Which Is Easier To Pull? (Railcars vs. Road Cars)

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

Imagine the room you’re in right now was filled to the top with gravel. (I promise I’m headed somewhere with this.) I don’t know the size of the room you’re in, but if it’s anywhere near an average-sized bedroom, that’s roughly 70 tons of material. Fill every room in an average-sized apartment, and now we’re up to 400 tons. Fill up an average-sized house. That’s 900 tons. Fill up 30 of those houses, that’s roughly 25,000 tons of gravel. A city block of just pure gravel. Imagine it with me… gravel… chicken soup for the civil engineer’s soul. And now imagine you needed to move that material somewhere else several hundred miles away. How would you do it? Would you put it in 25,000 one-ton pickup trucks? Or 625 semi-trucks? Imagine the size of those engines added together and the enormous volume of fuel required to move all that material. You know what I’m getting at here. That 25,000 tons is around the upper limit of the heaviest freight trains that carry raw materials across the globe. There are heavier trains, but not many.

I’m not trying to patronize you about freight trains. It’s not that hard to imagine how much they can move. But it is harder to imagine the energy it takes. Compare those 625 semi trucks to a handful of diesel locomotives, and the difference starts to become clear just by looking at engines and the fuel required to move that mountain of material. We’re in the middle of a deep dive series on railway engineering, and it turns out that a lot of the engineering decisions that get made in railroading have to do with energy. When you’re talking about thousands of tons per trip, even the tiny details can add up to enormous differences in efficiency, so let’s talk about some of the tricks that railroads use to minimize energy use by trains. And I even tried to pull a railcar myself. I’m Grady, and this is Practical Engineering. In today’s episode we’re running our own hypothetical railway to move apartments full of gravel (and other stuff too, I guess).

By energy, I’m not just talking about fuel efficiency either. If it was that simple, do you think there would be a 160-page report from the 1970s called “Resistance of a Freight Train to Forward Motion”? I’ll link it below for some lightweight bedtime reading. Management of the energy required to pull a train affects nearly every part of a railroad. Resistances add up as forces within the train, meaning they affect how long a train can be and where the locomotives have to be placed. Resistances vary with speed, so they affect how fast a train can move. Of course they affect the size and number of locomotives required to move a train from one point to another and how much fuel they burn. And they even affect the routes on which railroads are built. Let me show you what I mean. Here’s a hypothetical railroad with a few routes from A to B. Put yourself in the engineer’s seat and see which one you think is best. Maybe you’ll pick the straightest path, but did you notice it goes straight over a mountain range?

If you've ever read about the little engine that could, you’re familiar with one of the most significant obstacles railways face: grade. A train moving up a hill has to overcome the force of gravity on its load, which can be enormous. Grade is measured in rise over run, so a 1% grade rises 1 unit across a horizontal distance of a hundred units. There’s a common rule of thumb that you need 20 pounds or 9 kilograms of tractive effort (that’s pull from a locomotive) for every ton of weight times every percent of grade. By the way, I know kilograms are a unit of mass, not weight, but the metric world uses them for weight so I’m going to too in this video. And metric tonnes are close enough to US tons that we can just assume they’re equal for the purposes of this video.

A wheelchair ramp is allowed to have a grade of up to 8.3 percent in the US. Pulling our theoretical gravel train up a slope that steep would require a force of more than 5 million pounds or 2 million kilograms, way beyond what any railcar drawbar could handle. That’s why heavy trains have locomotives in the middle, called distributed power, to divide up those in-train forces. But it’s also why railway grades have to be so gentle, often less than half a percent. Next time you’re driving parallel to a railway, watch the tracks as you travel. The road will often follow the natural ground closely, but the tracks will keep a much more consistent elevation with only gradual changes in slope.

You might think, “So what?” We’ll spend the energy on the way up the mountain, but get it back on the other side. Once the train crests the top, we can just shut off the engines and coast back down. And that’s true for gentle grades, but on steeper slopes, a train has to use its brakes on the way down to keep from getting over the speed limit. So all that energy that went into getting the train up the hill, instead of being converted to kinetic energy on the way down, gets wasted as heat in the brakes. That’s why direct routes over steep terrain are rarely the best choice for railroads. So let’s choose an alternative route.

How about the winding path that avoids the steep terrain by curving around it? Of course, the path is longer, and that’s an important consideration we’ll discuss in a moment, but those curves also matter. Straight sections of track are often called tangent track. That’s because they connect tangentially between curved sections of rail that are usually shaped like circular arcs. Outside the US, curves are measured by their radius, the distance from the center of curvature and the center of the track. Of course, in the US, our systems of measurement are a little more old-fashioned. We measure the degrees of curvature between a 100-foot chord. A 1-degree curve is super gentle, appropriate for the highest speeds. Once you get above 5 degrees, the speed limit starts coming down, with a practical limit at slow-speed facilities of around 12 degrees. In an ideal world, you only have to accelerate a train up to speed once, but on a windy path with speed restrictions, slowing and accelerating back up to speed takes extra energy.

But those curves don’t just affect the speed of a train, they also affect the tractive effort required to pull a train around them. Put simply, curves add drag. As you might have seen in the previous video of this series, the wheels of most trains are conical in shape. This allows the inside and outside wheels to travel different distances on the same rigid axle. But it’s not a perfect system. Train wheels do slip and slide on curves somewhat, and there’s flange contact too. Listen closely to a train rounding a sharp curve and you’ll hear the flanges of each wheel squealing as they slide on the rail. A 1-degree curve might add an extra pound (or half a kilogram) of resistance for every ton of train weight (not much at all). A 5-degree curve quadruples that resistance and a 10-degree curve doubles it again. When you’re talking about a train that might weigh several thousand tons, that extra resistance means several thousand more pounds pulling back on the locomotives. It adds up fast. So, depending on the number of curves along the route, and more importantly, their degree of curvature, the winding path might be just as expensive as the one straight up the mountain and back down.

Sometimes terrain is just too extreme to conquer using just grades and curves. There comes a point in the design of a railroad where the cost of going around an obstacle like a mountain or a gorge is so great that it makes good sense and actually saves money to just build a bridge or a tunnel! Many of the techniques pioneered for railroad bridges influenced the engineering of the massive road bridges that stir the hearts of civil engineers around the world. And then there’s tunnels. You know how much I like tunnels. There are even spiral tunnels that allow trains to climb or descend on a gentle grade in a small area of land. I could spend hours talking about bridges and tunnels, but they’re not really the point of this video, so I’ll try to stay on track here. Hopefully you can see how major infrastructure projects might change the math when developing efficient railroad routes.

Of course, I’ve talked about grades, curves, and acceleration, but even pulling a train on a perfectly straight and level track without changing speed at all requires energy. In a perfect world, a wheel is a frictionless device and an object in motion would tend to stay in motion. But our world is far from perfect. I doubt you need that reminder. And there are several sources of regular old rolling resistance. Let me give you something to compare to.

I put a crane scale on a sling and hooked it to my grocery hauler in the driveway to demonstrate. This car just keeps showing up in demos on the channel. Doing my best to pull the car at a constant speed, I could measure the rolling resistance. With no friction, my car would just keep rolling once I got it up to speed, but those squishy tires and friction in the bearings mean I have to constantly pull to keep the car moving. It was pretty hard to keep this consistent, so the scale jumps around quite a bit, but it averages around 30 pounds or 14 kilograms. Very roughly, it’s about a percent of the car’s weight. I put half the car on the gravel road to compare the resistance, and it took about twice the force to keep it rolling. 60 pounds (around 2% of the car’s weight) is a little much for a civil engineer, so I had to get some help pulling. We tried it with a lighter car, but the scale must not have been working right.

At slow speeds like in the demo, drag mostly comes from the pneumatic rubber tires we use on cars and trucks. They’re great at gripping the road and handling uneven surfaces or defects, but they also squish and deform as they roll. Deforming rubber takes energy, and that’s energy that DOESN’T go into moving the load down the road. It’s wasted as heat. At faster speeds, a different drag force starts to become important: fluid drag from the air. I didn’t demo that in my driveway, but it’s just as important for trains as it is for cars. Let’s take a look back at that 1970s report to see what I mean.

One of the most commonly used methods for estimating train resistance is the Davis Formula, originally published in 1926 and modified in the 70s after roller bearings became standard on railcars. It says there are three main types of resistance in a train for a given weight. The first is mechanical resistance that only depends on the weight of the train. This comes from friction in the bearings and deflections of the wheels and track. Steel is a stiff material, but not infinitely so. As a steel wheel rolls over a steel track, they squish against each other creating a contact patch, usually around the size of a small coin. The pressure between the wheel and track in this contact patch can be upwards of 100,000 psi or 7,000 bar, higher than the pressure at the deepest places in the ocean. There is an entire branch of engineering about contact mechanics, so we’ll save that for a future video, but it’s enough to say that, just like the deformation of a rubber tire down a road, this deformation of steel wheels on steel rails creates some resistance.

The second component of resistance in the Davis formula is velocity dependent. The faster the train goes, the more resistance it experiences. This is mainly a result of the ride quality of the trucks. As the train goes faster, the cars sway and jostle more, creating extra drag. The final term of the Davis formula is air resistance. Drag affects the front, the back, and the sides of the train as it travels through the air. This is velocity dependent too, but it varies with velocity squared. Double the speed, quadruple the drag. Add all three factors together and you get the total resistance of the train, the force required to keep it moving at a constant speed.

But why use an equation when you can just measure the real thing. I took a little trip out to the Texas Transportation Museum in San Antonio to show you how this works in practice. Take a look at these classic Pullman passenger cars. You can see the square doors on the bearings where lubrication would have been added to the journal boxes by crews. This facility has a running diesel locomotive, a flat car outfitted with seats for passengers, and a caboose. This little train’s main job these days is to give rides to museum patrons, but today it’s going to help us do a little demonstration.

First [choo choo] we had to decouple the car from the caboose. Then we used the locomotive to move the flat car down the track. This car was built in 1937 and used on the Missouri Pacific railroad until it was acquired by the museum in the early 1980s. The painted labels have faded, but it weighs in the neighborhood of 20 tons empty (about 15 times the weight of my car). So I set up a small winch with the force gauge and attached it to the car. The locomotive provides an ideal anchor point for the setup. But on the first try, the scale maxed out before the car started to move. It turns out the rolling resistance of a rail car is pretty high if you don’t fully disengage the brakes first. Who would’ve thought?

Now that the wheels are allowed to turn, it’s immediately clear that the tracks aren’t perfectly level. Even without the car rolling at all, it’s pulling on the scale with around 100 pounds or 45 kilograms. Once I start the winch to pull the car, the force starts jumping around just like the car, but it averages around 150 pounds or 68 kilograms. If I subtract the force from the grade, the rolling resistance of the car, the force just required to keep it moving at a constant speed, is just about 50 pounds or 32 kilograms. That’s about the same force required to move my car on the gravel road even though this car is 15 times its weight. And it’s not far off from what the Davis Formula would predict either.

We tried this a few times, and the results were pretty much the same each time. This is an old rail car on an old railway, so there’s quite a bit of variation to try and average out of the results. Little imperfections in the wheels and rail make a huge difference when the rolling resistance is so low. A joint in the track can double or triple the force required to keep the car moving, if only for a brief moment. Kind of like getting a pebble under the wheel of a shopping cart: It seems insignificant, but if it’s happened to you, you know it’s not.

Watching the forces involved, I couldn’t help but wonder if I could move the car myself. But there was no safe way for me to start pulling the car once it was already moving. I would have to try and overcome the static friction first… aaaaand that turned out to be a little beyond my capabilities. If you look close, you can see the car budging, but I couldn’t quite get it started. On a different part of the track with the wheels at a different position, maybe I could have moved it, but considering most of the working out I do is on a calculator, this result might not be that surprising. Those joints between rails don’t only add drag, but maintenance costs too, but that’s the topic of the next episode in this series, so stay tuned if you want to learn more. It’s still remarkable that the rolling resistance between a 20 ton freight railcar and my little hatchback is in the same ballpark. And that’s a big part of why railways exist in the first place. Those steel wheels on steel rails get the friction and drag low enough that just a handful of locomotives can move the same load as hundreds or trucks with a lot less energy and thus a lot less cost.

HEAVY CONSTRUCTION of a Sewage Pump Station - Ep 5

This is the fifth and final episode of a five-part pilot series to gauge your interest in "How It's Made"-esque heavy construction videos. Drop a comment or send me an email to let me know what you think! Watch on YouTube above or ad-free on Nebula here.

Why There's a Legal Price for a Human Life

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

One of the very first documented engineering disasters happened in 27 AD in the early days of the Roman Empire. A freed slave named Atilius built a wooden amphitheater in a town called Fidenae outside of Rome. Gladiator shows in Rome were banned at the time, so people flocked from all over to the new amphitheater to attend the games. But the wooden structure wasn’t strong enough. One historian put it this way: “[Atilius] failed to lay a solid foundation and to frame the wooden superstructure with beams of sufficient strength; for he had neither an abundance of wealth, nor zeal for public popularity, but he had simply sought the work for sordid gain.” When the amphitheater fell, thousands of people were killed or injured. That historian put the number at 50,000, but it’s probably an exaggeration. Still, the collapse of the amphitheater at Fidenae is one of the most deadly engineering disasters in history.

Engineering didn’t really even exist at the time. Even with the foremost training in construction, Atilius would have had almost no ability beyond rules of thumb to predict the performance of materials, joints, or underlying soils before his arena was built. But there’s one thing about this story that was just as true then as it is today: The people in the amphitheater share none of the blame. They needn’t have considered (let alone verified) whether the structure they occupied was safe and sound. This idea is enshrined in practically every code of ethics you can find in engineering today: protection of the public is paramount. An engineer is not just someone who designs a structure; they are the person who takes the sole responsibility for its safety.

But if that were strictly true that safety is paramount, we would never engineering anything, because every part of the built environment comes with inherent risks. It’s clear that Atilius’s design was inadequate, and history is full of disasters that were avoidable in hindsight. But, it’s not always so obvious. The act of designing and building anything is necessarily an act of choosing a balance between cost and risks. So, how do engineers decide where to draw the line? I’m Grady, and this is Practical Engineering. Today, we’re exploring how safe is safe enough.

You might be familiar with the trolley problem or one of its variations. It’s a hypothetical scenario of an ethical dilemma. A runaway trolley is headed toward an unsuspecting group of five workers on the tracks. A siding only has a single worker. You, a bystander, can intervene and throw the switch to divert the trolley, killing only one person instead of five. But, if you do, that one person lost their life solely by your hand. There’s no right answer to the question, of course, but if you think harder about this ethical dilemma, you can find a way to blame an engineer. After all, someone engineered the safety plan for the track maintenance without an officer or lookout who could have warned the workers. And someone designed the brakes on that trolley that failed.

Hopefully, you never find yourself in such a philosophically ambiguous situation, but a large part of engineering involves making decisions that can be boiled down to a tug-of-war between cost and safety, and comparing those two can be an enormous challenge. On one side, you have dollars, and on the other, you have people. And you probably see where I’m going with this: sometimes you need a conversion factor. It sounds morbid, but it’s necessary for good decision-making to put a dollar price on the value of a human life. More technically, it’s the cost we’re willing to bear to reduce risks such that the expected number of fatalities goes down by one. But that’s not quite as easy to say.

Of course, no one is replaceable. You might say your life is priceless, but there are countless ways people signal how much value they put on their own safety. How much are people willing to pay for vehicles with higher safety ratings versus those that rank lower? How much life insurance do people purchase, and for what terms? What’s the difference in wages between people who do risky jobs and those who aren’t willing to? Economists much smarter than me can look at this type of data, aggregate it, and estimate what we call the Value of a Statistical Life or VSL. The US Department of Transportation, among many other organizations, actually does this estimation each year to help determine what safety measures are appropriate for projects like highways. The 2022 VSL is 12.5 million dollars.

Whether that number seems high or low, you can imagine how this makes safety decisions possible. Say you’re designing a new highway. There are countless measures that can be taken to make highways more safe for motorists: add a median, add a barrier, add rumble strips to warn drivers of lane diversions, increase the size of the clear zones, add guardrails, increase the radius of curves, cover the whole thing in bubble wrap, and so on. Each of these increases the cost of the highway, reducing the feasibility of building it in the first place. In other words, you don’t have the budget to make sure no one ever dies on this road. So, you have to decide which safety measures are appropriate and which ones may not be justified for the reduction in risk they provide. If you have a dollar amount for each fatality that a safety measure will prevent, it makes it much simpler to draw that line. You just have to compare the cost of the measure with the cost of the lives it saves.

But, really, It’s almost never quite so unequivocal. During the construction of the Golden Gate Bridge, the chief engineer required the contractor to put up an expensive safety net, not because it was the law, but just because it seemed prudent to protect workers against falls. The net eventually saved 19 people from plunging into the water below. That small group, who called themselves the Halfway to Hell Club, easily made up for the cost of that net, and that little example points to a dirty truth about the whole idea of weighing benefits and costs in terms of dollars: it’s predicated on the idea that we can actually know with certainty how much any one change to a structure will affect its safety over the long term (not to mention that we’ll know how much it actually costs, but I’ve covered that in a separate video). The truth is that we can only make educated guesses. Real life just comes with too many uncertainties and complexities. For example, in some northern places, the divots that form rumble strips on highways collect melted snow and de-icing salt, effectively creating a salt lick for moose and elk. What should be a safety measure, in some cases, can have the exact opposite effect, inviting hooved hazards onto the roadway. Humanity and the engineering profession have learned a lot of lessons like that the hard way because there was no other way to learn them. Sometimes, we have opportunities to be proactive, but it’s rare. As they say, most codes and regulations are written in blood. It’s a grim way to think about progress, but it’s true.

