effect, reclaim that treated sewage from the people upstream.

This isn’t theoretical - it’s just the reality of how humans interact with the water cycle. We’ve struggled with the problems it causes for ages. In 1906, Missouri sued Illinois in the Supreme Court when Chicago reversed their river, redirecting its water (and all the city’s sewage) toward the Mississippi River. If you live in Houston, I hate to break it to you, but a big portion of your drinking water comes from the flushes and showers in Dallas. There have been times when wastewater effluent makes up half of the flow in the Trinity River.

But the question is: if they can do it, why can’t we? If our wastewater effluent is already being reused by the city downstream to purify into drinking water, why can’t we just keep the effluent for ourselves and do the same thing? And the answer again is complicated.

It starts with what’s called an environmental buffer. Natural systems offer time to detect failures, dilute contaminants, and even clean the water a bit—sunlight disinfects, bacteria consume organic matter. That’s the big difference in one city, in effect, reclaiming water from another upstream. There’s nature in between. So a lot of water reclamation systems, called indirect potable reuse, do the same thing: you discharge the effluent into a river, lake, or aquifer, then pull it out again later for purification into drinking water. By then, it’s been diluted and treated somewhat by the natural systems.

Direct potable reuse projects skip the buffer and pipe straight from one treatment plant to the next. There’s no margin for error provided by the environmental buffer. So, you have to engineer those same protections into the system: real-time monitoring, alarms, automatic shutdowns, and redundant treatment processes.

Then there’s the issue of contaminants of emerging concern: pharmaceuticals, PFAS [P-FAS], personal care products - things that pass through people or households and end up in wastewater in tiny amounts. Individually, they’re in parts per billion or trillion. But when you close the loop and reuse water over and over, those trace compounds can accumulate. Many of these aren’t regulated because they’ve never reached concentrations high enough to cause concern, or there just isn’t enough knowledge about their effects yet. That’s slowly changing, and it presents a big challenge for reuse projects. They can be dealt with at the source by regulating consumer products, encouraging proper disposal of pharmaceuticals (instead of flushing them), and imposing pretreatment requirements for industries. It can also happen at the treatment plant with advanced technologies like reverse osmosis, activated carbon, advanced oxidation, and bio-reactors that break down micro-contaminants. Either way, it adds cost and complexity to a reuse program. But really, the biggest problem with wastewater reuse isn’t technical - it’s psychological.

The so-called “yuck factor” is real. People don’t want to drink sewage. Indirect reuse projects have a big benefit here. With some nature in between, it’s not just treated wastewater; it’s a natural source of water with treated wastewater in it. It’s kind of a story we tell ourselves, but we lose the benefit of that with direct reuse: Knowing your water came from a toilet—even if it’s been purified beyond drinking water standards—makes people uneasy. You might not think about it, but turning the tap on, putting that water in a glass, and taking a drink is an enormous act of trust. Most of us don’t understand water treatment and how it happens at a city scale. So that trust that it’s safe to drink largely comes from seeing other people do it and past experience of doing it over and over and not getting sick. The issue is that, when you add one bit of knowledge to that relative void of understanding - this water came directly from sewage - it throws that trust off balance. It forces you not to rely not on past experience but on the people and processes in place, most of which you don’t understand deeply, and generally none of which you can actually see. It’s not as simple as just revulsion. It shakes up your entire belief system.

And there’s no engineering fix for that. Especially for direct potable reuse, public trust is critical. So on top of the infrastructure, these programs also involve major public awareness campaigns. Utilities have to put themselves out there, gather feedback, respond to questions, be empathetic to a community’s values, and try to help people understand how we ensure water quality, no matter what the source is.

But also, like I said, a lot of that trust comes from past experience. Not everyone can be an environmental engineer or licensed treatment plant operator. And let’s be honest - utilities can’t reach everyone. How many public meetings about water treatment have you ever attended? So, in many places, that trust is just going to have to be built by doing it right, doing it well, and doing it for a long time. But, someone has to be first.

In the U.S., at least on the city scale, that drinking water guinea pig was Wichita Falls. They launched a massive outreach campaign, invited experts for tours, and worked to build public support. But at the end of the day, they didn’t really have a choice. The drought really was that severe. They spent nearly four years under intense water restrictions. Usage dropped to a third of normal demand, but it still wasn’t enough.

So, in collaboration with state regulators, they designed an emergency direct potable reuse system. They literally helped write the rules as they went, since no one had ever done it before. After two months of testing and verification, they turned on the system in July 2014. It made national headlines.

The project ran for exactly one year. Then, in 2015, a massive flood ended the drought and filled the reservoirs in just three weeks. The emergency system was always meant to be temporary. Water essentially went through three treatment plants: the wastewater plant, a reverse osmosis plant, and then the regular water purification plant. That’s a lot of treatment, which is a lot of expense, but they needed to have the failsafe and redundancy to get the state on board with the project. The pipe connecting the two plants was above ground and later repurposed for the city’s indirect potable reuse system, which is still in use today.

In the end, they reclaimed nearly two billion gallons of wastewater as drinking water. And they did it with 100% compliance with the standards. But more importantly, they showed that it could be done, essentially unlocking a new branch on the skill tree of engineering that other cities can emulate and build on.

Why are Smokestacks So Tall?

[Note that this article is a transcript of the video embedded above.]

“The big black stacks of the Illium Works of the Federal Apparatus Corporation spewed acid fumes and soot over the hundreds of men and women who were lined up before the red-brick employment office.” That’s the first line of one of my favorite short stories, written by Kurt Vonnegut in 1955. It paints a picture of a dystopian future that, thankfully, didn’t really come to be, in part because of those stacks.

In some ways, air pollution is kind of a part of life. I’d love to live in a world where the systems, materials and processes that make my life possible didn’t come with any emissions, but it’s just not the case... From the time that humans discovered fire, we’ve been methodically calculating the benefits of warmth, comfort, and cooking against the disadvantages of carbon monoxide exposure and particulate matter less than 2.5 microns in diameter… Maybe not in that exact framework, but basically, since the dawn of humanity, we’ve had to deal with smoke one way or another.

Since, we can’t accomplish much without putting unwanted stuff into the air, the next best thing is to manage how and where it happens to try and minimize its impact on public health. Of course, any time you have a balancing act between technical issues, the engineers get involved, not so much to help decide where to draw the line, but to develop systems that can stay below it. And that’s where the smokestack comes in. Its function probably seems obvious; you might have a chimney in your house that does a similar job. But I want to give you a peek behind the curtain into the Illium Works of the Federal Apparatus Corporation of today and show you what goes into engineering one of these stacks at a large industrial facility. I’m Grady, and this is Practical Engineering.

We put a lot of bad stuff in the air, and in a lot of different ways. There are roughly 200 regulated hazardous air pollutants in the United States, many with names I can barely pronounce. In many cases, the industries that would release these contaminants are required to deal with them at the source. A wide range of control technologies are put into place to clean dangerous pollutants from the air before it’s released into the environment. One example is coal-fired power plants. Coal, in particular, releases a plethora of pollutants when combusted, so, in many countries, modern plants are required to install control systems. Catalytic reactors remove nitrous oxides. Electrostatic precipitators collect particulates. Scrubbers use lime (the mineral, not the fruit) to strip away sulfur dioxide. And I could go on. In some cases, emission control systems can represent a significant proportion of the costs involved in building and operating a plant. But these primary emission controls aren’t always feasible for every pollutant, at least not for 100 percent removal.

There’s a very old saying that “the solution to pollution is dilution.” It’s not really true on a global scale. Case in point: There’s no way to dilute the concentration of carbon dioxide in the atmosphere, or rather, it’s already as dilute as it’s going to get. But, it can be true on a local scale. Many pollutants that affect human health and the environment are short-lived; they chemically react or decompose in the atmosphere over time instead of accumulating indefinitely. And, for a lot of chemicals, there are concentration thresholds below which the consequences on human health are negligible. In those cases, dilution, or really dispersion, is a sound strategy to reduce their negative impacts, and so, in some cases, that’s what we do, particularly at major point sources like factories and power plants.

One of the tricks to dispersion is that many plumes are naturally buoyant. Naturally, I’m going to use my pizza oven to demonstrate this. Not all, but most pollutants we care about are a result of combustion; burning stuff up. So the plume is usually hot. We know hot air is less dense, so it naturally rises. And the hotter it is, the faster that happens. You can see when I first start the fire, there’s not much air movement. But as the fire gets hotter in the oven, the plume speeds up, ultimately rising higher into the air. That’s the whole goal: get the plume high above populated areas where the pollutants can be dispersed to a minimally-harmful concentration. It sounds like a simple solution - just run our boilers and furnaces super hot to get enough buoyancy for the combustion products to disperse.

The problem with the solution is that the whole reason we combust things is usually to recover the heat. So if you’re sending a lot of that heat out of the system, just because it makes the plume disperse better, you’re losing thermodynamic efficiency. It’s wasteful. That’s where the stack comes in. Let me put mine on and show you what I mean. I took some readings with the anemometers with the stack on and off. The airspeed with the stack on was around double with it off. About a meter per second compared with two. But it’s a little tougher to understand why.

It’s intuitive that as you move higher in a column of fluid, the pressure goes down (since there’s less weight of the fluid above). The deeper you dive in a pool, the more pressure you feel. The higher you fly in a plane or climb a mountain, the lower the pressure. The slope of that line is proportional to a fluid’s density. You don’t feel much of a pressure difference climbing a set of stairs because air isn’t very dense. If you travel the same distance in water, you’ll definitely notice the difference. So let’s look at two columns of fluid. One is the ambient air and the other is the air inside a stack. Since it’s hotter, the air inside the stack is less dense. Both columns start at the same pressure at the bottom, but the higher you go, the more the pressure diverges.

It’s kind of like deep sea diving in reverse. In water, the deeper you go into the dense water, the greater the pressure you feel. In a stack, the higher you are in a column of hot air, the more buoyant you feel compared to the outside air. This is the genius of a smoke stack. It creates this difference in pressure between the inside and outside that drives greater airflow for a given temperature. Here’s the basic equation for a stack effect. I like to look at equations like this divided into what we can control and what we can’t. We don’t get to adjust the atmospheric pressure, the outside temperature, and this is just a constant. But you can see, with a stack, an engineer now has two knobs to turn: the temperature of the gas inside and the height of the stack.

I did my best to keep the temperature constant in my pizza oven and took some airspeed readings. First with no stack. Then with the stock stack. Then with a megastack. By the way, this melted my anemometer; should have seen that coming. Thankfully, I got the measurements before it melted. My megastack nearly doubled the airspeed again at around three-and-a-half meters per second versus the two with just the stack that came with the oven. There’s something really satisfying about this stack effect to me. No moving parts or fancy machinery. Just put a longer pipe and you’ve fundamentally changed the physics of the whole situation. And it’s a really important tool in the environmental engineer’s toolbox to increase airflow upward, allowing contaminants to flow higher into the atmosphere where they can disperse. But this is not particularly revolutionary… unless you’re talking about the Industrial Revolution.

When you look at all the pictures of the factories in the 19th century, those stacks weren’t there to improve air quality, if you can believe it. The increased airflow generated by a stack just created more efficient combustion for the boilers and furnaces. Any benefits to air quality in the cities were secondary. With the advent of diesel and electric motors, we could use forced drafts, reducing the need for a tall stack to increase airflow. That was kind of the decline of the forests of industrial chimneys that marked the landscape in the 19th century. But they’re obviously not all gone, because that secondary benefit of air quality turned into the primary benefit as environmental rules about air pollution became stricter.

Of course, there are some practical limits that aren’t taken into account by that equation I showed. The plume cools down as it moves up the stack to the outside, so its density isn’t constant all the way up. I let my fire die down a bit so it wouldn’t melt the thermometer (learned my lesson), and then took readings inside the oven and at the top of the stack. You can see my pizza oven flue gas is around 210 degrees at the top of the mega-stack, but it’s roughly 250 inside the oven. After the success of the mega stack on my pizza oven, I tried the super-mega stack with not much improvement in airflow: about 4 meters per second. The warm air just got too cool by the time it reached the top. And I suspect that frictional drag in the longer pipe also contributed to that as well. So, really, depending on how insulating your stack is, our graph of height versus pressure actually ends up looking like this. And this can be its own engineering challenge. Maybe you’ve gotten back drafts in your fireplace at home because the fire wasn’t big or hot enough to create that large difference in pressure.

You can see there are a lot of factors at play in designing these structures, but so far, all we’ve done is get the air moving faster. But that’s not the end goal. The purpose is to reduce the concentration of pollutants that we’re exposed to. So engineers also have to consider what happens to the plume once it leaves the stack, and that’s where things really get complicated.

In the US, we have National Ambient Air Quality Standards that regulate six so-called “criteria” pollutants that are relatively widespread: carbon monoxide, lead, nitrogen dioxide, ozone, particulates, and sulfur dioxide. We have hard limits on all these compounds with the intention that they are met at all times, in all locations, under all conditions. Unfortunately, that’s not always the case. You can go on EPA’s website and look at the so-called “non-attainment” areas for the various pollutants. But we do strive to meet the standards through a list of measures that is too long to go into here. And that is not an easy thing to do.

Not every source of pollution comes out of a big stationary smokestack where it’s easy to measure and control. Cars, buses, planes, trucks, trains, and even rockets create lots of contaminants that vary by location, season, and time of day. And there are natural processes that contribute as well. Forests and soil microbes release volatile organic compounds that can lead to ozone formation. Volcanic eruptions and wildfires release carbon monoxide and sulfur dioxide. Even dust storms put particulates in the air that can travel across continents. And hopefully you’re seeing the challenge of designing a smoke stack. The primary controls like scrubbers and precipitators get most of the pollutants out, and hopefully all of the ones that can’t be dispersed. But what’s left over and released has to avoid pushing concentrations above the standards. That design has to work within the very complicated and varying context of air chemistry and atmospheric conditions that a designer has no control over.

Let me show you a demo. I have a little fog generator set up in my garage with a small fan simulating the wind. This isn’t a great example because the airflow from the fan is pretty turbulent compared to natural winds. You occasionally get some fog at the surface, but you can see my plume mainly stays above the surface, dispersing as it moves with the wind. But watch what happens when I put a building downstream. The structure changes the airflow, creating a downwash effect and pulling my plume with it. Much more frequently you see the fog at the ground level downstream. And this is just a tiny example of how complex the behavior of these plumes can be. Luckily, there’s a whole field of engineering to characterize it.

There are really just two major transport processes for air pollution. Advection describes how contaminants are carried along by the wind. Diffusion describes how those contaminants spread out through turbulence. Gravity also affects air pollution, but it doesn’t have a significant effect except on heavier-than-air particulates. With some math and simplifications of those two processes, you can do a reasonable job predicting the concentration of any pollutant at any point in space as it moves and disperses through the air. Here’s the basic equation for that, and if you’ll join me for the next 2 hours, we’ll derive this and learn the meaning of each term… Actually, it might take longer than that, so let’s just look at a graphic. You can see that as the plume gets carried along by the wind, it spreads out in what’s basically a bell curve, or gaussian distribution, in the planes perpendicular to the wind direction.

But even that is a bit too simplified to make any good decisions with, especially when the consequences of getting it wrong are to public health. A big reason for that is atmospheric stability. And this can make things even more complicated, but I want to explain the basics, because the effect on plumes of gas can be really dramatic. You probably know that air expands as it moves upward; there’s less pressure as you go up because there is less air above you. And as any gas expands, it cools down. So there’s this relationship between height and temperature we call the adiabatic lapse rate. It’s about 10 degrees Celsius for every kilometer up or about 28 Fahrenheit for every mile up. But the actual atmosphere doesn’t always follow this relationship. For example, rising air parcels can cool more slowly than the surrounding air. This makes them warmer and less dense, so they keep rising, promoting vertical motion in a positive feedback loop called atmospheric instability. You can even get a temperature inversion where you have cooler air below warmer air, something that can happen in the early morning when the ground is cold. And as the environmental lapse rate varies from the adiabatic lapse rate, the plumes from stacks change.

In stable conditions, you usually get a coning plume, similar to what our gaussian distribution from before predicts. In unstable conditions, you get a lot of mixing, which leads to a looping plume. And things really get weird for temperature inversions because they basically act like lids for vertical movement. You can get a fanning plume that rises to a point, but then only spreads horizontally. You can also get a trapping plume, where the air gets stuck between two inversions. You can have a lofting plume, where the air is above the inversion with stable conditions below and unstable conditions above. And worst of all, you can have a fumigating plume when there are unstable conditions below an inversion, trapping and mixing the plume toward the ground surface. And if you pay attention to smokestacks, fires, and other types of emissions, you can identify these different types of plumes pretty easily.

Hopefully you’re seeing now how much goes into this. Engineers have to keep track of the advection and diffusion, wind speed and direction, atmospheric stability, the effects of terrain and buildings on all those factors, plus the pre-existing concentrations of all the criteria pollutants from other sources, which vary in time and place. All that to demonstrate that your new source of air pollution is not going to push the concentrations at any place, at any time, under any conditions, beyond what the standards allow. That’s a tall order, even for someone who loves gaussian distributions. And often the answer to that tall order is an even taller smokestack. But to make sure, we use software. The EPA has developed models that can take all these factors into account to simulate, essentially, what would happen if you put a new source of pollution into the world and at what height.

So why are smokestacks so tall? I hope you’ll agree with me that it turns out to be a pretty complicated question. And it’s important, right? These stacks are expensive to build and maintain. Those costs trickle down to us through the costs of the products and services we buy. They have a generally negative visual impact on the landscape. And they have a lot of other engineering challenges too, like resonance in the wind. And on the other hand, we have public health, arguably one of the most critical design criteria that can exist for an engineer. It’s really important to get this right. I think our air quality regulations do a lot to make sure we strike a good balance here. There are even rules limiting how much credit you can get for building a stack higher for greater dispersion to make sure that we’re not using excessively tall stacks in lieu of more effective, but often more expensive, emission controls and strategies.

In a perfect world, none of the materials or industrial processes that we rely on would generate concentrated plumes of hazardous gases. We don’t live in that perfect world, but we are pretty fortunate that, at least in many places on Earth, air quality is something we don’t have to think too much about. And to thank for it, we have a relatively small industry of environmental professionals who do think about it, a whole lot. You know, for a lot of people, this is their whole career; what they ponder from 9-5 every day. Something most of us would rather keep out of mind, they face it head-on, developing engineering theories, professional consensus, sensible regulations, modeling software, and more - just so we can breathe easy.