Look at fires and their consideration in modern building design. Insulated stairwells, sprinkler systems, emergency lights and signs, fire-resistant materials, and rated walls and doors - none of that stuff is free. It increases the cost of a building. But through years of studying the risks of fires through the tragedies of yesteryear, the powers at be decided that the costs of these measures to society (which we all pay in various ways) were worth the benefits to society through the lives they would save. And, by the way, there are countless safety measures that aren’t required in the building code or other regulations for the same reason.

Here’s an example: Earlier this year, a fuel tanker truck crashed into a bridge in Philadelphia, starting a fire and causing it to collapse. I made a video about it if you want more details. Even though there have been quite a few similar events in the recent past, bridge safety regulations don’t have much to say about fires. That’s because the risk of this kind of collapse is pretty well understood to take a bit of time. In almost every case, that timespan between when a fire starts and when it affects the structural integrity of the bridge is enough for emergency responders to arrive and close the road. Bridge fires, even if they end in a collapse, rarely result in fatalities. We could require bridges to be designed with fire-resistant materials, but (so far, at least), we don’t do it because the benefits through lives saved just wouldn’t make up for the enormous costs.

You can look at practically any part of the built world and find similar examples: flood infrastructure, railroads, water and wastewater utilities, and more. You know I have to talk about dams, and in the US, the federal agencies who own the big dams, mainly the Corps of Engineers and the Bureau of Reclamation, have put a great deal of thought and energy into how safe is safe enough. A dam failure is often a low-probability event but with high consequences, and those types of risks (like plane crashes and supervolcano eruptions) are the hardest for us to wrap our heads around. And dams can be enormous structures. They provide significant benefits to society, but the costs to upgrade them can be sky-high, so it’s prudent to investigate and understand which upgrades are worth it and which ones aren’t.

There’s an entire field of engineering that just looks at risk analysis, and federal agencies have developed a framework around dam safety decision-making by trying to put actual numbers to the probability of any part of a dam failing and the resulting consequences. Organizations around the world often use a chart like this, called an F-N chart, to put failure risks in context. Very roughly, society is less willing to tolerate a probability of failure the more people who might die as a result. Hopefully, that’s intuitive. So, a specific risk of failure can be plotted on this graph based on its probability and consequences. If the risks are too high, it’s justified to spend public money to reduce them. Below the line, spending more money to increase safety is just gold plating.

But above a certain number of deaths and below a certain probability, we kind of just throw up our hands. This box is really an acknowledgment that we aren’t brazen enough to suggest that society could tolerate any event where more than 1,000 people would die. The reality is that we’ve designed plenty of structures whose failure could result in so many deaths, but those structures’ benefits may outweigh the risks. Either way, such serious consequences demand more scrutiny than just plotting a point on a simple graph.

All this is, of course, not just true for civil structures, but every aspect of public safety in society. Workplace safety rules, labeling of chemicals, seatbelt rules, and public health measures around the world use this idea of the Value of a Statistical Life to justify the cost of reducing risks (or the savings of not reducing them). A road, bridge, dam, pipeline, antenna tower, or public arena for gladiatorial fights can always be made safer by spending more resources on design and construction. Likewise, resources can be saved by decreasing a structure’s strength, durability, and redundancy. Someone has to make a decision about how safe is safe enough. There’s a popular quote (unattributable, as far as I can tell) that gets the point across pretty well: “Any idiot can build a bridge that stands, but it takes an engineer to build a bridge that barely stands.” But there’s a huge difference between a bridge that barely stands and one that barely doesn’t. When it’s done correctly, people will consider you a good steward of the available resources. And, when it’s done poorly, your name gets put in the intro of online videos about structural failures. Thank you for watching, and let me know what you think.

HEAVY CONSTRUCTION of a Sewage Pump Station - Ep 4

This is the fourth episode of a five-part pilot series to gauge your interest in "How It's Made"-esque heavy construction videos. Drop a comment or send me an email to let me know what you think! Watch on YouTube above or ad-free on Nebula here.

Why Are Rails Shaped Like That?

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

Maybe more than any other type of infrastructure, railways have a contingent of devoted enthusiasts. “Railfans” as they call themselves; Or should say “ourselves”? Maybe it's the nostalgia of an earlier era or the simple appeal of seeing enormous machinery up close. But railroads and the trains that ride along them are just plain fascinating. Train drivers are often known as engineers, but operating a locomotive is far from the only engineering involved in railways. In fact, building and maintaining a railroad is a big feat full of complexity. And I’d like to share some of that complexity with you, starting where the rubber meets the road, or in this case, where the steel meets the… other steel? It might sound like a simple topic, but don’t say that to the attendees of the annual Wheel Rail Interaction Conference. This stuff is complicated, so this is the first in a series of videos I’m doing on the engineering behind railways. Why do the rails of railroads have such a weird shape? The answer is pretty ingenious. I’m Grady and this is Practical Engineering. In today’s episode, wer’e talking about train wheels and rails.

Why do we build railroads anyway? They might seem self-evident now and even kind of elementary. But modern railroads are the result of hundreds of years of innovation. And like many kinds of innovation, the development of railroads was really just a series of solving problems. For example, how can we move upwards of 100 tons per vehicle without tearing up the road in the process? Well, instead of compacted gravel, asphalt, or concrete, we can build the road out of steel. But steel is expensive, so rather than a ribbon, we can save cost by using two narrow steel rails directly below the wheels. But wooden or rubber tires have a lot of rolling resistance because they deform under load, and that resistance adds up with each individual train car. So, we use steel for the wheels too. I built this model to show exactly how this works. My wheels are plastic and rails are aluminum, but I think you’ll still get the point. Steel wheels on steel rails are just so much more efficient than…[wheel falls off track]

Well, there is the problem of turning, too. Just because you put a rail below a wheel doesn’t mean it will follow the same path. You have to have some way for the rail to correct the direction of the wheel and keep it on track, literally. And, if you look at railway wheels, the answer is obvious: flanges. The wheels on railway vehicles all have them: a lip that projects below the rail to guide the wheel as it rolls along, keeping the position side to side. You could put flanges on the outside of wheels like this, but if a horizontal force like a hard turn caused one of the wheels to lift, the flange won’t help keep the wheel on track. We put flanges on the insides of wheels so they can keep a train from derailing even if one wheel lifts off the track. Let’s put some flanges on my wheels and try that demo again. [wheels bind up on track].

You can see we haven’t fully solved the problem. Unlike a wheel that has a tiny contact point with the rail, a flange is a big surface that creates a lot of friction around every curve. If you’ve heard that characteristic squeal of a train going around a corner, that’s the sound of flanges rubbing and grinding along the side of a rail. Rails on tight curves are often made of higher-grade hardened steel compared to straight portions of the track, and sometimes they’re even greased up to minimize friction between flanges and the edges of rails. But, there’s a bigger problem at play in this demonstration than simple friction.

Instead of independent wheels, most railway cars use solid axles attached to both wheels called a wheelset. They need that design to withstand the incredible loads each axle carries, but it poses a problem around bends. A solid axle means both wheels turn at the same rate, but the length of the outer portion of track in any given curve is longer than the inside of the curve. Two wheels of the same diameter spinning at the same rate will, kind of obviously, have to roll the same distance. Since there’s a mismatch between the distances the wheels need to travel, solid-axeled wheelsets with cylindrical wheels would always experience some degree of slipping around a turn. That would not only create a bunch of additional friction, but also keep the wheels from following the curved path, and a flange can only do so much.

The trick to railway wheels is something that’s not so obvious at first glance. The wheels are actually conical. The profile of the wheel is wider on the inside next to the flange, and gently narrows toward the outside of the wheel. A wheelset with conical wheels will naturally tend to self-center itself between two rails. On a straight section of track, a wheel that rides up higher on one rail will naturally fall back down, keeping the wheelset roughly centered on the road. In a sense, conical wheels want to stay on the tracks. There’s always a little bit of wobble (exaggerated here), so trains actually move down tracks in a sinusoidal side-to-side pattern that you can sometimes feel if you’re paying attention. Incidentally, that helps the wheels wear evenly. But where it really counts is on a curve.

The turning forces on a train cause it to tend toward the outside track. This shifts the wheels over as well. The outer wheel will ride on the thicker part of its tread nearest to the flange, while the inner wheel will ride toward its edge, which has a smaller circumference. This way, the effective diameter of each wheel changes in a curve and solves the slip problem that cylindrical wheels would face. Take a look at the way these conical wheels that I 3D printed behave as they make this corner. You can see the outside wheel rolling on the wider part, effectively increasing its diameter and thus distance traveled per rotation. Conversely, the inside wheel rides on the narrower part of the cone, and so it has a smaller diameter and travels a shorter distance per rotation.

It really is kind of ingenious. Most vehicles have a differential gearbox to deal with this challenge of navigating curves; train cars just use some clever geometry. But that’s not the end of the story. You might even be thinking, “Richard Feynman already taught me this in the 80s… It’s nothing new.” But there’s more engineering involved in how train wheels and rails interact, including the interesting shape of modern rails. Think about that taper angle first. One standard in the US uses a 1:20 ratio. For the main part of the wheel, that means the outside diameter is roughly a quarter inch or 6 millimeters less than the inside diameter, and that difference has a big effect on the allowable radius of curves in a railroad. A steeper cone can navigate sharper curves, since there’s a bigger difference in the circumference from the inside to outside. You can see my wheelset can’t navigate this s-curve, despite the exaggerated conicity.

This challenge is partly solved with trucks, called bogies in the UK. You can kind of think of trucks as big rollerskates under each end of a train car. The trucks can rotate relative to the rest of the car, and they usually have some pretty serious springs and suspension systems to keep a smooth ride rolling. Most trucks keep the wheel sets parallel, but some can even allow them to ride radially with each curve.

However, even with trucks or bogies, wheels can overshoot their optimal orientation on the tracks. When the simple sinusoidal motion created by the tapered wheels is amplified by the speed of the car, the oscillation can violently slam the trucks side-to-side on the rails. This is called hunting behavior. The violent motion can even cause a train to derail. It’s worst with empty cars, and usually only happens at higher speeds, so a lot of engineering goes into developing wheel profiles and truck designs that raise the hunting onset speed so that it doesn’t limit how fast a train can go. That’s a lot of innovation on the wheel side, but what about the rails?

Just like all parts of a railroad, the rails themselves have evolved over time. Turns out there are a lot of shapes they can take and still serve the same basic function, but modern railway rails are shaped that way for a reason. Weight is equivalent to cost for big steel structures, so there’s nothing on these rails that isn’t absolutely necessary. In a sense, rails are I-beams, a shape that is well-known for its strength and something we see in plenty of other heavy load bearing steel structures. But there’s more to it than that. The bottom part of the rail, called the foot, distributes enormous loads, converting the extreme contact pressure of a steel wheel into something that can be withstood by a wooden or concrete tie. The web elevates the train above the ground, giving clearance for the flanges of the wheels and keeping everything clear of small debris that might end up on the tracks.

The head of the rail with where the action happens. This thick rounded section of steel takes an awful lot of abuse over its life, and thus experiences the bulk of the wear. An old rail section, especially on the high side of a curve, looks remarkably different than a newly forged rail. Here’s why: Theoretically, the speed of a spinning wheel exactly matches the speed of the rail at a mathematically precise point. But trains don’t care about math. For one, even steel wheels on steel rails deform a little bit as they roll. Rather than a single point, there is a small contact patch between the two. That tiny area, roughly the size of a small coin, carries all the weight of the train into the rail. But, because the contact patch is spread across the tapered wheel, the wheel is turning at many different speeds on the same piece of rail. Only the center of the contact patch actually moves at the exact speed of the train. This results in a small amount of grinding as the train moves along, slowly wearing down both the wheel and the rail. Eventually they start to conform to each other, and that’s mostly a bad thing.

Wheels can wear down to get a vertical face that wants to climb up the rail or a hollow profile with a quote-unquote “second flange” that takes the wrong direction at a switch. Most rail wheels have some amount of hollow to them, which changes how conical they actually are. Some wheels are even designed to be taken off and machined back into spec to extend their life. The best way to reduce this wear is to use hardened materials and reduce the size of the contact patch by curving the top of the rail so that the wheel only touches a tiny part of it as it rolls by. After that, it’s just a decision about how much wear you want before needing to replace the rail. The more metal you include in the rail head, the more it will cost, but the longer it will last. In fact, not all rails are equal. The lightest rails are used on straight sections and small commuter service lines. The largest rails are used on curves and heavy-haul freight tracks. Once they get worn down on the main line, they often get reinstalled for a second life in a yard or a siding where they can still bear train cars and locomotives at slow speeds.

So, rails are shaped in the funny way for a reason: they’re bulbous both to reduce the size of the contact patch and provide enough steel to wear away before needing to be replaced. And the shape of rails and wheels is still a topic of research and innovation. Just in the past few years, the standard profile of North American freight train wheels was updated to the new AAR-2A standard. Just a tiny change in the shape of the wheel was tested to have 40% less wear than the previous spec. That means trains will start seeing better steering, lower friction, better fuel consumption, and longer lasting infrastructure.

In many ways, railroads might seem like old technology, a solved problem that doesn’t need more engineering. But it’s just not true. Modern railroad companies use sophisticated software, like the Train Energy and Dynamics Simulator, to keep track of all the complexities involved in how wheels and rails interact. Simulators can let you adjust factors like train makeup, different track conditions, operating conditions, suspensions, and more to characterize how trains will handle and how much energy they’ll use. That’s the topic of the next video in this series, so stay tuned if you want to learn more.

In the 19th century, railway engineering was all about how to build railroads, finding routes through difficult terrain and efficient forms of construction. Modern rail engineering is all about getting the most out of the system. It might not look like much when you see a train passing by, but a huge amount of research, testing, and engineering went into the shape of those rails and wheels and we’re still improving them today.

HEAVY CONSTRUCTION of a Sewage Pump Station - Ep 3

This is the third episode of a five-part pilot series to gauge your interest in "How It's Made"-esque heavy construction videos. Drop a comment or send me an email to let me know what you think! Watch on YouTube above or ad-free on Nebula here.

Every Type of Railcar Explained in 15 Minutes

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

A train is a simple thing at first glance: a locomotive (or several) pull a string of cars along a railroad. But not all those railcars are equal, and there are some fascinating details if you take minute to notice their differences. I’m about to start a deep dive series on railway engineering, but I thought, before I do that, we should cover some of the basics first. How many of these cars have you spotted before? I’m Grady, and this is Practical Engineering. Let’s get started.

All trains have at least one locomotive that provides the power. They can pull from the front (called the head) of the train, push from the tail, or act as so-called distributed power somewhere in between. There’s a ton of types of locomotives, but they deserve their own video, so today I’ll focus mainly on the unpowered cars they push or pull. We’ll start with passenger cars, move on to freight, and then talk about a few of the more unusual cars you might be lucky enough to spot on the rails.

Unless you work in the railroad industry, passenger trains are the only ones you’ll ever get a chance to interact with. The standard passenger car or coach is what you’ve probably seen the most of: aisle in the center with rows of seats on either side. Some coach cars can be disconnected and rearranged, but most modern passenger cars come in “train sets” that are rarely split up in normal operation.

Some passenger cars are bilevel, also called double-decker. This can double the capacity of a car, but it’s kind of rare. That’s not only because of height and weight restrictions on railroads, but also because the added time it takes to load and unload the cars can cause congestion at busy stations.

Long-haul passenger trains may include a baggage car for checked luggage like the cargo hold of an airliner. In most cases, they’re designed to look like the rest of the passenger cars, although often with fewer windows since bags rarely enjoy the view. Combine cars have a section for passengers and one for luggage or freight.

Although tricky to identify from the outside, a common sight on passenger trains is a diner car, essentially a rolling restaurant. These cars gave rise to the quintessential American restaurant of the same name, many of which are converted railcars themselves. Some passenger routes even include a lounge car, a bar on rails, that sometimes even has live music.

If you’re sleepy after dinner, you might find yourself in a sleeper car. Open section cars have the beds in bunks with only a curtain for privacy. Most modern sleeping cars have private rooms and bathrooms akin to rolling hotels.

These days, especially in the US, passenger rail is used by people who find the journey itself to be the destination. Some passenger trains include dome cars for better sightseeing along the trip. A bulbous glass dome provides a panoramic view from the side of the car. Similarly, observation cars are sometimes included at the end of a train to give passengers a view out the back.

Of course, we can’t forget crew cars. All trains have a team of people who work aboard for operation, maintenance, and other tasks, and they sometimes need their own quarters for breaks or sleep. Especially in areas like Australia where there are huge stretches of rail without stops at cities, a whole second crew might wait in the crew car, ready to swap when the working time limits of the first crew are reached.

Passenger trains are cool, of course, but I’m more of a freight train railfan myself. There’s just something awesome about seeing a single car weighing sometimes more than 100 tons move almost effortlessly down the steel rails. And with the huge variety of types of freight that move overland come a huge variety of railcars.

Boxcars are a common sight with their huge sliding doors. They can be loaded by hand or forklift and accommodate a wide range of sizes and types of cargo that require protection from the elements. And they have a few variations too. A refrigerated boxcar is exactly what it sounds like: a giant insulated fridge or freezer on rails. They usually feature a diesel-powered refrigeration system that’s easy to spot from the outside.

If the goods being transported in a boxcar are relatively light, you end up completely filling the car before coming close to its weight capacity, sometimes called “cubing out” the car. To maximize the use of a boxcar for lightweight cargo, there are taller versions called High Cubes. Not all railroads can fit such a tall car because of tunnels or bridges, so you might see the excess height portion of the car marked in white to make sure it doesn’t inadvertently end up on a route without the necessary clearance.

If you want a train car full of cars, then you’re looking for an autorack, designed to carry consumer cars and trucks. Many have three levels and carry dozens of vehicles at once. Freight rail moves automobiles cheaper and with better protection compared to driving each one individually from factories to distribution centers. A few passenger trains pull autoracks as well, like the autotrain between Washington DC and Orlando. You can take your car on your rail trip and have it at your destination.

When it comes to freight cars, it doesn’t get much more straightforward than a flat car. A simple name for a simple function: just a rolling platform that can be used for all sorts of cargo, especially big stuff that needs to be loaded with a hoist or crane and cargo that can handle a little rain or snow. You might see flat cars used to transport heavy equipment and machinery, pipes or steel beams, or even see multiple flat cars outfitted to transport enormous wind turbine blades. Some flatcars feature bulkheads at the front and rear. These help keep loads like steel plates, pipes, and wood products from shifting forwards or backwards when the train accelerates or brakes.

Another flat car variant is the centerbeam car, used to haul lumber, plywood, wallboard and fencing. The central beam helps stiffen the car, making it possible to stack products higher. It also provides a place to secure the loads from either side of the car. Some centerbeam railcars hold enough lumber to frame out half a dozen houses!

Flatcars are also used for intermodal shipping, or using more than one mode of transportation like trucks, trains, and ships. Trailer-on-flatcar, or TOFC, isn’t exactly a distinct type of railcar, but it is a distinct use of one. A semi-trailer is lifted or driven onto a flatcar at one terminal, and it’s ready to connect back to a truck once it reaches the next intermodal facility to be driven to its final destination. This is sometimes called piggy-backing and it can be a cheaper alternative than trucking the trailer for its entire route.

Most intermodal freight these days comes in containers, standardized steel boxes that fit on trucks, trains, and ships. Container-on-flatcar, or COFC, again isn’t a different kind of car but simply a specialized use. The cast corners of steel containers have holes that make them easy to secure with latches or twist lock devices so they can be quickly loaded and unloaded.

One of the great advantages of containerization is that modern intermodal containers can be stacked. An interbox connector slots between the corner castings and holds each box together. But, you don’t see double-stacked containers on flatcars very often, because of height restrictions and issues with center of gravity. Instead, well cars recess the bottom of a container between the wheels, lowering the top of a double-stack and making it safer at speed. Not every line has the clearance, but well cars have made it possible to double-stack intermodal freight on a lot more routes than before.

Coils of sheet metal are used in countless manufacturing processes, so you can see them on freight railroads fairly frequently in coil cars. Steel coils are challenging to load and unload, and challenging to secure as well, so that’s why they get their own specialized cars. Many are covered with a hood to protect the steel or other metal cargo from the elements.

Gondola (GON-dola) cars, or gon-DO-la, depending on where you live, are used for bulk materials like scrap metal, sand, ore, and coal. They’re basically enormous wagons. Gondolas have to be loaded and unloaded from the top with a crane or bucket. Some can be turned upside down and unloaded using a rotary dumper. Look for the different color of paint on the side with the rotary coupler.

Hopper cars are like gondolas in that they’re loaded from the top, but they have sloped sides and bottoms that funnel material so they can be unloaded through hatches at the bottom. Hoppers can have open tops when carrying loads that aren’t sensitive to the weather, but covered hoppers are used for cargo that needs protection from the elements like sugar and grains.

Another option for unloading bulk goods is to tip it sideways. This is a side dump car, not very common to see. They’re mostly used to maintain the railroad itself, rather than move and deliver bulk goods to customers.

This next car is very rare, but it’s so cool I just had to include it. Behold the behemoth that is a Schnabel car. There are actually two cars with far more axles than normal, each sporting a heavy lift arm for truly enormous cargo, such as power transformers used in substations. One of the largest of these is used in the US to transport nuclear reactor containment vessels on 36 axles.

Tank cars are used to carry liquids and gases on rails. Like all railcars, there are plenty of variations, but in general, they’re split up into two types. Non-pressurized tank cars handle all kinds of liquids from milk to oil. They may have specialized coatings that match their specific cargo needs, can be insulated or even refrigerated, and they usually have a bottom outlet so that they can unload by gravity.

Pressurized cars are designed to transport liquids and gases under pressure. These tanks have thicker walls and higher standards for containment of cargo. Pressurized cars always have protective housings covering the fittings on top of the tank. But, some non-pressurized cars have them too, so you'll have to look for other subtle clues (or memorize the DOT classification numbers) to know which type each one is for sure. Tank cars designed for hazardous cargo are heavily regulated and have special features like reinforced ends called head shields, specialized couplings that reduce the impacts of a derailment, and pressure relief valves to minimize the chances of an explosion.

I can’t be totally comprehensive for this short video. If you can dream it, there’s probably a freight railcar of it somewhere, but that should be all of what you’re likely to see in the wild, plus a few that you’d be really lucky to spot. But passenger and freight cars aren't the only things you'll see on the tracks. Non-revenue cars are those used by the railroad companies themselves. After all, building and maintaining railroads is a complicated and expensive endeavor, and it takes a lot of interesting equipment to do it well. I’ll rattle some of these off, but every railroad is different in the type of equipment they use to keep things running smoothly.

Ballast is the name for the gravel bedding that railroad ties sit on. It distributes the enormous pressure of trains to the subgrade, provides lateral support to keep tracks from sliding side-to-side, and facilitates drainage to keep the subgrade from getting soggy. Ballast tampers shake and pack the ballast under the tracks, restoring the support if the ballast has settled and sometimes correcting the rail alignment too. Ballast regulators use blades and brushes to distribute the ballast material evenly around the tracks and keep excess ballast from covering the ties. A ballast cleaner picks up all the rock, separates it from any dirt, and replaces it on the tracks to improve its ability to drain water and lock together to support the railroad.

Rail Grinders do just that: grind the rails to restore their shape and remove irregularities that show up as rails wear down. A tie exchanger takes out the old ties and inserts new ones without having to remove the rails. A spiker drives the spikes that hold the rails tightly to the ties. A railroad crane is used for heavy lifting along the rails where it might be difficult to access with an overland crane. Some railways in the north use a rotary snowplow during severe winter weather to keep the tracks clear.

Sometimes you might see a work truck driving around on regular old paved roads with an extra set of flanged metal wheels. This is a road-rail vehicle also called hi-rail (since they can run both on the highway and the railroad). There’s a whole host of hi-rail vehicles out there, really any kind of work truck setup you can imagine on the highway could find itself doing work on the railroad. And this is probably the only rail vehicle you’ll have a chance of seeing without also seeing a railroad itself!

Railroads depend on large scales to measure the weight of equipment and cargo. And of course, if you’ve got a scale, you need a way to calibrate it, which is where the scale test car comes into play. These cars are basically rolling hunks of metal with very precisely known weights, kind of like a huge railroad version of the little weights you might have used in school science classes.

A particularly rare car that you’d be lucky to see is a track geometry car. They carefully measure the gauge, position, curvature, and alignment of the railroad, helping to ensure the safety and smoothness of tracks without interrupting service. Unlike manual measurements of rail geometry, the measurements of track geometry cars account for loading conditions since the car itself is a full-scale railroad car.

And finally, bringing up the rear, a train car we’ve all heard of, but one you won’t really see too much of any more: the caboose. Historically, cabooses housed crewmembers who had a host of jobs, from helping with switching and shunting cars around, to looking for damaged cars, dangling equipment, monitoring brakeline air pressure, and spotting overheating bearings and axles. With the advent of roller bearings and wayside defect detectors, the role of the caboose was diminished and eventually the laws requiring them on trains were relaxed. Today the last car of a freight train is often just a regular cargo car, but with a small device on the back called an End-Of-Train Device. The most sophisticated versions monitor brake line pressure and movement of the back of the train, relaying the information to the engineer at the head. And a flashing red light lets anyone know that that’s the whole train and there aren’t any cars inadvertently left behind on the tracks.

Trains are one of the most fascinating engineered systems in the world, and they’re out there, right in the open for anyone to have a look! Once you start paying attention, it's pretty satisfying to look for all the different types of railcars that show up on the tracks, and in future videos, I’m going to show you a lot more. If you’ve been inspired to keep your eye out, we put together a checklist that you can use to keep track of the cars you’ve seen. It’s linked below in the description, but that’s not all.

If you’ve watched my channel for any length of time, you know that almost every video I make is connected to something you can see in your own surroundings. You might even know I released a book about it: Engineering in Plain Sight: An Illustrated Field Guide to the Constructed Environment. And now, I’m launching a companion game too. This is Infrastructure Road Trip Bingo. Our brains have a stupendous capacity to ignore all the fascinating details that are hidden in plain sight, and road trips are the perfect opportunity to open your mind’s eye.

Infrastructure Road Trip Bingo is just what it sounds like: a spotting game to play with your fellow passengers. Each sheet has 24 engineered structures that you might see on a typical road trip. Some you’re sure to spot. Some you might need to try and influence the driver to take a special detour. Get a line of 5 before anyone else, and you win. All the icons were designed by the illustrator for my book, and there’s a cross reference table inside the cover if you want to learn more about a particular square. 100 tear-off sheets mean you’ll have plenty of chances to play and win, and the squares are randomized so that no game ends the same.

Is this a silly idea? Of course it is. But, what I’ve learned from you over all these years is that you’re enthusiastic about the built environment just like me. Engineering In Plain Sight hit the Publisher’s Weekly best seller list, and it’s still topping out categories on Amazon nearly a year later. So I wanted to give you a chance to put those observation skills to the test. Infrastructure Road Trip Bingo goes on pre-sale today, only on my website, and they’ll start shipping later this year. And if you still don’t have my book, you can get a copy bundled with your game for a huge discount as well. You can get it from any retailer, but if you buy from my website, I signed every single copy in our warehouse. These are awesome gifts, or treat yourself with something fun and cool, and support what we’re doing on Practical Engineering while you’re at it. That link’s in the description. Thank you for watching, and let me know what you think!

HEAVY CONSTRUCTION of a Sewage Pump Station - Ep 2

This is the second episode of a five-part pilot series to gauge your interest in "How It's Made"-esque heavy construction videos. Drop a comment or send me an email to let me know what you think! Watch on YouTube above or ad-free on Nebula here.

Do Droughts Make Floods Worse?

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

Do you remember the summer of 2022 when a record drought had gripped not only a large part of the United States, but most of Europe too? Reservoirs were empty, wildfires spread, crop yields dropped, and rivers ran dry. It seemed like practically the whole world was facing heatwaves and water shortages. But there was one video that warned against hoping for rain, at least not for big storm right at first. Rob Thompson, a meteorologist and professor at the University of Reading, shared a little backyard experiment: cups of water being inverted on top of grass with varying moisture levels in the soil. The results seemed to show that the dry soil absorbed the water much more slowly than the wet grass or normal summer conditions. This video was shared across the internet as a viral reminder that, contrary to what you might think, droughts can increase the impact of flooding. But is that actually true? Does dry soil absorb moisture more slowly than wet soil, and could a storm after a drought cause more runoff and worse damage than if the ground was already wet? No matter what your intuitions say, the answer’s a little more complicated than you might think. And of course, I built some garage demonstrations to show why. I’m Grady, and this is Practical Engineering. Today we’re exploring the relationship between droughts and floods.

Of all of the natural disasters we face, floods are among the worst. There have been more than 30 floods in the US since 1980 that caused over a billion dollars in damages each! And that’s not including hurricanes. In fact, floods are so impactful that I’ve already made a whole series of videos about how dangerous they are and many of the ways that engineers work to reduce the risk of flooding or at least reduce the damage they cause. Many of those flood infrastructure projects are based on a “design storm,” essentially a made up flood used to set the capacity or height of a structure. For example, the storm gutters on your street might be designed to carry the 25-year storm. Many spillways for dams are designed for a flood that is unlikely to ever occur called the Probable Maximum Flood. Of course, we just can’t run full-scale tests on flood infrastructure. Despite architects and contractors saying we always rain on the parade, civil engineers can’t call down a flood of a particular magnitude and duration from the heavens. And even if they could, it would be an ethical gray area. So, engineers who design water infrastructure instead use models to help estimate various magnitudes of flooding and predict how the built environment will respond.

There are all kinds of hydrologic models that can simulate just about every aspect of the water cycle you could imagine, but modeling basic storm hydrology is actually pretty simple. It’s usually broken into three steps. Precipitation is exactly what you would expect: how much rain actually falls from the sky and hits the surface of the earth? Transformation describes what actually happens to those raindrops as they run along the ground and the timing of how they combine and concentrate. But in between precipitation and transformation, there’s a third step. Because not all those rain drops run off and reach a river or stream. Some of them get stuck in puddles and ponds (called abstractions), some evaporate, and some soak into the ground.

I say all this to point out that the engineers and scientists who study flooding have put a great deal of thought and research into the how, where, how much, and why rainfall soaks into the ground. It’s the third leg of the “estimating how bad floods can be” stool (a stool, by the way, I spent a good part of my education and professional experience sitting on). And of course, there’s a litany of factors that affect how much precipitation is lost to infiltration into the earth versus how much runs off into rivers and creeks: temperature, vegetation, season, land use, soil type, and more. But one of the factors is more important than any other: soil moisture. And it shouldn’t be that surprising. How much water is being held between those tiny grains of silt, sand, or clay plays a pretty big role in how much more water can flow in.

Maybe you’re starting to see what I’m getting at here (and I promise the demos are coming but I think it’s important to know the theory first). One of the most beloved mathematical expressions of hydrologists everywhere is Horton’s equation. Looks a little intimidating, but it’s much simpler as a graph. This logarithmic curve shows the infiltration rate we can expect during a rain event of a given magnitude over time. At first, when the soil is driest, the rate of infiltration is highest. As rainfall continues to soak the soil, less water is absorbed, and the infiltration rate slowly approaches a steady state.

The inputs to Horton’s equation are fine for a laboratory, but they’re not really easy to estimate in a real world scenario, so most hydrologic models don’t use it. One of the simpler infiltration models actually used in engineering is the Curve Number method, originally developed by the Soil Conservation Service in the 1950s. Here, instead of esoteric laboratory variables, infiltration rates are tied to actual soil types and land uses we can estimate in the field, and this is meant to be dead simple. You too can be a civil engineer by simply picking the right number from a table and feeding it into a model. In fact, let’s try it out. My backyard is an open space, I would say in good condition, with mostly clay which is hydrologic soil group D. So my curve number should be 80.

I won't make you go through the calculations, because we can make the computer do them. This is the Hydrologic Modeling System, a free piece of software available from the US Army Corps of Engineers. I’ll plug in my backyard curve number, plug in a storm with a constant rainfall over a day, and push go. The bars show the total amount of precipitation for each time step. The red portion shows the losses and the blue portion shows the runoff. At first, all the precipitation goes toward losses as the rainfall gets caught in abstractions. But once the puddles fill up, some runoff starts to occur. You can see that, for a constant rate of precipitation, runoff increases over time, and infiltration goes down, just like we saw with Horton’s equation. I know we’re in the weeds just a bit, but I think it’s important to know that we have technically rigorous ways to describe our intuitions of how floods work. The Curve Number method (along with many others) are used across the world by engineers to characterize floods and even to calibrate hydrological models to actual floods. Of course models are never perfect, but at least they’re based on real science. Water fits into the spaces, the interstices, between soil particles. The more water there already, the harder it is for more water to flow in. But you don’t need a graph. You can see it for yourself.

I hammered a clear tube into my Curve Number 80 backyard, and we can watch the water flow into that clay soil with grass of quote-unquote “good condition.” This is actually a crude version of an actual scientific test apparatus called an infiltrometer, but this isn’t strictly scientific. The real test involves hammering the tube deeper to prevent lateral spread and maintaining a constant level to remove water pressure as a variable. But, hey, this is just a youtube demo, and I wanted to push my kid on the swing instead of babysitting the water level in a clear tube for 45 minutes.

I did take the time to graph the water level for the duration of the experiment so you can see the results more clearly. The level drops quickly at first and slows down to roughly a constant rate, just as the theory predicted. Some of that slowdown is because of the decreased water pressure over time, the variable I didn’t control, but it’s mostly because the soil became saturated, making it harder for water to infiltrate.

Just for fun, I ran another experiment in the garage with a tube full of sand. FYI, that’s roughly equivalent to “Natural Desert Landscaping” with an associated curve number of 63. Are you feeling like an expert at this yet? It’s a little harder to tell in the sand because the water flows so quickly, but it does in fact flow more quickly at the beginning before the sand is saturated. Once it saturates, the infiltration is more or less constant, just like we would expect. The reason for the sand demo is this: we’ve left out a key consideration so far which is the initial conditions. How much water is in the soil at the start of the event? If it’s a lot, you would assume there would be less infiltration. If the soil is dry, you would assume infiltration would be greater. Is it true? Let’s try it out!