The Most Implausible Tunneling Method

[Note that this article is a transcript of the video embedded above.]

The original plan to get I-95 over the Baltimore Harbor was a double-deck bridge from Fort McHenry to Lazaretto Point. The problem with the plan was this: the bridge would have to be extremely high so that large ships could pass underneath, dwarfing and overshadowing one of the US’s most important historical landmarks. Fort McHenry famously repelled a massive barrage and attack from the British Navy in the War of 1812, and inspired what would later become the national anthem. An ugly bridge would detract from its character, and a beautiful one would compete for it. So they took the high road by building a low road and decided to go underneath the harbor instead. Rather than bore a tunnel through the soil and rock below like the Channel Tunnel, the entire thing was prefabricated in sections and installed from the water surface above - a construction technique called immersed tube tunneling.

This seems kind of simple at first, but the more you think about it, the more you realize how complicated it actually is to fabricate tunnel sections the length of a city block, move them into place, and attach them together so watertight and safe that, eventually, you can drive or take a train from one side to the other. Immersed tube construction makes tunneling less like drilling a hole and more like docking a spacecraft. Materials and practices vary across the world, but I want to try and show you, at least in a general sense, how this works. I’m Grady, and this is Practical Engineering.

One of the big problems with bridges over navigable waterways is that they have to be so tall. Building high up isn’t necessarily the challenge; it’s getting up and back down. There are limits to how steep a road can be for comfort, safety, and efficiency, and railroads usually have even stricter constraints on grade. That means the approaches to high bridges have to be really long, increasing costs and, in dense cities, taking up more valuable space. This is one of the ways that building a tunnel can be a better option; They greatly reduce the amount of land at the surface needed for approaches. But traditional tunnels built using boring have to be installed somewhat deep into the ground, maintaining significant earth between the roof of the tunnel and the water for stability and safety. Since they’re installed from above, immersed tube tunnels don’t have the same problem. It’s basically a way to get the shortest tunnel possible for a given location, which often means the cheapest tunnel too. That’s a big deal, because tunnels are just about the most expensive way to get from point A to point B. Anything you can do to reduce their size goes a long way.

And there are other advantages too. Tunnel boring machines make one shape: a circle. It’s not the best shape for a tunnel, in a lot of ways. Often there’s underutilized space at the top and bottom - excavation you had to perform because of the machinery that is mostly just a waste. Immersed tubes can be just about any shape you need, making them ideal for wider tunnels like combined road and rail routes where a circular cross-section isn’t a good fit.

One of the other benefits of immersed tubes is that most of the construction happens on dry land. I probably don’t have to say this, but building stuff while underground or underwater is complex and difficult work. It requires specialty equipment, added safety measures, and a lot of extra expense. Immersed tube sections are built in dry docks or at a shipyard where it's much easier to deliver materials and accomplish the bulk of the actual construction work.

Once tunnel sections are fabricated, they have to be moved into place, and I think this is pretty clever. These sections can be enormous - upwards of 650 feet or 200 meters long. But they’re still mostly air. So if you put a bulkhead on either side to trap that air inside, they float. You can just flood the dry dock, hook up some tugboats, and tow them out like a massive barge. Interestingly, the transportation method means that the tunnel segments have to be designed to work as a watercraft first. The weight, buoyancy, and balance of each section are engineered to keep them stable in the water and avoid tipping or rolling before they have to be stable as a structure.

Once in place, a tunnel segment is handed over to the apparatus that will set it into place. In most cases, this is a catamaran-style behemoth called a lay barge. Two working platforms are connected by girders, creating a huge floating gantry crane. Internal tanks are filled with water to act as ballast, allowing the segment to sink. But when it gets to the bottom, it doesn’t just sit on the sea or channel floor below. And this is another benefit of immersed tube construction.

Especially in navigable waterways, you need to protect a tunnel from damage from strong currents, curious sea life, and ship anchors. So most immersed tube tunnels sit in a shallow trench, excavated using a clamshell or suction dredger. Most waterways have a thick layer of soft sediment at the surface - not exactly ideal as a foundation. This is another reason most boring machines have to be in deeper material. Drilling through soft sediment is prone to problems. Imagine using a power drill to make a nice, clean hole through pudding. But, at least in part due to being full of buoyant air, immersed tubes aren’t that heavy; in fact, in most cases, they’re lighter than the soil that was there in the first place, so the soft sediment really isn’t a problem. You don’t need a complicated foundation. In many cases, it’s just a layer of rock or gravel placed at the bottom of the trench, usually using a fall pipe (like a big garden hose for gravel) to control the location. This layer is then carefully leveled using a steel screed that is dragged over the top like an underwater bulldozer. Even in deep water, the process can achieve a remarkably accurate surface level for the tunnel segments to rest on.

The lowering process is the most delicate and important part of construction. The margins are tight because any type of misalignment may make it impossible for the segment to seal against its neighbor. Normally, you’d really want to take your time with this kind of thing, but here, the work usually has to happen in a narrow window to avoid weather, tides, and disruption to ship traffic. The tunnel section is fitted with rubber seals around its face, creating a gasket. Sometimes, the segment will also have a surveying tower that pokes above the water surface, allowing for measurements and fine adjustments to be made as it’s set into place. In some cases, the lowering equipment can also nudge the segment against its neighbor. In other cases, hydraulic jacks are used to pull the segments together. Divers or remotely operated submersibles can hook up the jacks. Or couplers, just like those used on freight trains, can do it without any manual underwater intervention. The jacks extend to couple the free segment to the one already installed, then retract to pull them together, compressing the gasket and sealing the area between the two bulkheads.

This joint is the most important part of an immersed tunnel design. It has to be installed blindly and accommodate small movements from temperature changes, settlement, and changes in pressure as water levels go up and down. The gasket provides the initial seal, but there’s more to it. Once in place, valves are opened in the bulkheads to drain the water between them. That actually creates a massive pressure difference between one side of the segment and the other. Hydrostatic force from the water pushes against the end of the tunnel, putting it in even firmer contact with its neighbor and creating a stronger seal. Once in its final place, the segment can be backfilled.

The tunnel segment connection is not like a pipe flange, where the joints are securely bolted together, completely restraining any movement. The joints on immersed tunnels have some freedom to move. Of course, there is a restraint for axial compression since the segments butt up against each other. In addition, keys or dowels are usually installed along the joint so that shear forces can transfer between segments, keeping the ends from shifting during settlement or small sideways movements. However, the joints aren’t designed to transfer torque, called moments. And there’s rarely much mechanical restraint to axial tension that might pull one joint away from the other. So you can see why the backfill is so important. It locks each segment into place. In fact, the first layer of backfill is called locking fill for that exact reason. I don’t think they make underwater roller compactors, and you wouldn’t want strong vibrations disturbing the placement of the tunnel segments anyway. So this material is made from angular rock that self-compacts and is placed using fall pipes in careful layers to secure each segment without shifting or disturbing it.

After that, general backfill - maybe even the original material if it wasn’t contaminated - can be used in the rest of the trench, and then a layer is placed over the top of everything to protect the backfill and tunnel against currents caused by ships and tides. Sometimes this top layer includes bands of large rock meant to release a ship’s anchor from the bottom, keeping it from digging in and damaging the tunnel.

Once a tunnel segment is secured in place, the bulkhead in the previous segment can be removed from the inside, allowing access inside the joint. The usual requirement is that access is only allowed when there are two or more bulkheads between workers and the water outside. A second seal, called an omega seal (because of its shape), then gets installed around the perimeter of the joint. And the process keeps going, adding segments to the tunnel until it’s a continuous, open path from one end to the other. When it reaches that point, all the other normal tunnel stuff can be installed, like roadways, railways, lights, ventilation, drainage, and pumps. By the time it’s ready to travel through, there’s really no obvious sign from inside that immersed tube tunnels are any different than those built using other methods.

This is a simplification, of course. Every one of these steps is immensely complicated, unique to each jobsite, and can take weeks to months, to even years to complete. And as impressive as the process is, it’s not without its downsides. The biggest one is damage to the sea or river floor during construction. Where boring causes little disturbance at the surface, immersed tube construction requires a lot of dredging. That can disrupt and damage important habitat for wildlife. It also kicks up a lot of sediment into suspension, clouding the water and potentially releasing buried contaminants that were laid down back when environmental laws were less strict. Some of these impacts can be mitigated: Sealed clamshell buckets reduce turbidity and mobilization of contaminated sediment. And construction activities can be scheduled to avoid sensitive periods like migration of important species. But some level of disturbance is inevitable and has to be weighed against the benefits of the project.

Despite the challenges, around 150 of these tunnels have been built around the globe. Some of the most famous include the Øresund Link between Denmark and Sweden, the Busan-Geoje tunnel in South Korea, the Marmaray tunnel crossing the Bosphorus in Turkey, of course, the Fort McHenry tunnel in Baltimore I mentioned earlier, and the BART Transbay Tube between Oakland and San Francisco. And some of the most impressive projects are under construction now, including the Fehmarn Belt between Denmark and Germany, which will be the world’s longest immersed tunnel. My friend Fred produced a really nice documentary about that project on The B1M channel if you want to learn more about it, and the project team graciously shared a lot of very cool clips used in this video too.