Again, the sand is maybe a little bit too porous for this demonstration, and my method for adding the water isn’t so precise either. But, just paying attention to how quickly the tube fills up with water with the valve fully opened, the dry sand takes longer. That’s because more of the water is infiltrating into the soil. The wet sand is like starting halfway down the Horton curve. But that wasn’t a super satisfying result, so I put some potting soil into the tube next (Curve Number 86). I ran it once dry, then ran it again after the soil was saturated, and lined the shots up side by side. This time you can clearly see the difference. Water infiltrates into the unsaturated soil much more quickly, but once it does, it infiltrates about the same rate as the already wet example.

We have a word for this: antecedent conditions. Most of the factors we talked about that affect infiltration rates don’t stay the same over time. They change. Many hydrologic models use average conditions as a starting point, but the real world isn’t very average. Vegetation is seasonal; temperatures fluctuate; watersheds experience fires and droughts (hint hint). How wet a watershed was before the storm is an important factor in determining how much runoff will occur. According to all the theory and practical examples I’ve shown, a wet watershed will absorb less precipitation, so flooding will be worse. And the opposite is true for a dry one. More water will soak in, making flooding less impactful. But, that seems contrary to the video I showed in the introduction, and do you really think I would make a video called “Do Droughts Make Floods Worse” if the answer was just, “no”?

It turns out that certain kinds of soil, when they become very dry, also become hydrophobic. They actually repel water. This is not a super-well-understood phenomenon, but it seems that under very dry conditions, waxes, plant root excretions, and the action of bacteria and fungi create a layer at the surface that reduces a soil’s affinity to water. If you’ve ever forgotten to water a houseplant for a while, you may have experienced this yourself. It’s hard to get the water to soak in at first, and many gardeners will actually fully submerge a potted plant to properly water it.

Because it’s a finicky phenomenon, I had a little trouble creating water repellent soil in the garage, but luckily, hydrophobicity is interesting enough to be a fun kids toy. I bought some hydrophobic sand and put a layer of it on top of my regular sand to simulate this effect of soil water repellency. You can see clearly that the repellent layer slows down the infiltration of the water. It still gets through, but it happens a lot more slowly compared to if it weren’t there. So, why doesn’t this effect show up in the theory (or least the theory of flood modeling)?

There are a few reasons: number 1 is that most hydrophobic soil effects disappear pretty quickly after the soil gets wet. It just doesn’t last that long, as you know if you’ve dealt with it in your potted plants. Number 2 is that it’s a phenomenon that hasn’t been well-characterized in terms of what soils experience repellency and under what conditions. There’s no nice table for an engineer to look up values. But number 3 is the biggest one: there are other antecedent factors that just end up being more important. Very high soil moisture before a flood is much more likely to lead to severe flooding than very low soil moisture in most cases. The extreme example of this is rain-on-snow flooding, which contributed to the 2022 flooding in Yellowstone National Park. But there is one big exception to this rule: fires.

When organic stuff burns, some of that volatile material creates hydrophobic properties in the underlying soil, reducing its ability to absorb rainfall. That effect plus the loss of vegetation on the surface means that the potential for flooding after a fire increases dramatically. Storms after wildfires are known to create massive floods, mudslides, and erosion, so there is a lot of research into understanding this phenomenon.

So what’s the answer? Are floods worse after a drought? Dry conditions do kill plants and grasses that slow down runoff, they create hydrophobic soils that briefly keep water from soaking into the ground. And, they often make fire conditions worse which in turn, can lead to more impactful floods. But droughts also leave the soil drier than average, increasing its ability to soak up rainfall. In many cases, a flood after a good soaking rainfall is going to generate far more runoff than a flood after a drought.

Rob told me he was completely surprised by the response to his video, especially since he only spent a few hours making it. His goal was to show that, under certain conditions, flash floods can be worse when the underlying soil is very dry. But I suspect if his demo lasted a little bit longer (and his setup was a little more rigorous), the results may have looked a little different. And on the other hand, most models used by engineers to estimate floods assume that infiltration always goes up as soil moisture goes down, completely neglecting the fact that some soils lose their affinity for water at very low moisture levels. One statistician famously said that, “All models are wrong, but some are useful.” And even something as simple as the flow of water into the soil has so many complexities to keep track of. Like most answers to simple questions in engineering and in life: the answer is that’s it’s complicated.

HEAVY CONSTRUCTION of a Sewage Pump Station - Ep 1

Check out our new series! This is the first episode of a five-part pilot series to gauge your interest in "How It's Made"-esque heavy construction videos. Drop a comment or send me an email to let me know what you think! Watch on YouTube above or ad-free on Nebula here.

Every Construction Machine Explained in 15 Minutes

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

We talk about a lot of big structures on this channel. But, it takes a lot of big tools to build the roads, dams, sewage lift stations, and every other part of the constructed environment. To me, there’s almost nothing more fun than watching something get built, and that’s made all the better when you know what all those machines do. So, in this episode, we’re going to try something a little bit different. I’m Grady, and this is Practical Engineering. Let’s get started!

A big part of construction is just shifting around soil and rock. If you’ve ever had to dig a hole, you know how limited human effort is in moving earth. Almost no major job site is complete without at least one excavator because they’re just so versatile. Depending on size, the heavy steel bucket of an excavator can match an entire day’s digging of one guy or girl with a single scoop. But excavators get used for more than just digging. They are a lifter, pusher, crane, and hammer all in one.

A skid steer is second only to an excavator when it comes to versatility. These little machines are often equipped with a bucket, but you can attach almost any type of tool as well. While there are often purpose built machines that can do the same job, none of them can convert from loader to mower to forklift to drill rig quite so quickly, and in tight confined spaces, a skid steer is the perfect tool.

A loader is one in many machines meant to carry soil and rock across a distance. They’re often articulated in the center for tighter turns and use a large bucket on the front for lifting and dumping. They’re meant to carry materials over short distances, like the length of a construction site.

Longer hauls use a dump truck. These trucks feature a large open-topped tub meant to withstand repeated loading with various heavy materials. A typical dump truck features a hydraulic cylinder that can lift the bed, tilting it at a steep angle and allowing material to dump out of the back.. Since dump trucks carry heavy loads, lots of them have auxiliary axles that can be lowered to distribute the weight over more tires and keep the truck in compliance with roadway and bridge weight limits. Articulated haulers are dump trucks used in off-road and difficult terrain.

If you want to move a lot of soil around a large construction site, another option is a scraper. Rather than loading from the ground into a dump truck, these machines do it all in one. A huge blade scrapes directly from the ground into a hopper. It’s carried directly to where it’s needed and unloaded with a hydraulic ejector, and these are often used on large embankments like for highways and dams.

Another Swiss army knife of the construction yard is the backhoe that is kind of a combination excavator and loader. Great for small sites where it doesn’t make sense to have two pieces of equipment.

And don’t forget the bulldozer that specializes in moving material at ground level. They can’t move material over large distances, but they can spread out literal tons with their tank-like tracks.

The last stop on the digging train is the trencher. There are a huge variety of styles and sizes, but ultimately they all specialize in digging long holes for pipes and utilities. Many use a tooth chain like a giant chainsaw for the Earth!

By the way, there are about a hundred different colloquial names for almost every piece of large equipment. Different sites, suppliers, regions, and countries use different words for the same machine; it’s part of the fun. One easy tip to sound like a pro is just to add the drive style to the front of the name. It’s not a loader, it’s a wheel loader, or a tracked excavator and so on.

Now let’s hit the road. Roadwork is something we’ve all seen, and while it can be a bit frustrating if you’re stuck in a traffic jam from it, roads might be the largest engineered structures on earth. Our modern lives depend on them, and it takes some pretty cool tools to get them built.

A grader is technically an earthwork tool, but it’s used mostly on roadways. The extra long wheelbase makes it well suited for precisely leveling surfaces and evening out bumps, leaving a nice even grade.

Once all that soil is in the right place, it needs to be solidified so it doesn’t settle over time. A roller compactor is the main tool for this job. There are a few varieties of these depending on the material being compacted. Smooth drums are used for most soils and asphalt. Sheep’s foot and padded drums have protrusions that work best on clay and silt. Pneumatic tire rollers are best to knead and seal the surface. And a lot of roller compactors have a vibration feature to shake the soil into place.

An asphalt paver is the machine where the road meets the road. Hot asphalt is loaded into the machine, which spreads it into an even layer onto the subgrade using a screed. Many paving machines have a wand that follows a stringline as a reference to the exact elevation required for the roadway.

If we’re talking about making a road out of concrete, then the tool for the job is a slip former. It’s usually more efficient and produces better quality of work when paving, curbs, and highway barriers are installed continuously rather than building forms and casting them in batches. Careful control of the mix makes it possible for a slip form machine to create long concrete structures without any formwork at all.

If we just added another layer of pavement to the road every time it started to wear out, pretty soon, we’d have walls! Roads are designed to be extraordinarily tough, so removing the top layer isn’t easy. That’s a job for an asphalt mill or planer. These specialized tools grind and remove the surface with a large rotating drum. The material is routed up a conveyor system and can be loaded into a following dump truck.

It’s actually fairly common to see multiple vehicles following one another in roadwork like this. An interesting example is the so-called paving train. On one end, we have a dump truck full of asphalt fresh from the plant. This is loaded into the asphalt paver, which continuously lays a layer of asphalt that is then compacted by one or more rollers. Workers on the ground also continuously monitor the process to ensure a nice even road surface.

Not everything at a construction site is a machine with wheels or tracks. A lot of equipment gets hauled in on a trailer, or is a trailer itself. A light tower lets you work outside of daylight hours, illuminating the site so you can work at night or underground. An air compressor enables the use of lots of tools on a job site, like jackhammers, sandblasters, and painting rigs. If you need electric power instead of compressed air, diesel generators offer access to power when grid service isn’t available.

So far, the actual material we’ve seen is in bulk like earth or asphalt. Often in construction, the materials we need to lift or move are objects like girders or concrete pipes. For that you need a crane or similar material-handling equipment.

This is a pipe layer. The name is a bit confusing since the workers that operate them are also often called pipe layers. And it's no surprise what kind of jobs they do. They specialize in handling large sections of pipe and precisely lowering them and placing them into trenches.

A telescopic handler, or a telehandler or teleporter is like an all-terrain forklift. The boom can have attachments like a bucket, pallet forks, or a winch, and it telescopes to make it easy to deliver materials and equipment exactly where you need it.

If you happen to be the load that needs elevating, then you’ll need a boom lift or its cousin, the scissor lift. The operator of these controls the platform while standing on it, allowing for very positioning of people that’s much more precise, and usually safer, than a ladder. Another relative of the boom lift is a bucket truck which has a boom lift in the back, used a lot of electric and utility work on poles.

Stepping up in size, we have road-rated all-terrain cranes. If you’ve passed a giant crane driving down the highway, it was one of these, since most other types of cranes have to be hauled to a site in pieces and assembled.

As the name implies, all-terrain cranes don’t require perfectly level, paved surfaces to get to work. However, if your job site is particularly rough, you need a rough-terrain crane. The giant rubber tires on these mean you’ll need to have them transported, but once rolling, they can go where highway-rated vehicles might struggle.

If the crane you’re looking at is mounted on tracks, you’ve got a crawler crane. These heavy-duty cranes, while slower and bulkier than all-terrain cranes and also requiring modular transport to job sites, can carry immense loads and extend to even greater heights than any of the cranes we’ve seen so far. Most crawler cranes can be configured according to the job with different lengths of booms, amounts of counterweight, and extensions called jibs. A particularly fun configuration is for demolition where a crawler crane might be fitted with a wrecking ball.

Most can move from place to place, but not all. Tower cranes use large counterbalanced horizontal booms with an integrated operator cab on top of a large, well… tower. Like most of the cranes we’ve seen so far, these come in a wide range of sizes but can be absolutely enormous, almost a construction project themselves requiring other cranes for assembly.

One way to build bridges uses a specialized crane called a launching gantry. You may have heard the term gantry before for a bridgelike overhead crane. These are in all kinds of industries. A launching gantry uses the existing structure of the bridge as a base and often lifts whole pre-built sections of the bridge.

Turning from the sky and looking underground, let’s talk about a few foundation-specific machines.

The biggest and heaviest structures are supported on bedrock or some deeper geological layer. Even if the usable soil is just clay for hundreds of feet, sinking deep subterranean columns or piles below a heavy structure can keep it from settling too much over time. One way to install a pile is to dig a very deep hole, place a reinforcing steel cage in the hole, then fill the whole thing with concrete. This is the exact job that a pile drill rig is designed to do. These large-scale drills are pretty closely related to the machines used for oil and gas exploration.

Another way to install piles is to drive them into the earth, the job of a pile driver. Just like the name implies, they repeatedly strike wooden, steel, or concrete piles to sink them to the required depth.

Speaking of concrete, there’s a whole subset of construction machines that are specifically designed to handle, transport, and place this important material. You’ve probably seen a mixer truck before, and I’ll forgive you for calling them cement trucks, even though cement is just one of the ingredients of a concrete mix. The truck can be loaded with dry materials and water, and the mixing occurs en route to the job site, since concrete generally has a limited time before it begins to cure.

Concrete is often placed directly from the truck using a chute, but that’s not always the easiest way. Concrete pumps are used to pump concrete to job site locations that are hard to access with a truck, often with a huge overhead boom. Since concrete is more than twice as dense as water, these pumps operate at extremely high pressures, sometimes over 100 times atmospheric pressure!

Finishing concrete is mostly a hand-tool job, but there are some machines for big jobs, like ride-on trowels, that speed up the job of floating a slab smooth once it has started to set up.

Big jobs with lots of concrete might just mix it onsite with a mobile batching plant. This is helpful if you need to produce vast volumes of concrete over a long period in a way that would be too inconvenient or maybe even impossible for mixer trucks to handle.

Sometimes concrete needs to be placed on a sloped or vertical surface to stabilize a rock face, shore up a tunnel, or even just install a pool! The catch-all term for the various varieties of sprayed concrete is shotcrete (although some pool installers might disagree). Shotcrete machines use compressed air to apply concrete to all kinds of surfaces in the construction world.

When projects require the installation of new or additional utility lines in areas that are already built up, the traditional method of digging trenches isn’t feasible. This kind of job calls for a directional drilling machine. While these are technically boring tools, they are anything but uninteresting. I actually have a dedicated video just to talk about how they work, and specifically how they steer that bit below the ground. Go check that out after this if you want to learn more.

Hopefully there have been a few machines in the list so far that are new to you, but if not, I have a few more specialized machines you might be lucky enough to spot on a site:

Fans of the channel might recognize a soil nail rig, a specialized machine that drills out more or less horizontal shafts in an earthen slope and then adds soil nails to greatly enhance stability.

Jobs that require grout often use mobile batch plants, called grout plants. You can even inject ground into the ground at high pressures using a hydraulic pump to fill voids and stabilize soils.

A wick drain machine installs prefabricated vertical drains into the soil at regular intervals to speed up drainage of water in clay soils which helps speed up the inevitable settling of the soil so construction can get started faster.

One option for repairing existing pipelines in place without trenching is cured-in-place pipe lining. Inverting a liner impregnated with epoxy-resin into an existing pipeline using air pressure essentially puts a brand new pipe inside an old or damaged line.

One of the least boring machines that you’d be really lucky to see above ground is a tunnel boring machine. These behemoths use a complicated face of various cutting tools followed by a material removal and shoring installation apparatus to efficiently bore full scale tunnels!

Obviously I can’t be exhaustive here. The construction industry is just full of machines. There is such a variety in the type and scale of projects that manufacturers are always coming up with new and improved equipment that can get a particular job done better. And lots of industries outside of construction use heavy machinery, including mines, oil and gas, and railroads. Let me know what you think I missed or if you want a similar list within a different industry. But I think this is a good starting point for any burgeoning construction spotter, and I hope it’s exhaustive enough that if you see something that didn’t make the list, you can puzzle out its purpose on your own. That part of the satisfaction of construction spotting anyway, so get out there and see what kinds of machines you can find.

Where Does Grounded Electricity Actually Go?

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

Imagine this scenario: You have a diesel-powered generator on a stand that is electrically isolated from the ground. Run a wire from the energized slot of an outlet to an electrode driven into the ground. Don’t connect anything to the ground or neutral slots. Now imagine starting the generator. What happens? Does current flow from the energized wire into the ground or not? Your answer depends completely on your mental model of what the earth represents in an electrical circuit. After all, the idea of a circuit is just an abstraction of some really complicated electromagnetic processes, and that’s even more true on the grand scale of the power grid. Grounding is one of the most confusing and misunderstood aspects of the grid, so you can be pardoned for being a little perplexed.

For example, if I run a wire from the positive side of a battery into the ground, nothing happens. But, when an energized power line falls from a pole, there’s definitely current flowing into the ground then. Cloud-to-ground lightning strikes move huge electrical currents into or out of the earth, but my little thought experiment of a generator connected to a grounding electrode won’t create any current at all. I’ll explain why in a minute. Even on a electrical diagram, ground is just this magical symbol that hangs off the circuit willy nilly. But, connections between an electrical circuit and the ground serve quite a few different and critical purposes. And I have some demonstrations set up in the studio to help explain. I think you’re going to look at the power grid in a whole new way after this, but just don’t try these experiments at home. I’m Grady and this is Practical Engineering. In today’s episode, we’re talking about electrical grounding.

Why do we ground electrical circuits in the first place? Maybe the easiest way to answer that question is to show you what happens when we don’t. For as much importance as it gets in the electrical code, it might surprise you that it’s not always such a big deal, and in some cases, can even be beneficial. After all, lots of small electrical circuits lack a connection to the ground, even if part of the circuit is literally called, “ground.” In that case, that term really just refers to a common reference point from which voltages are measured. That’s one thing that can be confusing about voltage: it doesn’t actually refer to a single wire or trace or location, but the difference in electrical potentials between two points. For convention, we pick a common reference point, assume it has zero potential to make the math simple, and call it ground, even if there’s no reference to the actual ground below our feet. On small, low voltage devices (like battery powered toys), the difference in potential between components on the circuit board and the actual earth isn’t all that important, but that’s not true for high voltage systems connected to the grid. Let me show you why:

This is a diagram of a typical power system on the grid. The coils of a generator are shown on the left. When a magnetic field rotates past these coils, it generates electric current on the conductors, and (very generally) this is how we get the three phase AC power that is the backbone of most electric grids today. Look at nearly any transmission line, and you’ll see three main conductors that (again, very generally) correspond to this diagram. But what you don’t see here is a connection to ground. Let me put another diagram underneath where distance is equal to voltage. You can see our three conductors all have the same phase-to-phase voltage, and they have the same phase-to-ground voltage too. Everything is balanced. But, in this example, that connection to the ground isn’t very strong, resulting just from the electromagnetic fields of the alternating current (called capacitive coupling).

Watch what happens during a ground fault. This could be a tree branch knocking down a power line or a conductor being blown into contact with a steel tower or any other number of problems that lead to a short between one phase and ground. Now, all of the sudden, that weak coupling force keeping the phase-to-ground voltages balanced is overpowered, and all the phases experience a voltage shift with respect to the ground. But, the phase-to-phase voltages don’t change. In fact, a ground fault on an ungrounded power system usually doesn’t cause any immediate problems. The motors and transformers and other loads on the system don’t really care about the phase-to-ground voltage because they’re hooked up between phases. This is one benefit of an ungrounded power system: in many cases it can keep working even during a ground fault. But, of course, there are some downsides too.

In the example I showed, the phase-to-ground voltages of the two unfaulted conductors rise to almost twice what they would be in a balanced condition. Here’s why that matters: Higher voltage requires more insulation which means more cost. Especially on large transmission lines where insulation means literally holding the conductors great distances away from each other and the ground, those costs can add up quick. It might seem like an esoteric problem for an electrical engineer, but in practice, it just means that ungrounded power systems can be a lot more expensive (a problem anyone can understand). But that’s just the start.

Look back at our diagram and you can see the faulted phase potential is equal to ground potential. In other words, their difference is zero. There’s no voltage, and when you have zero voltage, you also have zero current. No electricity is flowing from the conductor into the ground. Or at least not very much is. You still have the capacitive coupling between the unfaulted conductors that allows a little bit of current to flow, but it’s not much. And that matters because nearly all the devices that would protect a system from a problem (like a ground fault) need some current to flow.

If you know much about wiring in buildings, you might be familiar with the classic example of a toaster with a metal case. It could be any appliance, but let’s use a toaster. Under normal conditions, current flows from the live or hot wire through a heating element and into the neutral wire to return to the grid, completing the circuit. But, if something comes loose inside the toaster, the live or energized side of your electrical supply could come into contact with that metal case, making it energized too. This could start a fire, or in the worst case, shock someone who touches the case. So, many appliances are required to have another conductor attached to the housing, giving the current a parallel, low-resistance return path. That low resistance means lots of current will flow, triggering a breaker to shut off the circuit.

And, it’s not just the breakers in your house that work this way. Nearly all the protective devices, called relays, that monitor parts of the power grid for problems rely on fault current to tell the difference between normal electrical loads and short circuits. The simplest way to do that is make sure the fault current is much higher than the normal loads. In the case of the damaged toaster, that fault current flowed through a conductor that is called “ground” (but is actually just a parallel wire that connects to the neutral in your electrical panel). But, in the case of substations and transmission lines, the fault current path is the actual ground.

Let’s look back at the diagram and convert it to a grounded system. If I add a strong bond to ground at the generator, things don’t look much different in the unfaulted condition. But as soon as you add a phase-to-ground short circuit, the diagram looks much different. First, the other phases don’t experience a shift in their phase-to-ground potential. But secondly, there’s now a path for fault current to flow through the ground back to the source. And that’s the answer to the question in the title of this video: electrical current (in nearly all cases) doesn’t flow into the earth; it flows through the earth. The ground is really just another wire. Although not a great one. Let me show you an example.

I have a narrow acrylic box full of dry sand. I put a copper rod into the sand on either side of the box and connected a circuit with a lightbulb so that the current has to flow across the sand from one electrode to the other. When I turn on the switch, nothing happens. It turns out that dry sand is a pretty good insulator. In fact, soil and rock vary widely in how well they conduct electrical current. The resistivity changes with soil type, seasons, weather, temperature, and moisture content. For example, let’s try to wet this sand and see if it makes a difference. Still nothing. Even completely saturating the sand with tap water, only a tiny current flows. You can barely see anything in the lightbulb, but the current meter shows a tenth of an amp now.

Soil resistivity also changes with the chemical constituents in the soil, which is why I’m having trouble getting any current to flow through the sand. There just aren’t enough electrolytes. Even with a layer of standing water on top of the sand doesn’t conduct much current at all. If I add just a little bit of salt water to that standing water, immediately you see that the resistivity goes down and the lightbulb is able to light. And if I let that salt water soak into the soil, now the sand is able to conduct electricity too.

This resistivity of soil to conduct current is pretty important. Earth isn’t a great wire, but what it lacks in conductivity, it makes up for in size. You can kind of image current flowing from a ground electrode into the surrounding soil as a series of concentric shells, each representing a drop in voltage between the faulted conductor and the ground potential. Each shell has more surface area for current to flow and so has lower resistance until eventually there’s practically no resistance at all. But up close to the electrode, the shells are spaced tightly together toward a single point or line. That spacing is related to the resistance of the soil, and it can represent a pretty serious safety issue. Here’s a little demonstration I set up to show how this works.

This is a length of nichrome wire connected between mains voltage with a few power resistors in between to limit the current. When I flip the switch, electrical current flows through the wire, simulating a ground fault. This length of NiChrome wire is resistive to the flow of current just like the soil would be in a ground fault condition. You can see it heat up when I flip the switch. That means the electric potential along this wire is different at every point. I can show that just by measuring the voltage with a meter at a few different locations.

Remember that voltage is the difference in potential between two points, or in the case of Zap McBodySlam here, between two feet. When Zap steps on the wire, his legs are are at two different electric potentials, and unfortunately, human bodies are better conductors than the ground. That difference in electric potential creates a voltage that drives current up into one leg and down out of the other. In this case, I just have that voltage turning on a little light, but depending on how high that voltage is, and how well Zap is insulated from it, this step potential can be a matter of life or death. In fact, power line technicans are often encouraged to hop on one foot away from a ground fault to reduce the chance of a step potential. It sounds silly, but it might save their life.

Similarly, power technicians often come into contact with the metal cases around equipment regularly. So, if a ground fault happens on a piece of equipment, and the resistance of the grounding system is too high, there can be a voltage between the ground and the metal case, again creating the possibility of a voltage across a person’s body, called touch potential. The engineers who design power plants, substations, and transmission lines have to consider what touch potentials and step potentials can be safely withstood by a person and design grounding systems to make sure that they never exceed that level. For example, most substations are equipped not just with a single grounding electrode but a grid of buried conductors to minimize resistance in the earth connection. You might also notice that many substations use crushed rock as the ground surface. That’s not just because linesmen don’t like to mow the grass. It’s because the crushed rock, like the dry sand in my demo, doesn’t conduct electricity well and minimizes the chance of standing water.

But, not all power systems use the ground just as a safety measure. There are systems where the earth is actually the primary return path for current to flow. The ground is essentially the neutral line. Electrical distribution systems called “Single Wire Earth Return” or SWER are used in a few places around the world to deliver electrical power in rural areas. Using the earth as a return path can save cost, since you only have to run a single wire, but of course there are safety and technical challenges too.

Similarly, there are some high voltage transmission lines across the word that use direct current (like a battery) instead of AC. We’ll save a detailed discussion of these systems for another day, because there is a lot of fascinating engineering involved. But, I did want to mention them here, because many of these lines are equipped with really elaborate grounding systems. Although most High Voltage DC transmission lines use two conductors (positive and negative), some only use one with the return current flowing through the earth or the sea. And, even the bipolar lines often include grounding systems so they they can use ground return during and outage or emergency if one pole is out of service. For example, the Pacific DC Intertie that carries power from the pacific northwest into Los Angeles has elaborate grounding systems at both ends. In Oregon, over 1000 electrodes are buried in a ring with a circumference of 2 miles or 3.2 kilometers. In California, the grounding system consists of huge electrodes submerged in the Pacific Ocean a few miles off the shore.

Unlike AC return currents that generally follow a path that matches the transmission line, DC currents can flow through the entire earth. In essence, the electrodes are completely decoupled. That does mean they’re susceptible to some environmental issues though. They create magnetic fields that can affect compass readings and magneto-sensitive fish like salmon and eels. In ocean electrodes, the current can cause electrolysis, breaking down seawater into toxic chemicals like chloroform and bromoform. And, stray electrical currents in the ground can flow into pipelines and other buried structures, causing them to corrode. This is also a problem with some electric trains that use the rail as a return path. You may have heard that electricity takes the path of least resistance, but that’s not really true. Electricity takes all the paths it can in accordance with their relative conductivity. So, even though a big steel rail is a lot more conductive than the earth, return current from traction motors can and does flow into ground, sometimes corroding adjacent pipelines, and occasionally interfering with buried telecommunication lines too.

I’ve conveniently left out lightning from this discussion until now. Unlike a conventional circuit where current is alway moving, lightning is a type of static electricity. It’s not flowing… until it is. And unlike fault current that only uses the ground as a conduit, the current from a lightning strike really does just flow into the ground, or most frequently, out of the ground and into the atmosphere, restoring an imbalance of charge created by the movement of air or water… or something else. We really don’t understand lightning that well. But an additional and vital reason we ground electrical systems is so that, if lightning strikes, that current has a direct path to the ground. If it didn’t, it might arc across gaps or build up charge in the system, creating a fire or damaging equipment.

It’s not just lighting, ground faults, and circuit return current that flows through the earth. Lots of other natural mechanisms cause current to flow below our feet, including solar wind, changes in earth’s magnetic field, and more. These are collectively known as a telluric currents, and they intermingle below the surface with the currents that we send into the ground.

A common question I get about the electrical grid is how to know specifically which power plant serves a city or a building. It’s kind of like asking what tree or plant created the oxygen that you breath. Technically, it’s more likely to be one close to you than very far away, but that’s not quite how it works. Power gets intermingled on the grid - that’s why it’s called the grid in the first place - and it just flows along the lines in accordance with the difference in potential. And the ground works in a similar way. You can’t necessarily draw lines of current flow between sources and loads, lightning strikes, and telluric phenomena. The truth of how current flows in the ground is a little more complicated than that; it all kind of mixes together down there to some extent. But above the surface, it really isn’t so complicated. Current doesn’t flow to the ground; it flows through the ground and back up. If there is electricity moving into the ground from an energized conductor, go back to the source of that conductor and see what’s happening. For the grid, it’s probably a transformer or electrical generator, in either case, a simple coil of wire. And, the electrical current flowing out of the coil has to be equal to the electrical current flowing into it, whether that current is coming from one of the other phases, a neutral line, or an electrode buried in the ground.

Philadelphia I-95 Bridge Collapse Explained

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

On Jun 11, 2023, a fuel tanker truck caught fire on an exit underneath Interstate 95 in Northeast Philadelphia. The fire severely damaged the northbound bridge, eventually causing it to collapse. Sadly, the driver of the truck was killed in the crash, but fortunately there were no other injuries or deaths. Although it didn’t collapse, PennDOT officials said that the southbound bridge was also compromised in the fire and had to be demolished. All of I-95 through a major part of Philly was shut down for a couple of weeks, and (as of this writing) the off-ramp underneath it will likely will be closed for the near future as the bridges are rebuilt. Fires at bridges haven’t really been a major concern for transportation engineers in the past, but increasingly, they’re becoming a more serious problem. The cost to rebuild I-95 may pale in comparison to the indirects costs of having the highway shut down for so long. Or maybe not - it’s hard to say. Let’s talk about what happened and how engineers think about fire hazards at bridges. I’m Grady and this is Practical Engineering. In today’s episode, we’re talking about the I-95 bridge collapse.

The details in the intro are really all the details we know at the moment. A tanker truck crashed below the bridge, eventually leading it to collapse. There are some wild videos taken by motorists on I-95 during the fire, probably only minutes before the bridge fell, with the road deck sagging significantly. Fortunately, emergency crews were able to shut down the highway before anyone was seriously injured. The National Transportation Safety Board has had a crew on site to begin their investigation, but knowing the meticuluous pace at which they work, it will likely be a year or more before we get their report. But, the basics are pretty clear already. And in fact, even though we don’t know all the details of this particular event, we’ve seen similar collapses on several occasions. And the sequence of events is almost always the same.

In 2002, fire caused the main span of the I-20 interchange in Birmingham, Alabama to sag by 3 meters or 10 feet, necessitating replacement of the bridge. Cause of the fire? A crashed fuel tanker. In 2006, a temporary part of the Brooklyn Queens Expressway in New York collapsed during a fire. Again, the cause of the fire was crashed tanker truck under the bridge. 2007: The MacArthur Maze Interchange in Oakland, California collapsed during a fire from a crashed fuel tanker. 2009: A bridge over I-75 in Detroit collapsed after a tanker truck crashed into the overpass. 2013: A diesel tanker crash damaged a bridge in Harrisburg, Pennsylvania that had to be demolished. 2014: A gasoline tanker exploded on I-65 in Tennessee, destroying two overpass bridges. Of course, this isn’t just a US phenomenon. In 2012, a tanker overturned in Rouen (roo-AHN), France damaging the Mathilde bridge over the Seine (sehn) River and requiring part of it to be replaced. And of course, bridge fires don’t only come from tanker truck crashes. In 2017, a massive fire under I-85 in Atlanta, Georgia that resulted in collapse happened because someone set fire to construction materials stored below the bridge.

Incredibly PennDot was able to reopen this bridge a mere two weeks after it collapsed with a pretty clever solution. Rather than wait until the original bridges could be rebuilt to get I-95 back open, they decided to simply build a temporary embankment instead. After the demolition of the fire-damaged bridges was complete, the less-critical off-ramp below the bridges remained closed so that crews from PennDOT’s emergency contractor, Buckley & Company, could fill the area in and simply pave over the top. My friend Rob, built a little model of this on his channel you should check out after this.

This temporary highway wasn’t built using soil or crushed rock, the typical backfill material used in roadway embankments (at least not mostly). That stuff is heavy, so most roadway embankments have to be built slowly to allow time for the foundation to settle as each layer is added to the top, a process that can take months or even years. (Not an option in this case.) Plus there are sewer lines below the existing road that could have been overloaded by a mountain of backfill on top. Instead, the design called for lightweight backfill called foamed glass aggregate. I have a whole video we produced earlier this year about different types of lightweight backfills and how they work, so check that out if you want to learn more. This foamed glass aggregate is not cheap, many times the cost of standard backfill. But, it’s strong enough in compression to support the overlying roadway without overloading the foundation below which would lead to settlement over time or damage to underground pipes. I actually have some of it here in the studio. It feels kind of like floral foam, just a lot stronger.

The other innovative design aspect of the temporary embankment is that it leaves room on either side for the permanent repairs to the bridge. Eventually the City needs this off-ramp back open for travel, after all. The emergency embankment is sited in the center of the right-of-way to give as much space as possible for the next phase of the repairs that will replace the bridges. That also required that both sides of the embankment have a retaining wall, in this case mechanically stabilized earth walls that use reinforcing elements between each layer of backfill to keep the tall structure from collapsing. I’ve also done a few videos explaining MSE retaining walls if you want to learn more about them. The basics are easy to see in this drone footage. The reinforcement turns the backfill itself into a stable wall, making it able to both withstand vertical loads and hold back the rest of the embankment backfill. I built a little MSE cube many years back and put one of my car tires on top to show how strong it really is. Looks like the cube built by PennDOT will hold up even more cars than mine!

To their credit, PennDOT kept a live feed of construction going for most of the project. You can see the flurry of activity as workers and equipment build the embankment up to the level of the highway on either side. Traffic was rerouted onto the temporary embankment starting June 24th. But, why did a fire cause so much damage in the first place?