There’s something about immersed tube tunnels that I can’t quite get over. At a glance, it’s dead simple - basically like assembling lego blocks. But the reality is that the process is so complicated and intricate, more akin to building a moon base. Giant concrete and steel segments floated like ships, carefully sunk into enormous trenches, precisely maneuvered for a perfect fit while completely submerged in sometimes high-traffic areas of the sea, with tides, currents, wildlife, and any number of unexpected marine issues that could pop up. And then you just drive through it like it’s any old section of highway. I love that stuff.

When Abandoned Mines Collapse

[Note that this article is a transcript of the video embedded above.]

In December of 2024, a huge sinkhole opened up on I-80 near Wharton, New Jersey, creating massive traffic delays as crews worked to figure out what happened and get it fixed. Since then, it happened again in February 2025 and then again in March. Each time, the highway had to be shut down, creating a nightmare for commuters who had to find alternate routes. And it’s a nightmare for the DOT, too, trying to make sure this highway is safe to drive on despite it literally collapsing into the earth. From what we know so far, this is not a natural phenomenon, but one that’s human-made. It looks like all these issues were set in motion more than a century ago when the area had numerous underground iron mines. This is a really complex issue that causes problems around the world, and I built a little model mine in my garage to show you why it’s such a big deal. I’m Grady and this is Practical Engineering.

We’ve been extracting material and minerals from the earth since way before anyone was writing things down. It’s probably safe to say that things started at the surface. You notice something shiny or differently colored on the side of a hill or cliff and you take it out. Over time, we built up knowledge about what materials were valuable, where they existed, and how to efficiently extract them from the earth. But, of course, there’s only so much earth at the surface. Eventually, you have to start digging. Maybe you follow a vein of gold, silver, copper, coal or sulfur down below the surface. And things start to get more complicated because now you’re in a hole. And holes are kind of dangerous. They’re dark, they fill with water, they can collapse, and they collect dangerous gases. So, in many cases, even today, it makes sense to remove the overburden - the soil and rock above the mineral or material you’re after. Mining on the surface has a lot of advantages when it comes to cost and safety.

But there are situations where surface mining isn’t practical. Removing overburden is expensive, and it gets more expensive the deeper you go. It also has environmental impacts like habitat destruction and pollution of air and water. So, as technology, safety, and our understanding of soil and rock mechanics grew, so did our ability to go straight to the source and extract minerals underground.

One of the major materials that drove the move to underground mining was coal. It’s usually found in horizontal formations called seams, that formed when vast volumes of paleozoic plants were buried and then crushed and heated over geologic time. At the start of the Industrial Revolution, coal quickly became a primary source of energy for steam engines, steel refining, and electricity generation. Those coal seams vary in thickness, and they vary in depth below the surface too, so many early coal mines were underground.

In the early days of underground mining, there was not a lot of foresight. Some might argue that’s still true, but it was a lot more so a couple hundred years ago. Coal mining companies weren’t creating detailed maps of their mines, and even if they did, there was no central archive to send them to. And they just weren’t that concerned about the long-term stability of the mines once the resources had been extracted. All that mattered was getting coal out of the ground. Mining companies came and went, dissolved or were acquired, and over time, a lot of information about where mines existed and their condition was just lost. And even though many mines were in rural areas, far away from major population centers, some weren’t, and some of those rural areas became major population centers without any knowledge about what had happened underneath them decades ago.

An issue that confounds the problem of mine subsidence is that in a lot of places, property ownership is split into two pieces: surface rights and mineral rights. And those rights can be owned by different people. So if you’re a homeowner, you may own the surface rights to your land, while a company owns the right to drill or mine under your property. That doesn’t give them the right to damage your property, but it does make things more complicated since you don’t always have a say in what’s happening beneath the surface.

There are myriad ways to build and operate underground mines, but especially for soft rock mining, like coal, the predominant method for decades was called “room and pillar”. This is exactly what it sounds like. You excavate the ore, bringing material to the surface. But you leave columns to support the roof. The size, shape, and spacing of columns are dictated by the strength of the material. This is really important because a mine like this has major fixed costs: exploration, planning, access, ventilation, and haulage. It’s important to extract as much as possible, and every column you leave supporting the roof is valuable material you can’t recover. So, there’s often not a lot of margin in these pillars. They’re as small as the company thought they could get away with before they were finished mining.

I built a little room and pillar mine in my garage. I’ll be the first to admit that this little model is not a rigorous reproduction of an actual geologic formation. My coal seam is just made of cardboard, and the bright colors are just for fun. But, I’m hoping this can help illustrate the challenges associated with this type of mine. I’ve got a little rainfall simulator set up, because water plays a big role in these processes. This first rainfall isn’t necessarily representative of real life, since it’s really just compacting the loose sand. But it does give a nice image of how subsidence works in general. You can see the surface of the ground sinking as the sand compacts into place.

But you can also see that as the water reaches the mine, things start to deform. In a real mine, this is true, too. Stresses in the surrounding soil and rock redistribute over time from long-term movements, relaxation of stresses that were already built up in the materials before extraction, and from water.

I ran this model for an entire day, turning the rainfall on and off to simulate a somewhat natural progression of time in the subsurface. By the end of the day, the mine hadn’t collapsed, but it was looking a great deal less stable than when it started. And that’s one big thing you can learn from this model - in a lot of cases, these issues aren’t linearly progressive. They can happen in fits and starts, like this small leak in the roof of the mine. You get a little bit of erosion of soil, but eventually, enough sand built up that it kind of healed itself, and, for a while, you can’t see any evidence of any of it at the surface. The geology essentially absorbed the sinkhole by redistributing materials and stresses so there’s no obvious sign at the surface that anything wayward is happening below.

In the US, there were very few regulations on mining until the late 19th century, and even those focused primarily on the safety of the workers. There just wasn’t that much concern about long-term stability. So as soon as material was extracted, mines were abandoned. The already iffy columns were just left alone, and no one wasted resources on additional supports or shoring. They just walked away.

One thing that happens when mines are abandoned is that they flood. Without the need to work inside, the companies stop pumping out the water. I can simulate this on my model by just plugging up the drain. In a real soft rock mine, there can be minerals like gypsum and limestone that are soluble in water. Repeated cycles of drying and wetting can slowly dissolve them away. Water can also soften certain materials and soils, reducing their mechanical strength to withstand heavy loads, just like my cardboard model. And then, of course, water simply causes erosion. It can literally carry soil particles with it, again, causing voids and redistribution of stresses in the subsurface. This is footage from an old video I did demonstrating how sinkholes can form.

The ways that mine subsidence propagates to the surface can vary a lot, based on the geology and depth of the mine. For collapses near the surface, you often see well-defined sinkholes where the soil directly above the mine simply falls into the void. And this is usually a sudden phenomenon. I flooded and drained my little mine a few times to demonstrate this. Accidentally flooded my little town a few times in the process, but that’s okay. You can see in my model, after flooding the mine and draining it down, there was a partial failure in the roof and a pile of sand toward the back caved in. And on the surface, you see just a small sinkhole. In 2024, a huge hole opened right in the center of a sports complex in Alton, Illinois. It was quickly determined that part of an active underground aggregate mine below the park had collapsed, leading to the sinkhole. It’s pretty characteristic of these issues. You don’t know where they’re going to happen, and you don’t know how the surface soils are going to react to what’s happening underneath.

Subsidence can also look like a generalized and broader sinking and settling over a large area. You can see in my model that most of the surface still looks pretty flat, despite the fact that it started here and is now down here as the mine supports have softened and deformed. This can also be the case when mines are deeper in the ground. Even if the collapse is sudden, the subsidence is less dramatic because the geology can shift and move to redistribute the stresses. And the subsidence happens more slowly as the overburden settles into a new configuration. In all cases, the subsidence can extend laterally from the mine, so impacted areas aren’t always directly above. The deeper the mine, the wider the subsidence can be.

I ran my little mine demo for quite a few cycles of wet and dry just to see how bad things would get. And I admit I used a little percussion at the end to speed things along. Let’s say this is a simulation of an earthquake on an abandoned mine. [Beat] You can see that by the end of it, this thing has basically collapsed.

And take a look at the surface now. You have some defined sinkholes for sure. And you also have just generalized subsidence - sloped and wavy areas that were once level. And you can imagine the problems this can cause. Structures can easily be damaged by differential settlement. Pipes break. Foundations shift and crack. Even water can drain differently than before, causing ponding and even changing the course of rivers and streams for large areas. And even if there are no structures, subsidence can ruin high-value farm land, mess up roads, disrupt habitat, and more.

In many cases, the company that caused all the damage is long gone. Essentially they set a ticking time bomb deep below the ground with no one knowing if or when it would go off. There’s no one to hold accountable for it, and there’s very little recourse for property owners. Typical property insurance specifically excludes damage from mine subsidence. So, in some places where this is a real threat, government-subsidized insurance programs have been put in place. Eight states in the US, those where coal mining was most extensive, have insurance pools set up. In a few of those states, it is a requirement in order to own property. The federal government in the US also collects a fee from coal mines that goes into a fund that helps cover reclamation costs of mines abandoned before 1977 when the law went into effect.

That federal mining act also required modern mines to use methods to prevent subsidence, or control its effects, because this isn’t just a problem with historic abandoned mines. Some modern underground soft rock mining doesn’t use the room and pillar method but instead a process called longwall mining. Like everything in mining, there are multiple ways to do it. But here’s the basic method: Hydraulic jacks support the roof of the mine in a long line. A machine called a shearer travels along the face of the seam with cutting drums. The cut coal falls onto a conveyor and is transported to the surface. The roof supports move forward into the newly created cavity, intentionally allowing the roof behind them to collapse. It’s an incredibly efficient form of mining, and you get to take the whole seam, rather than leaving pillars behind to support the roof. But, obviously, in this method, subsidence at the surface is practically inevitable.

Minimizing the harm that subsidence creates starts just by predicting its extent and magnitude. And, just looking at my model, I think you can guess that this isn’t a very easy problem to solve. Engineers use a mix of empirical information, like data from similar past mining operations, geotechnical data, simplified relationships, and in some cases detailed numerical modeling that accounts for geologic and water movement over time. But you don’t just have to predict it. You also have to measure it to see if your predictions were right. So mining companies use instruments like inclinometers and extensometers above underground mines to track how they affect the surface. I have a whole video about that kind of instrumentation if you want to learn more after this.

The last part of that is reclamation - to repair or mitigate the damage that’s been done. And this can vary so much depending on where the mine is, what’s above it, and how much subsidence occurs. It can be as simple as filling and grading land that has subsided all the way to extensive structural retrofits to buildings above a mine before extraction even starts. Sinkholes are often repaired by backfilling with layers of different-sized materials, from large at the bottom to small at top. That creates a filter to keep soil from continuing to erode downward into the void. Larger voids can be filled with grout or even polyurethane foam to stabilize the ground above, reducing the chance for a future collapse.

I know coal - and mining in general - can be a sensitive topic. Most of us don’t have a lot of exposure to everything that goes into obtaining the raw resources that make modern life possible. And the things we do see and hear are usually bad things like negative environmental impacts or subsidence. But I really think the story of subsidence isn’t just one of “mining is bad” but really “mining used to be bad, and now it’s a lot better, but there are still challenges to overcome.” I guess that’s the story of so many things in engineering - addressing the difficulties we used to just ignore. And this video isn’t meant to fearmonger. This is a real issue that causes real damages today, but it’s also an issue that a lot of people put a great deal of thought, effort, and ultimately resources into so that we can strike a balance between protection against damage to property and the environment and obtaining the resources that we all depend on.

When Kitty Litter Caused a Nuclear Catastrophe

[Note that this article is a transcript of the video embedded above.]

Late in the night of Valentine’s Day 2014, air monitors at an underground nuclear waste repository outside Carlsbad, New Mexico, detected the release of radioactive elements, including americium and plutonium, into the environment. Ventilation fans automatically switched on to exhaust contaminated air up through a shaft, through filters, and out to the environment above ground. When filters were checked the following morning, technicians found that they contained transuranic materials, highly radioactive particles that are not naturally found on Earth. In other words, a container of nuclear waste in the repository had been breached. The site was shut down and employees sent home, but it would be more than a year before the bizarre cause of the incident was released. I’m Grady, and this is Practical Engineering.

The dangers of the development of nuclear weapons aren’t limited to mushroom clouds and doomsday scenarios. The process of creating the exotic, transuranic materials necessary to build thermonuclear weapons creates a lot of waste, which itself is uniquely hazardous. Clothes, tools, and materials used in the process may stay dangerously radioactive for thousands of years. So, a huge part of working with nuclear materials is planning how to manage waste. I try not to make predictions about the future, but I think it’s safe to say that the world will probably be a bit different in 10,000 years. More likely, it will be unimaginably different. So, ethical disposal of nuclear waste means not only protecting ourselves but also protecting whoever is here long after we are ancient memories or even forgotten altogether. It’s an engineering challenge pretty much unlike any other, and it demands some creative solutions.

The Waste Isolation Pilot Plant, or WIPP, was built in the 1980s in the desert outside Carlsbad, New Mexico, a site selected for a very specific reason: salt. One of the most critical jobs for long-term permanent storage is to keep radioactive waste from entering groundwater and dispersing into the environment. So, WIPP was built inside an enormous and geologically stable formation of salt, roughly 2000 feet or 600 meters below the surface. The presence of ancient salt is an indication that groundwater doesn’t reach this area since the water would dissolve it. And the salt has another beneficial behavior: it’s mobile.

Over time, the walls and ceilings of mined-out salt tend to act in a plastic manner, slowly creeping inwards to fill the void. This is ideal in the long term because it will ultimately entomb the waste at WIPP in a permanent manner. It does make things more complicated in the meantime, though, since they have to constantly work to keep the underground open during operation. This process, called “ground control,” involves techniques like drilling and installing roof bolts in epoxy to hold up the ceilings. I have an older video on that process if you want to learn more after this. The challenge in this case is that, eventually, we want the roof bolts to fail, allowing a gentle collapse of salt to fill the void because it does an important job.

The salt, and just being deep underground in general, acts to shield the environment from radiation. In fact, a deep salt mine is such a well-shielded area that there’s an experimental laboratory located in WIPP across on the other side of the underground from the waste panels where various universities do cutting-edge physics experiments precisely because of the low radiation levels. The thousands of feet of material above the lab shield it from cosmic and solar radiation, and the salt has much lower levels of inherent radioactivity than other kinds of rock. Imagine that: a low-radiation lab inside a nuclear waste dump.

Four shafts extend from the surface into the underground repository for moving people, waste, and air into and out of the facility. Room-and-pillar mining is used to excavate horizontal drifts or panels where waste is stored. Investigators were eventually able to re-enter the repository and search for the cause of the breach. They found the source in Panel 7, Room 7, the area of active disposal at the time. Pressure and heat had burst a drum, starting a fire, damaging nearby containers, and ultimately releasing radioactive materials into the air.

On activation of the radiation alarm, the underground ventilation system automatically switched to filtration mode, sending air through massive HEPA filters. Interestingly, although they’re a pretty common consumer good now, High Efficiency Particulate Air, or HEPA, filters actually got their start during the Manhattan Project specifically to filter radionuclides from the air.

The ventilation system at WIPP performed well, although there was some leakage past the filters, allowing a small percentage of radioactive material to bypass the filters and release directly into the atmosphere at the surface. 21 workers tested positive for low-level exposure to radioactive contamination but, thankfully, were unharmed. Both WIPP and independent testing organizations confirmed that detected levels were very low, the particles did not spread far, and were extremely unlikely to result in radiation-related health effects to workers or the public. Thankfully, the safety features at the facility worked, but it would take investigators much longer to understand what went wrong in the first place, and that involved tracing that waste barrel back to its source.

It all started at the Los Alamos National Laboratory, one of the labs created as part of the 1940s Manhattan Project that first developed atomic bombs in the desert of New Mexico. The 1970s brought a renewed interest in cleaning up various Department of Energy sites. Los Alamos was tasked with recovering plutonium from residue materials left over from previous wartime and research efforts. That process involved using nitric acid to separate plutonium from uranium. Once plutonium is extracted, you’re left with nitrate solutions that get neutralized or evaporated, creating a solid waste stream that contains residual radioactive isotopes.

In 1985, a volume of this waste was placed in a lead-lined 55-gallon drum along with an absorbent to soak up any moisture and put into temporary storage at Los Alamos, where it sat for years. But in the summer of 2011, the Las Conchas wildfire threatened the Los Alamos facility, coming within just a few miles of the storage area. This actual fire lit a metaphorical fire under various officials, and wheels were set into motion to get the transuranic waste safely into a long-term storage facility. In other words, ship it down the road to WIPP.

Transporting transuranic wastes on the road from one facility to another is quite an ordeal, even when they’re only going through the New Mexican desert. There are rules preventing the transportation of ignitable, corrosive, or reactive waste, and special casks are required to minimize the risk of radiological release in the unlikely event of a crash. WIPP also had rules about how waste can be packaged in order to be placed for long-term disposal called the Waste Acceptance Criteria, which included limits on free liquids. Los Alamos concluded that barrel didn’t meet the requirements and needed to be repackaged before shipping to WIPP. But, there were concerns about which absorbent to use.

Los Alamos used various absorbent materials within waste barrels over the years to minimize the amount of moisture and free liquid inside. Any time you’re mixing nuclear waste with another material, you have to be sure there won’t be any unexpected reactions. The procedure for repackaging nitrate salts required that a superabsorbent polymer be used, similar to the beads I’ve used in some of my demos, but concerns about reactivity led to meetings and investigations about whether it was the right material for the job. Ultimately, Los Alamos and their contractors concluded that the materials were incompatible and decided to make a switch. In May 2012, Los Alamos published a white paper titled “Amount of Zeolite Required to Meet the Constraints Established by the EMRTC Report RF 10-13: Application of LANL Evaporator Nitrate Salts.” In other words, “How much kitty litter should be added to radioactive waste?” The answer was about 1.2 to 1, inorganic zeolite clay to nitrate salt waste, by volume.

That guidance was then translated into the actual procedures that technicians would use to repackage the waste in gloveboxes at Los Alamos. But something got lost in translation. As far as investigators could determine, here’s what happened: In a meeting in May 2012, the manager responsible for glovebox operations took personal notes about this switch in materials. Those notes were sent in an email and eventually incorporated into the written procedures:

“Ensure an organic absorbent is added to the waste material at a minimum of 1.5 absorbent to 1 part waste ratio.”