We, collectively, put a tremendous amount of research and engineering into the fire resistance of buildings and tunnels, but when it comes to fires at bridges, we know a lot less. In fact, most bridges in the world are designed with little, if any, consideration for fire resistance. Neither the Eurocode or the US bridge design criteria address fires or have any guidelines or requirements for how or when to engineer against them. Of course, we think about thermo-mechanical behavior of bridges all the time. I have a video all about thermal expansion and contraction of large structures. But, when you get above a few hundred degrees, there just hasn’t been much consideration. And the reasons for that are kind of obvious, at least at first glance. Less then 3% of US bridge failures between 1980 and 2012 resulted from fire. Compare that to hydraulic damage from scour and flooding that makes up nearly 50% of all failures. That alone isn’t enough reason to ignore fires in the design codes. After all, earthquakes make up only 2% of those failures, and we spend considerable resources and engineering to design bridges against seismic loads. But, you also have to consider safety. Even when bridges collapse due to fire, people are rarely injured because most places have robust emergency response capabilities. Roads are closed well before a fire is able to significantly weaken a structure. The relative infrequency of serious fires at bridges and their unlikelihood of causing a public safety issue mean that we just don’t devote a lot of resources to the problem right now… at least not proactive resources.

The National Fire Protection Association does have some guidance for fires at bridges, but it’s nebulous. They don’t recommend what fire loads should be considered, how to protect a bridge against fire, or how to analyze a structure after a fire. And, the guidelines only apply to bridges longer than 1000 feet or 300 meters. When you think about bridges, you often think about these long-span structures over major valleys or waterbodies. They’re iconic, but they’re also just the tip of the iceberg when it comes to bridges. In the US alone, there are over half a million bridges in service today, and nearly all of them are short-span bridges used mainly for grade separation (to let streams of traffic cross each other without interruption). They’re overpasses, structures you traverse every day without even noticing. But you definitely notice when one of these bridges is taken out of service. Bridges used for grade separation are more vulnerable to fires because, unlike the ones over waterways, a tanker truck can crash underneath one where the fire is most likely to cause damage. But protecting them is not as easy as it might seem.

A robust engineering guideline for design of bridges against fires would actually be pretty complicated. There are so many different variables and scenarios, and we really don’t have any collective agreement about what level of protection is appropriate. What would be the fuel source, footprint, flame height, intensity, and duration of the fire? With that information, we can try to predict the response. How does the heat transfer from the fire to the structural elements through radiation and convection, and how much do the structural elements increase in temperature as a result? These are tough questions to answer on their own, but they still don’t get to the heart of the matter, because what we really care about is how those structural elements respond. What happens to the material properties of steel and concrete when they increase in temperature way beyond what they were designed to handle? And more importantly, how does the overall structure behave? You have thermal expansion, weakening of materials, loss of stiffness, load redistribution, and a lot more. This is an extremely complicated scenario just to characterize through engineering, let alone to design protections against.

And the biggest question right now seems to be “Should we?” Bridge fires are primarily economic problems. As I mentioned, they rarely result in injuries or life safety concerns because the roadway is closed ahead of failure. But that doesn’t mean there aren’t impacts, and if you regularly drive on I-95 in Philadelphia (or any of the other roadways I mentioned before), you definitely know what I’m talking about. Replacing a bridge is an expensive endeavor, but the indirect costs that come with having a major highway closed are often higher. When the MacArthur Maze in Oakland collapsed from a tanker fire, the indirect costs of having the bridge out was estimated in the millions of dollars per day, way more than the cost of reconstruction. In fact, the rebuilding job was bid with a bonus to the contractor for each day ahead of schedule they were able to finish the job. SFGate has a great story about how they got that bridge reopened in just 26 days that I’ll link below.

Road construction often seems slow, and part of the reason is to keep the costs down. It’s not very efficient to dedicate expensive resources like equipment, engineers, and specialty construction crews to a single project. Instead, resources get spread across many jobs so that people, crews, vendors, and equipment can stay busy. Even if seemingly slow progress is often frustrating to see, it’s usually less a result of incompetence or corruption and more just government agencies trying to be good stewards of limited public resources. But a major bridge failure changes that math. Fabricators, equipment suppliers, painters, truckers, operators, and laborers are all willing to set aside their other obligations for the right price. And government agencies will happily devote their engineers and inspectors to sit and wait for questions or problems to arise on a single job if the politicians can deliver the funds for it. In the industry, they call it “accelerated construction.” It comes at a steep price, but sometimes that price is worth it.

Like the MacArthur Maze, I-95 is a busy stretch of roadway, carrying roughly 150,000 vehicle trips per day. Some of that traffic was redistributed to other routes, but some of the capacity was simply lost while the roadway was out. That means deliveries were cancelled, workers had trouble reaching their jobs, emergency response times went up, and more. The gridlock was not as apocalyptic as predicted, but there were still some major slowdowns. In most large American cities, unexpectedly closing a major highway has real economic consequences through lost commercial shipping, lost productivity, lost retail sales, more wear and tear on roadways not meant to accommodate the detour traffic, and a lot more. And those indirect costs play into the consideration in whether or not its worth it to include fire protection in the design of highway bridges.

But what’s on the other side of that equation? Of course it would have been worth the cost to protect this one bridge in Philly from a tanker fire if we knew it was going to happen, but would it have been worth the cost of protecting all the bridges just in case? Or is that gold-plating our infrastructure where it’s not really needed? We know adding highway capacity induces traffic demand, but we also know the corollary. Reducing capacity decreases traffic demand as people find alternatives to making trips in cars, and maybe a highway bridge outage isn’t quite as big a deal as the politicians and news coverage suggest. And maybe investing in some diversity in our transportation infrastructure and giving people better alternatives to driving can do more good than putting that money toward protecting bridges against the unlikely event of a fire.

Like a lot of things in engineering, the costs and risks and alternatives aren’t that easy to weigh out. Your answer might depend on how many fuel tanker trucks you see on your everyday commute. The International Associaiton for Bridge and Structural Engineering has a group working on guidelines for designing bridges against fire hazards. That’s a long way from incorporating fire protection in the design codes, but it will at least give engineers some tools to include fire resistance in designs where the situation calls for it. That group is scheduled to finish their work later in 2023, but hopefully PennDot is able to get I-95 fully repaired before then.

Why Is Desalination So Difficult?

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

This is the Carlsbad Desalination Plant outside of San Diego, California. It produces roughly ten percent of the area’s fresh water, around 50 million gallons or 23,000 cubic meters per day. Unlike most treatment plants that clean up water from rivers or lakes, the Carlsbad plant pulls its water directly from the ocean. Desalination, or the removal of salt from seawater, is one of those technologies that has always seemed right on the horizon. It might surprise you to learn that there are more than 18,000 desalination plants operating across the globe. But, those plants provide less than a percent of global water needs even though they consume a quarter of all the energy used by the water industry.

I live like 100 miles away from the nearest sea, so it’s easier for me to mix up my own batch of seawater right here in the studio. There are two main ways we use to desalinate water, and I’ve got some garage demonstrations to show you exactly how they work. Will the dubious chemistry set or the cheapest pressure washer I could find work better? Let’s track the energy use and other complications for both these demos so we can compare at the end of the video. Dumping that salt into a bucket of water may seem like no big deal, but reversing the process is a lot more complicated than you might think. I’m Grady, and this is Practical Engineering. In today’s episode, we’re talking about desalination.

Earth is a watery place. Zoom out and the stuff is practically everywhere. It doesn't seem fair that the word “drought” is even in our lexicon. And yet, the scarcity of water is one of the most widespread and serious challenges faced by people around the world. The oceans are a nearly unlimited resource of water with this seemingly trivial caveat, which is that the water is just a little bit salty. It’s totally understandable to wonder why that little bit of salt is such an enormous obstacle.

How much salt is in seawater anyway? You’ve heard of “percent,” but have you ever heard of “per mille”? Just add another circle below the slash and now, instead of parts per hundred, this symbol means parts per thousand, which is the perfect unit to talk about salinity. The salinity of the ocean actually varies a little bit geographically and through the seasons, but in general, every liter of seawater usually has around 35 grams of dissolved salt. In other words, 35 parts per thousand or 35 permille. That means, for this bucket, I need about this much salt to match the salinity of seawater.

I didn’t get it dead on, but this is close enough for our demo. Looks like a lot of salt, but I could dissolve about 10 times that much in this water before the solution becomes saturated and won’t hold any more. So, compared to how salty it could be, seawater isn’t that far from freshwater. But, compared to how salty it should be (in order to be okay to drink and such), it has a ways to go. Normal saline solution used in medicine is 9 parts per thousand because it’s approximately isotonic to your blood. That means it won’t dehydrate or overhydrate your cells. But (unless it’s masked by a bunch of sugar) even that concentration of salt in water isn’t going to taste very good.

Most places don’t put legal limits on dissolved solids for drinking water, but the World Health Organization suggests anything more than 1 part per thousand is usually unacceptable to consumers. It doesn’t taste good. 500 parts per million (or half permille) is generally the upper limit for fresh water (and that includes all dissolved solids combined, not just salt). But that means seawater desalination has to remove (or in industry jargon, reject) more than 98 percent of the salt in the water. That’s the reason why there are really only two main technologies in desalination. But neither of them are particularly sophisticated, at least in their simplest form, so I’m going to try some do-it-yourself desalination to show you how this works.

The oldest and most straightforward way to separate salt and water is distillation, and this is my very basic setup to do just that. All you chemists and laboratory professionals are probably shaking your heads right now, but this is just to illustrate the basics. On the left, I have a flask of my homemade seawater sitting in sand, in a pot, on a hot plate. Salt doesn’t like to be a gas, at least not under the conditions we normally live in on earth. Water, on the other hand, can be convinced into its gaseous state with some heat from a conventional hotplate. And that’s what I’m doing here, just adding some heat to the system. And I’m tracking exactly how much heat using this Kill A Watt meter.

Once the water is converted to steam, it is effectively separated from the salt. All I have to do is condense the vaporized water back into its liquid form. This pump moves ice water through the condenser to encourage that process… if the tube doesn’t slip out of the beaker and spill ice water all over the table.

In my receiving flask on the right, I should have distilled water that is nearly salt free. Testing it out with the meter, the dissolved solids are practically nil, just a few parts per million. But it took nearly 2 hours to get only 200 milliliters of water, and right about a kilowatt-hour of electricity too.

Water usage in the US varies quite a bit, but a rough estimate is 300 gallons (or 1,100 liters) per day per household. To produce that much water using my distillation setup here, I would have to scale it up nearly 500 times this size, and it would consume nearly 6,000 kilowatt-hours in a day (assuming the same efficiency I got in the demo). At the average residential US electricity price, it’s roughly 800 dollars per day! That’s an expensive shower. Could this be made more efficient? I don’t think so.

No, obviously it can. My garage demo has very little going for it in terms of efficiency. It’s about as basic as distillation gets. There’s lost heat going everywhere. Modern distillation setups are much more efficient at separating liquids, especially because they can take advantage of waste heat. In fact they are often co-located with coal or gas-fired power plants for this exact reason. And there’s a lot of technology just in minimizing the energy consumption of distillation, including reuse of the heat released during condensation, using stages to evaporate liquids more efficiently, and using pumps to lower the pressure and encourage further evaporation through mechanical means. But the thermal efficiency isn’t the only challenge with distillation.

Take a look at the flask that held the seawater after all the water boiled away and you can see the salt deposits building up, even after distilling only a small amount of water. These scale deposits reduce the efficiency of boiling because heat doesn’t transfer through them very easily, which means they would have to be cleaned off regularly. One alternative is a flash evaporator that sends the liquid stream through an expansion valve to force it to evaporate at temperatures lower than boiling, which minimizes the buildup of scale. Flash evaporators are the workhorses of desalination plants that use distillation, and especially in the middle east, plants like this have been reliably producing fresh water for decades now, but they’re not the only way to get the job done.

The other primary type of desalination uses membranes. You may have heard of the phenomenon called osmosis, where a solution naturally diffuses through a barrier. But you can reverse the osmotic process, moving a solution from high concentration to low with pressure… usually a lot of pressure. Let me show you what I mean. Luckily there are commercially available seawater membranes that don’t cost an arm and a leg. That’s because these systems are frequently used in boats and ships to make freshwater while at sea. But why spend thousands of dollars on a working watermaker when you have the rudimentary plumbing skills of a civil engineer?

Here’s the membrane I’m using for this demo. It’s wrapped in a spiral so you get lots of surface area in a small package. It is kind of like a filter that lets water pass through while holding back the dissolved solids, but at a much tinier scale. It’s generally a lot more efficient than thermal distillation, so most modern desalination plants use reverse osmosis (or RO) for primary separation. But, as you’ll see, it still uses a lot of energy, way more than a typical raw water treatment plant.

It takes a lot of pressure to force seawater through a membrane, in my case about 600 psi or 40 times normal atmospheric pressure. Even small RO systems use high-pressure pumps designed for continuous use, because this is not a fast process. Instead of springing for a nice pump well-suited for the application, I’m using the cheapest power washer I could find at the local hardware store. The instructions didn’t say not to run saltwater through it.

The membrane sits inside this high pressure housing that keeps it from unraveling under the immense forces inside. That’s if you hook everything up correctly… I had to redo a few connections when the housing sprung a leak during early testing.

A booster pump delivers the seawater from the bucket to the pressure washer, then the pressure washer sends it into the housing. Unlike a typical filter, not all the feed water flows through the membrane. Instead, most of it flows past the membranes and comes out on the other side just a little bit more concentrated with salt. This is called the brine and we’ll talk more about it in a minute. The water that does make it through the membrane, called the permeate, comes out in the center of the housing. You can see on my flow meters that, if I close the valve on the brine discharge line, it increases the pressure in the housing, forcing more of the water through the membrane. The meter on the left is brine discharge, and the one on the right is the permeate line. As I close the valve, the brine flow goes down and the permeate flow goes up. Of course I could close the brine flow all the way down, but you still need some water to carry the salt away or it will just foul up the membrane.

Typically you need to run water through these membranes for several hours before they settle into their best performance. My little power washer wasn’t quite up to the task of running for that long, but even after roughly half an hour, I was getting water with one to two parts per thousand of dissolved solids through this crude setup. That’s not high quality drinking water, but it’s definitely drinkable!

I ran this experiment a few times at different pressures, but the results didn’t vary too much. For this run, the combined power for the booster pump and the pressure water was around 1200 watts, and it took about five minutes to produce a liter (or quarter of a gallon). Going back to our residential household, it would take four pressure washers running non-stop and consume more than 100 kilowatt-hours in a day. That’s a huge improvement over the distillation demo, even considering the water quality wasn’t quite as good, but it’s still 15 dollars a day or more than 5,000 dollars per year just to separate salt from water.

It won’t surprise you to learn that, just like my crude distillation demo, my reverse osmosis via pressure washer demo is also not nearly as efficient as it could be on a larger scale. Modern RO plants use huge racks of high quality membrane units and high efficiency pumps. They also recover the energy from the brine stream before it leaves the system back out to sea, saving the precious kilowatt-hours already consumed by the pumps. To separate a cubic meter or 264 gallons, of seawater from its salt, my power washer RO system would take about a hundred kilowatt-hours. The newest RO plants can do it with just one or two.

But, even though the separation step is energy intensive, it’s not the only energy requirement in a seawater desal plant, and it’s definitely not the only cost. I’m using tap water in my demonstration, but these plants don’t start with that. Raw seawater not only has salt, but also dirt, algae, organic matter, and other contaminants too. Those constituents can foul or damage evaporators or membranes, so all desal plants use a pretreatment process to remove them first. That takes energy and cost to keep up with the various chemical feeds and filters before the water even reaches the salt separation process. And, even with good pretreatment, the RO membranes or evaporators have to be taken out of service for cleaning regularly, and eventually they have to be replaced. Additionally, you usually can’t send RO permeate or distilled water directly to customers. It’s too clean! It normally goes through a post-treatment process to add minerals, since most people prefer the taste over just pure water. Plus it gets disinfectant so that it can’t be contaminated on its way through the distribution system. And don’t forget about that brine.

All that salt that didn’t come out of the product stream is now packed into a smaller volume of water, making it more concentrated than before. Modern desalination plants generally recover about half of the intake flow, which means their brine stream is about twice the concentration of normal seawater. It’s a waste product that is actually pretty tough to get rid of. You can’t just discharge that super-saline waste directly back into the sea because of the environmental impacts, particularly on the plants and animals near the sea floor (since the concentrated solution usually sinks). To avoid environmental impacts, most brine discharge lines either use diffusers to spread out the salty solution so it dilutes faster or they blend the brine with some other stream of water like power plant cooling lines or wastewater effluent so it’s diluted before being released. When that’s not possible, some plants have to inject the saltwater into the ground (an expensive endeavor that only adds to operational costs).

With all the complications of separating salt from seawater, it’s easy to let one’s mind drift toward alternatives like harnessing renewable sources of energy. Like, what if we could use solar power to not only distill seawater but also carry it inland toward major cities and release it onto the ground where it could easily be collected. But now we’ve just re-invented the water cycle, which is already how we humans get the vast majority of the water we use to drink, cook, and bathe. It’s not like dams, reservoirs, canals, pumping stations, and surface water intakes don’t have their own enormous costs and environmental impacts. But, if mother nature isn’t dropping enough water for your particular populated area, you can build and operate a pretty long pipeline for the immense costs and energy required to desalinate seawater.

And that’s the problem with desalination. It’s kind of like the nuclear power of water supply. It seems so simple on the surface, but when you add up all the practical costs and complexities, it gets really hard to justify over other alternatives. It’s also harder to compare costs between those alternatives because of desal’s unique problems. It’s just a newer technology, so it’s harder to predict hidden technical, legal, political, and environmental challenges. For example, because of the high energy demands, desalination can strongly couple water costs with electricity costs. During a drought, the cost of hydropower goes up because there’s less water available, increasing overall energy costs and thus making desalination less viable right when you need it most.