Did you hear that? The white paper’s requirement to use an inorganic absorbent became “...an organic absorbent” in the procedures. We’ll never know where the confusion came from, but it could have been as simple as mishearing the word in the meeting. Nonetheless, that’s what the procedure became. Contractors at Los Alamos procured a large quantity of Swheat Scoop, an organic, wheat-based cat litter, and started using it to repackage the nitrate salt wastes. Our barrel first packaged in 1985 was repackaged in December 2013 with the new kitty litter. It was tested and certified in January 2014, shipped to WIPP later that month, and placed underground. And then it blew up. The unthinkable had happened; the wrong kind of kitty litter had caused a nuclear disaster.

While the nitrates are relatively unreactive with inorganic, mineral-based zeolite kitty litter that should have been used, the organic, carbon-based wheat material could undergo oxidation reactions with nitrate wastes. I think it’s also interesting to note here that the issue is a reaction that was totally unrelated to the presence of transuranic waste. It was a chemical reaction - not a nuclear reaction - that caused the problem. Ultimately, the direct cause of the incident was determined to be “an exothermic reaction of incompatible materials in LANL waste drum 68660 that led to thermal runaway, which resulted in over-pressurization of the drum, breach of the drum, and release of a portion of the drum’s contents (combustible gases, waste, and wheat-based absorbent) into the WIPP underground.” Of course, the root cause is deeper than that and has to do with systemic issues at Los Alamos and how they handled the repackaging of the material.

The investigation report identified 12 contributing causes that, while individually did not cause the accident, increased the likelihood or severity of it. These are written in a way that is pretty difficult for a non-DOE expert to parse: take a stab at digesting contributing cause number 5: “Failure of Los Alamos Field Office (NA-LA) and the National Transuranic (TRU) Program/Carlsbad Field Office (CBFO) to ensure that the CCP [that is, the Central Characterization Program] and LANS [that is, that is the contractor, Los Alamos National Security] complied with Resource Conservation and Recovery Act (RCRA) requirements in the WIPP Hazardous Waste Facility Permit (HWFP) and the LANL HWFP, as well as the WIPP Waste Acceptance Criteria (WAC).”

Still, as bad as it all seems, it really could have been a lot worse. In a sense, WIPP performed precisely how you’d want it to in such an event, and it’s a really good thing the barrel was in the underground when it burst. Had the same happened at Los Alamos or on the way to WIPP, things could have been much worse. Thankfully, none of the other barrels packaged in the same way experienced a thermal runaway, and they were later collected and sealed in larger containers.

Regardless, the consequences of the “cat-astrophe” were severe and very expensive. The cleanup involved shutting down the WIPP facility for several years and entirely replacing the ventilation system. WIPP itself didn’t formally reopen until January of 2017, nearly three full years after the incident, with the cleanup costing about half a billion dollars.

Today, WIPP remains controversial, not least because of shifting timelines and public communication. Early estimates once projected closure by 2024. Now, that date is sometime between 2050 and 2085. And events like this only add fuel to the fire. Setting aside broader debates on nuclear weapons themselves, the wastes these weapons generate are dangerous now, and they will remain dangerous for generations. WIPP has even explored ideas on how to mark the site post-closure, making sure that future generations clearly understand the enduring danger. Radioactive hazards persist long after languages and societies may have changed beyond recognition, making it essential but challenging to communicate clearly about risks.

Sometimes, it’s easy to forget - amidst all the technical complexity and bureaucratic red tape that surrounds anything nuclear - that it’s just people doing the work. It’s almost unbelievable that we entrust ourselves - squishy, sometimes hapless bags of water, meat, and bones - to navigate protocols of such profound complexity needed to safely take advantage of radioactive materials. I don’t tell this story because I think we should be paralyzed by the idea of using nuclear materials - there are enormous benefits to be had in many areas of science, engineering, and medicine. But there are enormous costs as well, many of which we might not be aware of if we don’t make it a habit to read obscure government investigation reports. This event is a reminder that the extent of our vigilance has to match the permanence of the hazards we create.

Why Are Beach Holes So Deadly?

[Note that this article is a transcript of the video embedded above.]

Even though it’s a favorite vacation destination, the beach is surprisingly dangerous. Consider the lifeguard: There aren’t that many recreational activities in our lives that have explicit staff whose only job is to keep an eye on us, make sure we stay safe, and rescue us if we get into trouble. There are just a lot of hazards on the beach. Heavy waves, rip currents, heat stress, sunburn, jellyfish stings, sharks, and even algae can threaten the safety of beachgoers. But there’s a whole other hazard, this one usually self-inflicted, that usually doesn’t make the list of warnings, even though it takes, on average, 2-3 lives per year just in the United States. If you know me, you know I would never discourage that act of playing with soil and sand. It’s basically what I was put on this earth to do. But I do have one exception. Because just about every year, the news reports that someone was buried when a hole they dug collapsed on top of them. There’s no central database of sandhole collapse incidents, but from the numbers we do have, about twice as many people die this way than from shark attacks in the US.

It might seem like common sense not to dig a big, unsupported hole at the beach and then go inside it, but sand has some really interesting geotechnical properties that can provide a false sense of security. So, let’s use some engineering and garage demonstrations to explain why. I’m Grady and this is Practical Engineering.

In some ways, geotechnical engineering might as well be called slope engineering, because it’s a huge part of what they do. So many aspects of our built environment rely on the stability of sloped earth. Many dams are built from soil or rock fill using embankments. Roads, highways, and bridges rely on embankments to ascend or descend smoothly. Excavations for foundations, tunnels, and other structures have to be stable for the people working inside. Mines carefully monitor slopes to make sure their workers are safe. Even protecting against natural hazards like landslides requires a strong understanding of geotechnical engineering. Because of all that, the science of slope stability is really deeply understood. There’s a well-developed professional consensus around the science of soil, how it behaves, and how to design around its limitations as a construction material. And I think a peek into that world will really help us understand this hazard of digging holes on the beach.

Like many parts of engineering, analyzing the stability of a slope has two basic parts: the strengths and the loads. The job of a geotechnical engineer is to compare the two. The load, in this case, is kind of obvious: it’s just the weight of the soil itself. We can complicate that a bit by adding loads at the top of a slope, called surcharges, and no doubt surcharge loads have contributed to at least a few of these dangerous collapses from people standing at the edge of a hole. But for now, let’s keep it simple with just the soil’s own weight.

On a flat surface, soils are generally stable. But when you introduce a slope, the weight of the soil above can create a shear failure. These failures often happen along a circular arc, because an arc minimizes the resisting forces in the soil while maximizing the driving forces. We can manually solve for the shear forces at any point in a soil mass, but that would be a fairly tedious engineering exercise, so most slope stability analyses use software. One of the simplest methods is just to let the software draw hundreds of circular arcs that represent failure planes, compute the stresses along each plane based on the weight of the soil, and then figure out if the strength of the soil is enough to withstand the stress. But what does it really mean for a soil to have strength?

If you can imagine a sample of soil floating in space, and you apply a shear stress, those particles are going to slide apart from each other in the direction of the stress. The amount of force required to do it is usually expressed as an angle, and I can show you why. You may have done this simple experiment in high school physics where you drag a block along a flat surface and measure the force required to overcome the friction. If you add weight, you increase the force between the surfaces, called the normal force, which creates additional friction. The same is true with soils. The harder you press the particles of soil together, the better they are at resisting a shear force. In a simplified force diagram, we can draw a normal force and the resulting friction, or shear strength, that results. And the angle that hypotenuse makes with the normal force is what we call the friction angle. Under certain conditions, it’s equal to the angle of repose, the steepest angle that a soil will naturally stand.

If I let sand pour out of this funnel onto the table, you can see, even as the pile gets higher, the angle of the slope of the sides never really changes. And this illustrates the complexity of slope stability really nicely. Gravity is what holds the particles together, creating friction, but it’s also what pulls them apart. And the angle of repose is kind of a line between gravity’s stabilizing and destabilizing effects on the soil. But things get more complicated when you add water to the mix.

Soil particles, like all things that take up space, have buoyancy. Just like lifting a weight under water is easier, soil particles seem to weigh less when they’re saturated, so they have less friction between them. I can demonstrate this pretty easily by just moving my angle of repose setup to a water tank. It’s a subtle difference, but the angle of repose has gone down underwater. It’s just because the particle’s effective weight goes down, so the shear strength of the soil mass goes down too. And this doesn’t just happen under lakes and oceans. Soil holds water - I’ve covered a lot of topics on groundwater if you want to learn more. There’s this concept of the “water table” below which, the soils are saturated, and they behave in the same way as my little demonstration. The water between the particles, called “pore water” exerts pressure, pushing them away from one another and reducing the friction between them. Shear strength usually goes down for saturated soils. But, if you’ve played with sand, you might be thinking: “This doesn’t really track with my intuitions.” When you build a sand castle, you know, the dry sand falls apart, and the wet sand holds together.

So let’s dive a little deeper. Friction actually isn’t the only factor that contributes to shear strength in a soil. For example, I can try to shear this clay, and there’s some resistance there, even though there is no confining force pushing the particles together. In finer-grained soils like clay, the particles themselves have molecular-level attractions that make them, basically, sticky. The geotechnical engineers call this cohesion. And it’s where sand gets a little sneaky.

Water pressure in the pores between particles can push them away from each other, but it can also do the opposite. In this demo, I have some dry sand in a container with a riser pipe to show the water table connected to the side. And I’ve dyed my water black to make it easier to see. When I pour the water into the riser, what do you think is going to happen? Will the water table in the soil be higher, lower, or exactly the same as the level in the riser? Let’s try it out. Pretty much right away, you can see what happens. The sand essentially sucks the water out of the riser, lifting it higher than the level outside the sand. If I let this settle out for a while, you can see that there’s a pretty big difference in levels, and this is largely due to capillary action. Just like a paper towel, water wicks up into the sand against the force of gravity.

This capillary action actually creates negative pressure within the soil (compared to the ambient air pressure). In other words, it pulls the particles against each other, increasing the strength of the soil. It basically gives the sand cohesion, additional shear strength that doesn’t require any confining pressure. And again, if you’ve played with sand, you know there’s a sweet spot when it comes to water. Too dry, and it won’t hold together. Too wet, same thing. But if there’s just enough water, you get this strengthening effect. However, unlike clay that has real cohesion, that suction pressure can be temporary. And it’s not the only factor that makes sand tricky.

The shear strength of sand also depends on how well-packed those particles are. Beach sand is usually well-consolidated because of the constant crashing waves. Let’s zoom in on that a bit. If the particles are packed together, they essentially lock together. You can see that to shear them apart doesn’t just look like a sliding motion, but also a slight expansion in volume. Engineers call this dilatancy, and you don’t need a microscope to see it. In fact, you’ve probably noticed this walking around on the beach, especially when the water table is close to the surface. Even a small amount of movement causes the sand to expand, and it’s easy to see like this because it expands above the surface of the water. The practical result of this dilatant property is that sand gets stronger as it moves, but only up to a point. Once the sand expands enough that the particles are no longer interlocked together, there’s a lot less friction between them. If you plot movement, called strain, against shear strength, you get a peak and then a sudden loss of strength.

Hopefully you’re starting to see how all this material science adds up to a real problem. The shear strength of a soil, basically its ability to avoid collapse, is not an inherent property: It depends on a lot of factors; It can change pretty quickly; And this behavior is not really intuitive. Most of us don’t have a ton of experience with excavations. That’s part of the reason it’s so fun to go on the beach and dig a hole in the first place. We just don’t get to excavate that much in our everyday lives. So, at least for a lot of us, it’s just a natural instinct to do some recreational digging. You excavate a small hole. It’s fun. It’s interesting. The wet sand is holding up around the edges, so you dig deeper. Some people give up after the novelty wears off. Some get their friends or their kids involved to keep going. Eventually, the hole gets big enough that you have to get inside it to keep digging. With the suction pressure from the water and the shear strengthening through dilatancy, the walls have been holding the entire time, so there’s no reason to assume that they won’t just keep holding. But inside the surrounding sand, things are changing.

Sand is permeable to water, meaning water moves through it pretty freely. It doesn’t take a big change to upset that delicate balance of wetness that gives sand its stability. The tide could be going out, lowering the water table and thus drying the soil at the surface out. Alternatively, a wave or the tide could add water to the surface sand, reducing the suction pressure. At the same time, tiny movements within the slopes are strengthening the sand as it tries to dilate in volume. But each little movement pushes toward that peak strength, after which it suddenly goes away. We call this a brittle failure because there’s little deformation to warn you that there’s going to be a collapse. It happens suddenly, and if you happen to be inside a deep hole when it does, you might be just fine, like our little friend here, but if a bigger section of the wall collapses, your chance of surviving is slim. Soil is heavy. Sand has about two-and-a-half times the density of water. It just doesn’t take that much of it to trap a person.

This is not just something that happens to people on vacations, by the way. Collapsing trenches and excavations are one of the most common causes of fatal construction incidents. In fact, if you live in a country with workplace health and safety laws, it’s pretty much guaranteed that within those laws are rules about working in trenches and excavations. In the US, OSHA has a detailed set of guidelines on how to stay safe when working at the bottom of a hole, including how steep slopes can be depending on the types of soil, and the devices used to shore up an excavation to keep it from collapsing while people are inside. And for certain circumstances where the risks get high enough or the excavation doesn’t fit neatly into these simplified categories, they require a professional engineer be involved.

So does all this mean that anyone who’s not an engineer just shouldn’t dig holes at the beach. If you know me, you know I would never agree with that. I don’t want to come off too earnest here, but we learn through interaction. Soil and rock mechanics are incredibly important to every part of the built environment, and I think everyone should have a chance to play with sand, to get muddy and dirty, to engage and connect and commune with the stuff on which everything gets built. So, by all means, dig holes at the beach. Just don’t dig them so deep. The typical recommendation I see is to avoid going in a hole deeper than your knees. That’s pretty conservative. If you have kids with you, it’s really not much at all. If you want to follow OSHA guidelines, you can go a little bigger: up to 20 feet (or 6 meters) in depth, as long as you slope the sides of your hole by one-and-a-half to one or about 34 degrees above horizontal. You know, ultimately you have to decide what’s safe for you and your family. My point is that this doesn’t have to be a hazard if you use a little engineering prudence. And I hope understanding some of the sneaky behaviors of beach sand can help you delight in the primitive joy of digging a big hole without putting your life at risk in the process.

This Bridge’s Bizarre Design Nearly Caused It To Collapse

[Note that this article is a transcript of the video embedded above.]

This is the Washington Bridge that carries I-195 over the Seekonk River in Providence, Rhode Island… or at least, it was the Washington Bridge. You can see that the westbound span is just about completely gone. In July of 2023, that part of the bridge, although marked as being in poor condition, received a passing inspection. Six months later, the bridge was abruptly closed to traffic because it was in imminent danger of collapse. Now, the whole thing has nearly been torn down as part of an emergency replacement project. Rhode Islanders who need to travel between Providence and East Providence have suffered through more than a year of traffic delays from the loss of this important link, and business owners have seen major downturns. If you live in the area, you’re probably tired of seeing it in the news. But it hasn’t had a lot of coverage outside the state. And I think it’s a really fascinating case study in the complexities of designing, building, and taking care of bridges, including some lessons that apply to designing just about anything. I’m Grady, and this is Practical Engineering.

The original bridge over the Seekonk River was finished in 1930. Part of that old bridge now serves as a pedestrian crossing and bike link. It’s a nice bridge: concrete and stone multiple arch spans give it a graceful look over the river. In 1959, when I-195 expanded to include this road, it quickly filled with traffic. The old bridge just wasn’t big enough, at least according to the standards of the time. So, a new bridge to carry the westbound lanes was planned, with the federal government picking up most of the bill.

Since the feds were paying, they wanted a simple, inexpensive steel girder bridge. But Rhode Island refused. The state didn’t want a plain, stark, utilitarian structure right next to their historic and elegant multi-arch bridge. It took years to come to an agreement, but eventually, they met in the middle with the Federal Bureau of Roads agreeing to include false concrete arch facades between each of the exterior piers, matching the style of the eastbound bridge. But by that time, the field of bridge engineering had shifted.

The Interstate Highway system in the US started in 1956 with the idea of an interconnected freeway system with no at-grade intersections. Every road and rail crossing required grade separation, and that meant we started building a lot of bridges. We’re up to around 55,000 today, and that’s just on the interstates. With steel in short supply, a new kind of bridge girder was coming into vogue made from pre-stressed concrete. In simple reinforced concrete structures, the rebar is just cast inside. It takes some deflection of the concrete before the steel can take on any of the internal stress within the member. For beams, the amount of deflection needed to develop the strength of the steel often leads to cracks, which eventually lead to corrosion as water reaches the steel. But if you can load up the steel before the beam is put into service, in other words, “prestress” it, you can stiffen the beam, making it less likely to crack under load. I have a whole video going into more detail about prestressed concrete if you want to learn more after this. If you’ve already seen it, then you know there are two main ways to do it.

In some structures, the reinforcing steel is tensioned before the concrete is cast. This “pre-tensioning” is usually done in facilities with specialized equipment that can apply and hold those extreme forces while the concrete cures. Alternatively, you can do it on-site by running steel tendons through hollow tubes in the concrete. Once it’s cured, jacks are used to stress the tendons, a process called post-tensioning.

The engineers for the westbound lanes of the Washington Bridge took advantage of this relatively new construction method, using both post-tensioned and pre-tensioned beams. While most of the grade separation bridges on interstate highways were rigidly standardized, this was a bridge unlike practically any other in the United States. It had 18 spans of varying structural types. Except for the navigation span for boats that used steel girders, the rest of the bridge passing over the water used cantilever beams.

Rather than having the end of the beam sit on the pier like most beam bridges do, called simply supported, the primary beams in the Washington Bridge were supported at their center, cantilevering out in both directions. The pre-tensioned drop-in concrete girders were suspended between the cantilever arms. Those cantilever beams were post-tensioned structural members. Five steel cables were run in hollow ducts from one end to the other, then tensioned to roughly 200,000 pounds (nearly a meganewton each), and locked off at anchorages on both ends. Then the ducts were filled with grout to bond the strands to the rest of the concrete member and protect them against corrosion.

Most of the cantilever beams in the Washington Bridge were balanced, meaning they had roughly the same load on either side. But at the west abutment and navigation span, that wasn’t true. You can see that these beams support a drop-in girder on one end, but the steel girders over the navigation span are simply-supported on their piers. Since the cantilever beams weren’t balanced, designers needed an alternative way to keep them from rotating atop the pier. So steel rods called tie-downs were installed on each of the unbalanced cantilevers.