Of course, desalination is a viable solution in many situations, especially in places with large populations and severe water scarcity. All the biggest plants are in middle eastern countries like Saudi Arabia and the UAE. That’s because they really have no choice. But it can also be viable in areas with a lot of variability in climate like California, Texas, and Florida. In these cases, a desalination plant is just one element in a diverse portfolio of resources, all with different risk profiles. Yes, the desalinated water is more expensive than other options like rivers, reservoirs, and groundwater supplies. But it can be more reliable too, providing water during drought conditions when the other sources are limited or completely unavailable. And, a lot of these costs and complexities get simpler when you’re not pulling salt out of seawater. There are sources of water that have some salt (but not as much as the ocean) like estuaries and brackish groundwater. In places where such a supply is available, desalination can be a much more cost effective source of fresh water.

Another way to make desal projects more viable is to let the private sector take on the risks. Many of the largest desalination plants are partnerships with private water companies rather than being financed, built, and operated by the utility like what’s done for a typical treatment plant. Partnering with a private company allows a utility to offload the financing costs and operational risks in return for the stability of a simple water purchase agreement. You pay for it, build it, and operate it, and we’ll just buy the water from you. This type of arrangement also keeps government boards from having to weigh in on complicated technical issues and innovations where there’s just not as much precedence to lean on as there is with more established types of water infrastructure projects.

The private company running the Carlsbad plant in San Diego County I mentioned earlier is working on a major project scheduled to finish in 2024: a new standalone seawater intake required after the power plant next door shut down in 2018. Bonds issued for the project were upgraded to rating of triple-B by Fitch, meaning the facility has a relatively stable outlook with a lower chance of defaulting. That’s just one rating agency’s assessment of just one project on just one membrane plant, but it gives some confidence that the technology of desalination is making progress, and that it might become a bigger and bigger part of the world’s limited supply of fresh water in the future.

Was Starship’s Stage Zero a Bad Pad?

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

On April 20, 2023, SpaceX launched it’s first orbital test flight of its Starship spacecraft from Boca Chica on the gulf coast of Texas. You probably saw this, if not live, at least in the stunning videos that followed. Thanks to NASA Space Flight for giving me permission to use their footage in this video. Starship launched aboard the Super Heavy first stage booster, and was the tallest and most powerful rocket ever launched at the time. There was no payload; this was a test flight with the goal of gathering data not just on the rocket, but all the various systems involved. The rocket itself was exciting to watch: some of the engines failed to ignite, and a few more flamed out early during the launch. About 40 kilometers above the ground, the rocket lost steering control and the flight termination system was triggered, eventually blowing up the whole thing.

But a lot of the real excitement was on the ground. Those Raptor engines put out about twice the thrust of the Saturn V rocket used in the Apollo Program. And, all that thrust, for several moments, was directed straight into the concrete base of the launch pad, or as SpaceX calls it, Stage Zero. And that concrete base wasn’t really up for the challenge. Huge chunks of earth and concrete can be seen flying hundreds of meters through the air during the launch, peppering the gulf more than 500 meters away. A fine rain of debris fell over the surrounding area, and the damage seen after the road opened back up was surprising. Tanks were bent up. Debris was strewn across the facility. And the launch pad itself now featured an enormous crater below.

Although the FAA’s mishap report hasn’t been released yet, there’s plenty of information available to discuss. Rocket scientists and aeronautical engineers get a lot of well-deserved attention on youtube and around the nerdy content-sphere, but when it comes to the design and construction of launchpads like stage zero, that’s when civil engineers get to shine! What happened with Stage Zero and how do engineers design structures to withstand some of the most extreme conditions humans have ever created? I’m Grady, and this is Practical Engineering. Today we’re talking about launch pads.

Humans have been launching spacecraft for over 65 years now. And so far, pretty much the only way we have to propel a payload to the incredibly high speeds and altitudes that task requires, is rockets. Rockets produce enormous amounts of thrust by burning fuel and oxidizer in what amounts to a carefully (or not so carefully) controlled explosion. By throwing all that mass out the back, they’re able to accelerate forwards. But what happens to that mass once it’s expelled? When the rocket is flying through the sky, the gasses from its engines eventually slow and dissipate into the atmosphere. But, most rockets (especially the big ones), don’t get to start in the sky. Instead, they’re launched from the earth’s surface, and the small part of the earth’s surface directly below them can take a heck of a beating. Hot and corrosive gasses move at incredible speeds and often carry abrasive particles along with them. To call a rocket launch “thunderous” is often an understatement, because the sound waves generated are louder than a lightning strike and they last longer too.

Dealing with these extreme loading conditions isn’t your typical engineering task. It’s niche work. You’re not going to find a college course or textbook covering the basics of launch pad design. Instead, engineers who design these structures work from multiple directions. They use first principles to try and bracket the physics of a launch. They look at what’s worked and what hasn’t worked in the past. They use computational fluid dynamics, in other words, simulations, that help characterize the velocities and temperatures and sound pressures so that they can design structures to withstand them. But eventually, you have to use tests to see if your intuitions and estimations hold up in the real world.

It’s no surprise that one of the world leaders in successful launchpad engineering is NASA. And their historic Launchpad 39A in Cape Canaveral, Florida is a perfect case study. This pad, and it’s sister 39B, were originally built for the enormous Saturn V rocket, the cornerstone of the Apollo Space Program that first sent astronauts to the moon. Just like the SpaceX facility in Boca Chica, 39A is situated on a coast with the water out to the east. Most rockets launch in that direction to take advantage of earth’s rotation. The earth itself is already rotating to the east, so it makes sense to go ahead and take advantage of that built-in momentum. But some rockets blow up before they make it into space. So it’s best to choose a launch location with a huge stretch of unpopulated area to the east, like an ocean!

Launch Complex 39 was constructed on Merrit Island, a barrier island east of Orlando. NASA decided early on that water was the best way to move the first and second stages of the Saturn V rocket, so several miles of canals were dredged out. Over three quarters of a million cubic yards of sand and shells were produced by this dredging and used as fill for construction. Some of that material was used to build a special road called a crawlerway connecting the Vehicle Assembly Building to the launchpad. But a lot of it was used to construct a flat topped pyramid 80 feet or 24 meters tall. This structure would ultimately become the launchpad. If you’re a fan of the channel, you might already be thinking what I’m thinking. Huge piles of material like this settle over time, and I have a video all about that you can check out after this! NASA engineers let this structure sit before the rest of the launchpad was built. It’s a good thing too, because it settled about 4 feet, well over a meter!

Why did NASA bother building such a massive hill when they could have simply built the pad on the existing ground? It was all about the flame deflector: a curved steel structure that would redirect the tremendous plume of rocket exhaust exiting the Saturn V during launch into a monumental concrete trench. This would keep the plume from damaging the sensitive support structures around the pad or undermining its foundation.

But why not put the trench into the existing ground rather than building a massive artificial hill? The answer is groundwater. Siting a launch pad so near to the coast comes with the challenge of being basically at sea level. If you’ve ever dug a hole at the beach, you know the exact problem the launchpad engineers were facing. Imagine trying to install expensive and delicate technology inside that hole. Of course we build structures below the water table all the time, and I have a video about that topic too. But with the cost and complexity of dewatering the subsurface, especially considering the extreme environment in which pumps and pipes would be required to operate, it just made more sense to build up. On top of that gigantic hill, thousands of tons of concrete and steel were installed to bear the loads of the launch support structures, the weight of the rocket itself while filled with thousands of pounds of fuel and oxidizer, and of course the dynamic forces during a full scale launch.

But that’s not all. Along with the enormous flame trench, and the associated flame diverters, which have gone through various upgrades throughout the years, NASA employed a water deluge system. This is a test of the current system on pad 39B. During a launch, huge volumes of water are released through sprayers to absorb the heat and acoustic energy of the blast, further reducing the damage it causes on the surrounding facility. Check out this incredible historical slow-motion footage of a Space Shuttle Launch. You’ll notice a copious flow of water both under the main engines on the right, and under the enormous solid rocket boosters on the left. In fact, a lot of the billowing white clouds you see during launches are from the deluge system as water’s rapidly boiled off by the extreme temperatures.

39A has seen a lot of launches over the years, more than 150. The first launch was the unmanned Apollo 4 in 1967, the first ever launch of the Saturn V. The bulk of the moon missions and space shuttles launched from 39A, and more recently SpaceX themselves have launched dozens of their Falcon 9 and several Falcon Heavy rockets from the historic pad! But when you compare it to the Stage Zero structure in Boca Chica, at least its configuration during the first orbital test, the differences are obvious: No flame diverter; no water deluge system; just the world’s most powerful rocket pointed square at a concrete slab on the ground. And, I think the results came as a surprise to no one who pays attention to these things. Elon himself tweeted in 2020 that leaving out the flame diverter could turn out to be a mistake.

That concrete, by the way, isn’t just the ready-mix stuff you buy off the shelf at the hardware store. I have a whole video about refractory concrete that’s used to withstand the incredible heat of furnaces, kilns, and rocket launches. This concrete has to be strong, erosion resistant, insulating, resistant to thermal shock, and immune to exposure from saltwater since launchpads are usually near the coast. NASA used a product called Fondu Fyre at 39A and SpaceX uses Fondag. But even that fancy concrete was no match for those raptor engines. Even during the static test fire, there was some damage to the concrete pad, and that was only at about half power. The orbital test and the full force of the rocket completely disintegrated the protective pad and cratered the underlying soil, spraying debris particles for miles.

In a call after the launch, Elon said that, although things looked bad on the surface, the damage to the launch pad could be repaired pretty quickly, noting that the outcome of the test was about what he expected. And even though many might have expected the extensive damage to the pad and surrounding area, it sure wasn’t mentioned in the Environmental Assessment required before SpaceX could get a license to perform the test, whose sole purpose was to document all the environmental impacts that would be associated with building the facility and launching rockets there. Nowhere in the nearly 200-page report is a discussion of the enormous debris field that resulted from the test, and yet there are actually quite a few laws against stuff like this.

For just one example, there are federal rules about filling in wetlands, of which there are many surrounding the launch facility. If you can’t do it with a bulldozer, you probably can’t do it with a rocket, and spraying significant volumes of soil and concrete into the surrounding area likely has the regulator’s attention for that reason alone, not to mention the public safety aspects of the showering debris. The launch also caused a fire in the nearby state park. The FAA has effectively grounded Starship pending their mishap investigation, and several environmental groups have already sued the FAA over the fallout of the launchpad’s destruction.

Even if the FAA comes back with no required changes moving forward, SpaceX themselves aren’t planning to do that again, and they’ve already shared their plans for the future. An enormous, watercooled steel plate design is already well under construction as of this writing. This design is, again, very different than what we see at other launch pads, basically an upside-down shower head directly below the vehicle. That’s the nature of SpaceX and why many find them so exciting. Unlike NASA that spends years in planning and engineering, SpaceX uses rapid development cycles and full-scale tests to work toward their eventual goals. They push their hardware to the limit to learn as much as possible, and we get to follow along. They’re betting it will pay off to develop fast instead of carefully. But this wasn’t just a test of the hardware. It was also a test of federal regulations and the good graces of the people who live, work, play, and care about the Boca Chica area. And, SpaceX definitely pushed those limits as well with their first orbital test. It’s still yet to be seen what they’ll learn from that.

How Flood Tunnels Work

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

This is Waterloo Park in downtown Austin, Texas, just a couple of blocks away from the state capitol building. It’s got walking trails, an ampitheater, Waller Creek runs right through the center, and it has this strange semicircular structure right on the water. And this is Ladybird Lake, formerly Town Lake, about a mile away. Right where Waller Creek flows into the lake, there’s another strange structure. You saw the title of this video, so you know what I’m getting at here. It turns out these two peculiar projects are linked, not just by the creek that runs through downtown Austin, but also by a tunnel, a big tunnel. The Waller Creek Tunnel is about 26 feet (or 8 meters in diameter) and runs about 70 feet or 21 meters below downtown Austin. It’s not meant for cars or trains or bikes or buses or even high voltage oil filled cables, and it’s not even meant to carry fresh water or sewage. Its singular goal is to quickly get water out of this narrow downtown area during a flood. It’s designed with a peak flow rate of 8,500 cubic feet per second. That’s 240 cubic meters per second, or enough to fill a cubic olympic sized swimming pool in about 10 seconds. And the way it works is pretty fascinating.

Most major cities use underground pipes as drains to get rid of stormwater runoff so it doesn’t flood streets and inundate populated areas. But, a storm drain only has so much capacity, and a lot of places across the world have taken the idea a few steps further in scale. As I always say, the only thing cooler than a huge tunnel is a huge tunnel that carries lots of water and protects us from floods. And I built a model flood tunnel from acrylic, so you can see how these structures work and learn just a few of the engineering challenges that come with a project like this. I’m Grady, and this is Practical Engineering. In today’s episode, we’re talking about flood tunnels.

Floods are natural occurrences on earth, and in fact, in many places they are beneficial to the environment by creating habitat and carrying nutrient rich sediments into the floodplain, the area surrounding a creek or river that is most vulnerable to inundation. But, floods are not beneficial to cities. They are among the most disruptive and expensive natural disasters worldwide. If a flood swells a creek or river in a scattered residential neighborhood, it’s not ideal for the few homeowners who are impacted, but if a flood strikes the dense urban core of a major city, the consequences can be catastrophic with millions of dollars of damage and entire systems shutdown. What that means in practice is that we’re often willing to spend millions of dollars on flood infrastructure to protect densely populated areas, opening the door to more creative solutions. And heavily developed downtown areas demand resourceful thinking because they lack the space for traditional protection projects and they often predate modern urban drainage practices.

We can’t change the amount of water that falls during a flood, so we’re forced to develop ways to manage that water once it’s on the ground. The main way we mitigate flooding is just to avoid development within the floodplain. Don’t build in the areas of land most at risk of inundation during heavy storms. Seems simple, but it’s not an option for most downtown areas that have been developed since well before the advent of modern flood risk management. Another way we manage flooding is storing the water in large reservoirs behind dams, allowing it to be released slowly over time. Again, not an option in downtown areas where creating a reservoir could mean demolishing swaths of expensive property. A third flood management strategy is bypassing - sending the water around developed areas where it will cause fewer impacts. Once again, not an option in downtown areas where there is no alternative path for the water to go… unless you start thinking in the third dimension. Tunnels allow us to break free from the confines of the earth’s surface and utilize subterranean space to allow floodwaters past developed areas to be released further downstream. Let me show you how this works.

This is my model downtown business district. It’s got buildings, landscaping, and a beautiful river running right through the center. I have a flow meter and valve to control how much water is moving through that beautiful river, and here on the downstream side is a little dam to create some depth. Take a look at many major cities that have rivers running through them, and you’ll often see a weir or dam just like this to maintain some control over the upstream level, keeping water deep enough for boats or in some cases, just for beauty like the RiverWalk in downtown San Antonio. I put some blue dye and mica powder in the water to make it easier to visualize the flow.

I also have a big clear pipe with an inlet upstream of the developed area and an outlet just below the dam. Looking at this model, it might seem like a flood bypass tunnel is as simple as slapping a big pipe to where you want the flood waters to go, but here’s the thing about floods: most of the time, they’re not happening. In fact, almost all of the time, there isn’t a flood. And if you’re the owner of a flood bypass tunnel, that means almost all of the time you’re responsible for a gigantic pipe full of water below your city that has no real job except to wait. Watch what happens when I turn down the flow rate in my model to something you might see on a typical day. If we just leave the city like it is, all the flow goes into the tunnel, draining the channel like a bathtub and leaving the water along the downtown corridor to stagnate.

Standing water creates an environmental hazard. Without motion, the water doesn’t mix, and so it loses dissolved oxygen that is needed for fish and bacteria that eat organic material. Without dissolved oxygen, rivers become dead zones with little aquatic life and full of smelly, rotting organic material. Stagnant water also creates a breeding ground for mosquitoes, and is just unpleasant to be around. It’s not something you want in an urban core. The answer to this issue is gates, a topic I have a whole other video about. I can show how this works in my model. If you equip your gigantic flood bypass tunnel with gates on the inlet, you can control how much water goes into the tunnel versus what continues in the river. I just used this piece of foam to close off most of the tunnel entrance. I still have some water moving through there, but most continues in the river, keeping it from getting stagnant. This is why, on many flood bypass tunnels, you’ll see interesting structures at the inlets. Here’s the one in Austin again, and here’s the one just down the road in San Antonio. In addition to screening for trash and debris (and keeping people out) the main purpose of these structures is to regulate how much water goes into the tunnel.

But, some creeks and rivers don’t just have low flows during dry times, they have no flows. Intermittent streams only flow at certain times of the year and ephemeral streams only flow after it rains. Take a look at the stream gage for Waller Creek in Austin. Except for the days with rain, the flow in the creek is essentially zero. But, if you’re worried about stagnant water and lack of habitat on the surface, you want more water running in the river. You definitely don’t want to divert any of the scarce flows available into the tunnel. But you can’t just close the tunnel off completely, because then the water inside the tunnel will stagnate instead. You might think, “So what? It’s down there below the ground where we don’t have to worry about it.” Well, as soon as the next big flood comes and you open the gates to your tunnel, you’re going to push a massive slug of disgusting stagnant water out the other end, creating an environmental hazard downstream. So, in addition to gates on the upstream end, some flood tunnels, including the one in Austin, are equipped with pumps to recirculate water back upstream. I put a little pump in the model to show how this works. The pump pulls water from the river downstream and delivers it back upstream of the tunnel entrance. This allows you to double dip on benefits during low flows: you keep water moving in the tunnel so it doesn’t stagnate and you actually increase the flow in the river, improving its quality.