In December 2023, the now 57-year-old westbound bridge was in the middle of a 64-million-dollar construction project to repair damaged concrete, widen the deck for another lane of traffic, and add a new off-ramp, with the goal of extending the bridge’s life by 25 years. One of the engineers involved in that project was on site and noticed something unusual under the navigation span. Some of the tie-down rods on the unbalanced cantilevers were completely broken.

The finding was serious, so three days later, a more detailed inspection of the structure was carried out, discovering that half of the unbalanced cantilevers at piers 6 and 7 - the piers on either side of the navigation span - were not performing as designed. The Rhode Island Department of Transportation closed the bridge to traffic that day while the state could investigate the issue and come up with a solution.

The closure snarled traffic on a crossing that was already regularly congested. Westbound traffic was eventually rerouted onto the eastbound bridge, with the lanes narrowed to fit more vehicles. The state put up an interactive dashboard where you can look at travel times by route and time of day and view live webcams to try and help travelers and commuters decide how and when to get across the Seekonk River. Still, the closure has had an enormous impact on the Providence area, impacting travel times and economic activity in the area for more than a year now.

The state was fully expecting to implement some kind of emergency repair project, essentially a retrofit that would replace the broken tie-downs on the unbalanced cantilevers. The project was designed, and the contractor started installing work platforms below the bridge in January 2024. As they got access to the underside of the bridge, things started looking worse. Deteriorating concrete on the beams threatened to complicate the installation of the new tie-downs, so the state decided to do a more detailed investigation. They tested concrete in the beams, used ground penetrating radar and ultrasound to inspect the tendons inside, and even drilled into the beams to observe the actual condition of the post-tensioned cables. What they uncovered was a laundry list of serious issues.

In addition to the failed tie-down rods, there were major problems with the beams themselves. The concrete was soft and damaged, in part because of freeze-thaw action. Like most concrete from the 1960s, there was no air entrainment in the concrete beams. This requirement in most modern concrete mixes, especially in northern climates, introduces tiny air bubbles that act like cushions to reduce damage when water freezes. Without air, concrete exposed to water and freezing conditions will spall, crack, and deteriorate over time.

The post-tensioning system was also in bad condition. The anchorages at the end of the beams were corroded, and voids and soft grout were found within the cable ducts. When the inspectors drilled into the beams to reach one of the cables, they saw that the poor grout job had allowed water inside the duct, corroding the cable itself.

Most of the damage was related to the condition and location of the joints in the bridge deck, which allowed water and salty snow melt to leak down onto the structure below. If you saw my video on the Fern Hollow Bridge collapse in Pittsburgh, it was a similar situation. When the engineers analyzed the strength of the bridge, considering its actual condition, the results weren’t good.

With no traffic, the beams met the minimum requirements in the bridge code. When traffic loads were applied, it was a totally different story. The code does not allow any tension to occur in a post-tensioned member, but you can see in the graph that the top of the beam is in tension across a large portion of its length. Worse than that, the engineers found that the beams were in a condition where failure would happen before you could see significant cracking in the concrete. In other words, if the beam was in structural distress, it likely wouldn’t be caught during an inspection. There could be no warning before a potential failure. In short, this was not a bridge worth widening. It wasn’t even safe to drive on.

A big question here is: Why didn’t any of this get caught in inspections? And that mostly has to do with access. Only some of these tie-downs were visible to inspectors. The rest were embedded in concrete diaphragms that ran laterally between the beams. But it’s not clear if any special attention was paid to them, given their structural importance in the bridge. Looking through all the past inspection reports, there’s very little mention of the tie-down rods at all, and only a few pictures of them. The state actually used this photo from the July 2023 inspection, 5 months prior to when it was observed to be broken, to show that this tie-down wasn’t broken then, suggesting that maybe a large truck had caused the damage in a single event. But you can clearly see that, if it were fractured at that time, that break would be obscured by the pier in the photo. Same thing with this one; the fracture is at the very top of the rod, so it’s impossible to see if it was there in July. There’s no easy way to know how long this had been an issue. At least for these outside tie-rods, you have bare steel, exposed and mostly uncoated, directly beneath a leaky joint in the road deck. This is easy to say in hindsight, but if I’m an inspector and I understand the configuration of this bridge, I’m making sure to put eyes on every one of these visible tie-downs, or at least state clearly and explicitly that the access wasn’t enough to fully document their condition.

And it’s even worse for the post-tensioned anchorages in the beams. Those drop-in girders sat essentially flush with the ends of the beams, making it impossible to inspect their condition, let alone perform maintenance or repairs. Seismic retrofits installed in 1996 made access and visibility even tougher. And this is a perfect case study in the risks that hidden elements can pose. If you’ve ever done a renovation project on an older house, you know exactly how this goes. You start to change a light fixture, and next thing you know there’s a backhoe in your front yard. The bridge widening project uncovered the situation with the tie rods. The repairs to the tie rods revealed issues with the post-tension system in the beams. Investigation into that problem revealed further structural issues, and pretty quickly, you have a much bigger problem on your hands than you set out to fix in the first place. You’re trying to keep the public informed about what’s going on and predict how long the bridge is going to be closed at the same time that the situation is unraveling before your eyes.

The engineers looked at a bunch of options to repair all these issues, but the complexity of implementing any fixes just made it infeasible. Just to get to the beams, you’d have to demo the entire road deck and remove the drop-in girders. Since things have shifted, there was no way to know how the load had redistributed, so even taking the deck would come with risks. Then, with the state of the concrete in the beams, it wasn’t a sure bet that they could even support any external strengthening. And even if you did get it repaired, you would still have all the same issues with access and visibility. The report put it in plain words: the options for repair were “limited, complex, and [did] not completely mitigate the identified risks with the structure.” So, eventually, the state decided to demolish the entire thing and start over.

And that’s where it stands (or doesn’t stand) right now. Demolition is well underway, but that’s not the end of the mess. The state put out a request for proposals to design and build the replacement project in April 2024 with an aggressive schedule to finish construction by August 2026. Not a single contractor bid on the job, likely due to the difficult schedule and the inherent risks. The state planned to leave the substructure of the bridge (the piers and piles) intact, giving the replacement contractor the option to reuse it as a part of their design. It seems that no one could get comfortable with that idea, and I don’t blame them, considering how each milestone in this saga has only revealed new bad news about the condition of the bridge. In October, the state decided to just demo the substructure, too, adding it to the existing contract. They started a new solicitation process, this time with two stages, to try and find a contractor willing to take on this project. The two finalists were announced in December, and they expect to award a contract this summer of 2025. But, in the midst of just trying to figure out what to do with the bridge, the fight over who’s responsible for all this chaos started.

In August of 2024, the state filed a lawsuit against 13 companies, including firms that did the bridge inspections, alleging that they should have identified these structural issues earlier. At one point the attorney general stopped the demolition work to preserve evidence for the lawsuit, extending the timeline for a month. Then in January, the US Department of Justice disclosed that they’re investigating the state of Rhode Island under the False Claims Act, which comes into play when federal funds are misused or fraudulently obtained. The dual legal battles—one against the engineering firms and another potentially implicating the state—turned what was already a logistical and financial nightmare into a high-stakes showdown, with millions of dollars and public trust hanging in the balance. Then in February, this video came out showing the demolition contractor dropping huge pieces of the cantilever beams onto the barges below, sparking a workplace safety investigation from OSHA.

A fellow YouTube engineer, Casey Jones, has been covering a lot of the more detailed aspects of the situation if you want to keep up with the story, and I also have to shout out the local journalists who have done some fantastic work to keep the public apprised of the situation where maybe the State has faltered. This saga is far from over, and we’re probably going to learn a lot more in the coming months and years. Maybe the inspectors really did neglect their duties to identify major problems. Maybe the state has some issues with its inspection and review program. Probably there’s a little bit of both. But also, this bridge had some bizarre design decisions that made a lot of these problems inevitable.

Putting critical structural elements, like tie-downs and post-tension anchorages, where they can’t be inspected or repaired is essentially like planting a time bomb. We’re fortunate it was caught before it blew up. And a lot of those design decisions were driven by a roughly five-million-dollar (adjusted for inflation) battle between Rhode Island and the federal government over the visual appearance of the bridge in 1965. Now, it will cost roughly 20 times that just to tear the bridge down, and who knows how much to rebuild.

This situation is a mess! It’s an embarrassment for the state, a nightmare for the engineers and contractors who have worked on the bridge in the past, and a major problem for all the residents of Rhode Island who depend on this bridge. Every time I talk about failures, I get so much feedback about how bad US infrastructure is. And I don’t want to sugarcoat this situation, but I do want to put it in context. This is one of roughly 617,000 bridges in the US, and in some ways, it’s a success story: A serious problem was identified before it became a disaster, and the final outcome should be what was needed all along - replacing a bridge that had reached the end of its design life.

It’s not a bizarre situation that an old bridge was old. It happens all the time, and although sometimes the roadwork is frustrating, we generally understand that structures don’t last forever and eventually need to be replaced. But just like engineers design structures to be ductile, to fail with grace and warning, we want and need projects like this to happen in an orderly fashion. We should be able to recognize when replacement is necessary, plan ahead for the project, do a good job informing the public, and execute the job on a timeline that doesn’t require panic, chaos, and emergency contracts, and the Washington Bridge is a perfect case study in why that’s so important.