That’s 99 percent of managing a flood bypass tunnel: maintaining the infrastructure during normal flows. But of course, all that trouble is worth it the moment a big flood comes. Let’s turn the model all the way up and see how it performs. You can see the tunnel collecting flows, moving them downstream, and delivering them below the dam away from the developed area. The tunnel is adding capacity to the river, allowing a good proportion of the flood flows to completely bypass the downtown area. Of course, the river still rose during the flood, but it hasn’t overtopped the banks, so the city was protected. Let’s plug the tunnel and see what would happen without it. Turning up the model to full blast causes the stream to go over the bank and flood downtown. In this case, it’s not a huge difference, but even a few inches of floodwaters backing up into buildings is enough to create enormous damages and huge costs for repairs. Without any margin for increased flows, a big peak in rainfall can even wash buildings and cars away.

So, comparing flood levels between the two alternatives flowing at the same rate, it’s easy to see the benefits of a flood bypass tunnel. It resculpts the floodplain, lowering peak levels and pulling property and buildings out of the most vulnerable areas, making it possible to develop more densely in urban areas, not to mention creating habitat, improving water quality, and maintaining a constant flow in the river during dry times. Of course, a tunnel is an enormous project itself, and flood bypass tunnels are truly one of the most complicated and expensive ways to mitigate flood risks, but they’re also one of the only ways to manage flood risks in heavily populated areas.

I’ve been referencing projects in central Texas because that’s where I live, but despite their immense cost and complexity, flood bypass tunnels have been built across the world. One of the most famous is the Tokyo Metropolitan Area Outer Underground Discharge Channel that features this enormous cathedral of a subsurface tank. Unlike my model that works by gravity alone, the Tokyo tunnel needs huge pumps to get the water back out and into the Edogawa River. And some tunnels aren’t just for stormwater. Many older cities don’t have separated sewers for stormwater and wastewater, so everything flows to the treatment plants. That means when it rains, these plants see enormous influxes of water that must be treated before it can be released into rivers or the ocean. One of the largest civil engineering projects on earth has been in design and construction in Chicago since the 1970s and isn’t scheduled for completion until 2029. The Tunnel and Reservoir Plan (or TARP) includes four separate tunnel systems that combine with a number of storage reservoirs to keep Chicago’s sewers from overflowing into and polluting local waterways. And we keep finding value in tunnels where other projects wouldn’t be feasible. After record breaking floods from Hurricane Harvey in 2017, Houston started looking into the viability of using tunnels to reduce the impacts from future downpours. A 2.5-million-dollar engineering study was finished in 2022 suggesting that a system of tunnels might be a feasible solution to remove tens of thousands of structures from the floodplain. If they do move forward with any of the eight tunnels evaluated, that will complete the superfecta of major metropolitan areas in Texas with large flood bypass tunnels, but represent just one more of the many cities across the world that that have maximized the use of valuable land on earth’s surface by taking advantage of the space underneath.

Merrimack Valley Gas Explosions: What Really Happened?

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

On September 13, 2018, a pipeline crew in the Merrimack Valley in Massachusetts was hard at work replacing an aging cast iron natural gas line with a new polyethylene pipe. Located just north of Boston, the original cast iron system was installed in the early 1900s and due for replacement. To maintain service during the project, the crew installed a small bypass line to deliver natural gas into the downstream pipe while it was cut and connected to the new plastic main line. By 4:00 pm, the new polyethylene main had been connected and the old cast iron pipe capped off. The last step of the job was to abandon the cast iron line. The valves on each end of the bypass were closed, the bypass line was cut, and the old cast iron pipe was completely isolated from the system. But it was immediately clear that something was wrong.

Within minutes of closing those valves, the pressure readings on the new natural gas line spiked. One of the fittings on the new line blew off into a worker's hand. And as they were trying to plug the leak, the crew heard emergency sirens in the distance. They looked up and saw plumes of smoke rising above the horizon. By the end of the day, over a hundred structures would be damaged by fire and explosions, several homes would be completely destroyed, 22 people (including three firefighters) would be injured, and one person would be dead in one of the worst natural gas disasters in American history. The NTSB did a detailed investigation of the event that lasted about a year. So let’s talk about what actually happened, and the ways this disaster changed pipeline engineering so that hopefully something like it never happens again. I’m Grady, and this is Practical Engineering. In today’s episode, we’re talking about the 2018 Merrimack Valley natural gas explosions.

Like many parts of the world, natural gas is an important source of energy in homes and businesses in the United States. It’s a fossil fuel composed mostly of methane gas extracted from geologic formations using drilled wells. The US has an enormous system of natural gas pipelines that essentially interconnect the entire lower 48 states. Very generally, gathering lines connect lots of individual wells to processing plants, transmission lines connect those plants to cities, and then the pipes spread back out again for distribution. Compressor stations and regulators control the pressure of the gas as needed throughout the system. Most cities in the US have distribution systems that can deliver natural gas directly to individual customers for heating, cooking, hot water, laundry, and more. It’s an energy system that is in many ways very similar to the power grid, but in many ways quite different, as we’ll see.

Just like a grid uses different voltages to balance the efficiency of transport with the complexity of the equipment, a natural gas network uses different pressures. In transmission lines, compressor stations boost the pressure to maximize flow within the pipes. That’s appropriate for individual pipelines where it’s worth the costs for higher pressure ratings and more frequent inspections, but it’s a bad idea for the walls of homes and businesses to contain pipes full of high-pressure explosive gas. So, where safety is critical, the pressure is lowered using regulators.

Just a quick note on units before we get too far. There are quite a few ways we talk about system pressures in natural gas lines. Low pressure systems often use inches or millimeters of water column as a measure of pressure. For example, a typical residential natural gas pressure is around 12 inches (or 300 millimeters) of water, basically the pressure at which you would have to blow into a vertical tube to get water to raise that distance: roughly half a psi or 30 millibar. You also sometimes see pressure units with a “g” at the end, like “psig.” That “g” stands for gauge, and it just means that the measurement excludes atmospheric pressure. Most pressure readings you encounter in life are “gauge” values that ignore the pressure from earth’s atmosphere, but natural gas engineers prefer to be specific, since it can make a big difference in low pressure systems.

The natural gas main line in the Merrimack Valley being replaced had a nominal pressure of 75 psi or about 5 bar, although that pressure could vary depending on flows in the system. Just for comparison, that’s 173 feet or more than 50 meters of water column. But, the distribution system, the network of underground pipes feeding individual homes and businesses, needed a consistent half a psi or 30 millibar, no matter how many people were using the system. The device that made this possible was a regulator. There are lots of different types of regulators used in natural gas systems, but the ones in the Merrimack valley use pilot-operated devices, which are pretty ingenious. It’s basically a thermostat, but for pressure instead of temperature. The pilot is a small pressure regulating valve that supports the opening or closing of the larger primary valve. If the pilot senses an increase or decrease in pressure from the set point, it changes the pressure in the main valve diaphragm, causing it to open or close. This all works without any source of outside power just using the pressure of the main gas line.

Columbia Gas’s Winthrop station was just a short distance south of where the tie-in work was being done on the day of the event. Inside, a pair of regulators in series was used to control the pressure in the distribution system. One of these regulators, known as the worker, was the primary regulator that maintained gas pressure. A second device, called the monitor, added a layer of redundancy to the system. The monitor regulator was normally open with a setpoint a little higher than the worker so it could kick in if the worker ever failed, and, at least in theory, make sure that the low-pressure system never got above its maximum operating level of about 14 inches of water column or 35 millibar. But, in this worker/monitor configuration, the pilots on the two regulators can’t use the downstream pressure right at the main valve. For one, the reading at the worker would be affected by any changes in the downstream monitor. And for two, measuring pressure right at the valve can be inaccurate because of flow turbulence generated by the valve itself. It would be kind of like putting your thermostat right in front of a register; it wouldn’t be getting an accurate reading. So, the pilots were connected to sensing lines that could monitor the pressure in the distribution system a little ways downstream of the regulator station.

The worker and monitor regulators were both functioning as designed on September 13, and yet, they allowed high pressure gas to flood the system, leading to a catastrophe. How could that happen? The NTSB’s report is pretty clear. Tying a natural gas line while it’s still in service, called a hot tie-in, is a pretty tricky job that requires strict procedures. Here are the basic steps: First a bypass line was installed across the upstream and downstream parts of the main line. Then balloons were inserted into the main to block gas from flowing into the section to be cut. Once the gas was purged from the central section, it was cut out and removed while the bypass line kept gas flowing from upstream to downstream. The line to be abandoned got a cap, and the new plastic tie in was attached to the downstream main. Once the tie-in was complete, the crew switched the upstream gas service from the old cast iron line over to the new plastic line and deflated the last balloon so that gas could flow. The upstream cast iron line was still pressurized, since it was still connected to the in-service line through the bypass. But, as soon as the crew closed the valves on the bypass, the old cast iron line was fully isolated, and the pressure inside the line started to drop, as planned.

What that crew didn’t know is that when that plastic main line was installed 2 years back, a critical error had been made. The main discharge line at the regulator station had been attached to the new polyethylene pipe, but the sensing lines had been left on the old cast iron main. It hadn’t been an issue for the previous 2 years, since both lines were being used together, but this tie-in job was the first of the entire project that would abandon part of the original piping. Within minutes of isolating the old cast iron pipe, its pressure began to drop. To a regulator, there’s no difference between a pressure drop from high demands on the gas system and a pressure drop from an abandoned line, and they respond the same way in both cases: open the valves. In a normal situation, the increased gas flow would result in higher pressure in the sensing lines, creating a feedback loop. But this was not a normal situation. It’s the equivalent of putting your thermostat in the freezer. Even as pressure in the distribution system rose, the pressure in the sensing lines continued to drop with the abandoned line. The regulators, not knowing any better, kept opening wider and wider, eventually flooding the distribution system with gas at pressures well above its maximum rating.

By the time things went sideways, the crew at the tie-in had taken most of their equipment out of the excavation. But as one worker was removing the last valve, it blew off into his hand as gas erupted from the hole. The crew heard firefighters racing throughout the neighborhood and saw the smoke from fires across the horizon. The overpressure event had started a chain of explosions, mostly from home appliances that weren’t designed for such enormous pressures. The emergency response to the fires and explosions strained the resources of local officials. Within minutes, the fire departments of Lawrence, Andover, and North Andover had deployed well over 200 firefighters to the scenes of multiple explosions and fires, and help from

neighboring districts in Massachusetts, New Hampshire, and Maine would quickly follow. The Massachusetts Emergency Management Agency activated the statewide fire mobilization plan, which brought in over a dozen task forces in the state, 180 fire departments, and 140 law enforcement agencies. The electricity was shut off to the area to limit sources of ignition to help prevent further fires, and of course, natural gas service was shut off to just under 11,000 customers.

By the end of the day, one person was dead, 22 were injured, and over 50,000 people were evacuated from the area. And while they were allowed back into their homes after three days, many were uninhabitable. Even those lucky enough to escape immediate fire damage were faced with a lack of gas service as miles of pipelines and appliances had to be replaced. That process ended up taking months, leaving residents without stoves, hot water, and heaters in the chilly late fall in New England.

NTSB had several recommendations stem from their investigation. At the time of the disaster, gas companies were exempt from state rules that required the stamp of a licensed professional engineer on project designs. Less than three months after NTSB recommended the exemption be lifted, a bill was passed requiring a PE stamp on all designs for natural gas systems, providing the public with better assurance that competent and qualified engineers would be taking responsibility for these inherently dangerous projects. And actually, NTSB issued the same recommendation and sent letters to the governors of 31 states with PE license exemptions, but most of those states still don’t require a PE stamp on natural gas projects today. There were recommendations about emergency response as well, since this event put the area’s firefighters through a stress test beyond what they had ever experienced.

NTSB also addressed the lack of robustness of low pressure gas systems where the only protection against overpressurization is sensing lines on regulators. It’s easy to see in this disaster how a single action of isolating a gas line could get past the redundancy of having two regulators in series and quickly lead to an overpressure event. This situation of having multiple system components fail in the same way at the same time is called a common mode failure, and you obviously never want that to happen on critical and dangerous infrastructure like natural gas lines. Interestingly and somewhat counterintuitively, one solution to this problem is to convert the low-pressure distribution system to one that uses high pressure. Because, in this kind of system, every customer has their own regulator, essentially eliminating the chance of a common mode failure and widespread overpressure event.

Most importantly, the NTSB did not mince words on who they found at fault for the disaster. They were clear that the training and qualification of the construction crew, or the condition of the equipment at the Winthrop Avenue regulator station were NOT factors in the event. Rather, they found that the probable cause was Columbia Gas of Massachusetts’ weak engineering management that did not adequately plan, review, sequence, and oversee the project.

To put it simply, they just forgot to include moving the sensing lines when they were designing the pipeline replacement project, and the error wasn’t caught during quality control or constructability reviews. NiSource, the parent company of Columbia Gas (of Massachusetts), estimated claims related to the disaster exceeded $1 billion, an incredible cost for weak engineering management. Ultimately, Columbia Gas pleaded guilty to violating federal pipeline safety laws and sold their distribution operations in the state to another utility. They also did a complete overhaul of their engineering program and quality control methods.

All those customers hooked up to natural gas lines didn’t have a say in how their gas company was managed; they didn’t have a choice but to trust that those lines were safe; and they probably didn’t even understand the possibility that those lines could overpressurize and create a dangerous and deadly condition in the place where they should have felt most safe: their own homes. The event underscored the crucial responsibility of engineers and (more importantly) the catastrophic results when engineering systems lack rigorous standards for public safety.

Why Bridges Need Sensors (and other structures too)

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

Almost immediately after I started making videos about engineering, people started asking me to play video games on the channel. Apparently there’s roughly a billion people who watch online gaming these days, and some of them watch silly engineering videos too! And there’s one game that I get recommended even more than minecraft: Polybridge. So I finally broke down one evening after the kids went to bed and gave it a try. I’m really not much of a gamer, but I have to admit that I got a little addicted to this game (hashtag not-an-ad). I admit too that there really is a lot of engineering involved. You have different materials that give your structure different properties. The physics are RELATIVELY accurate. You get a budget to spend on each project. And your score is based on the efficiency of your design. But there’s one way this game is not like real structural engineering at all: if your bridge collapses, you get to try again!

In the real world, we can’t design a dam, a building, a transmission line pylon, or a bridge, spend all that money to build it, watch how it performs, tear it down, and build it back better if we’re not happy with the first iteration. Structures have to work perfectly on the first try. Of course we have structural design software that can simulate different scenarios, but it’s only as powerful as your inputs, which are often just educated guesses. We don’t know all the loads, all the soil conditions, or all the ways materials and connections will change over time from corrosion, weathering, damage, or loading conditions. There are always going to be differences between what we expect a structure to do and what actually happens when it gets built. Hopefully engineers use factors of safety to account for all that uncertainty, but you don’t have to dig too deep into the history books to find examples where an engineer neglected something that turned out to matter a lot, sometimes to the detriment of public safety. So what do you do?

We can’t build a project then watch the cars and trucks drive over with the pretty green and red colors on each structural member to see how they’re performing in real time… except you kind of can, with sensors. It turns out that plenty of types of infrastructure, especially those that have serious implications for public safety, are equipped with instruments to track their performance over time and even save lives by providing an early warning if something is going wrong.

I love sensors. To me, it’s like a superpower to be able to measure something about the world that you can’t detect with just your human senses. Plus I’m always looking for an opportunity to exercise my inalienable right to take measurements of stuff and make cool graphs of the data. So I have a bunch of demonstrations set up to show you how engineers employ these sensors to compare the predicted and actual performance of structures, not just for the sake of delightful data visualization, but sometimes even to save lives. I’m Grady, and this is Practical Engineering. In today’s episode, we’re talking about infrastructure instrumentation.

And what better place to start than with a big steel beam? In fact, this is the biggest steel beam that my local metals distributor would willingly load on top of my tiny car. One of the biggest questions in polybridge and real world engineering is this: How much stress is each structural member experiencing? Of course, this is something we can estimate relatively quickly. So let’s do the engineer thing and predict it first. Beam deflection calculations are structural engineering 101, so we can do some quick recreational math to predict how much this thing flexes under different amounts of weight. And we can use my weight as an example: about 180 pounds or 82 kilograms. The calculation is relatively simple. You can choose your preferred unit system and pause here if you want to go through them. Standing at the beam’s center, I should deflect it by about 2 thousandths of an inch or about 60 microns, around the diameter of the average human hair. In other words, I am a fly on the wall of this beam (or really a fly on the flange). I’m barely perceptible. In fact, it would take more than 100 of me to deflect this beam beyond what would normally be allowed in the structural code. And it would take a lot more than that to permanently bend it. But 2 thousandths of an inch isn’t nothing, so, let’s check our math.

I put my dial indicator underneath the beam, and added some weight. I started with 45 pound or 20 kilogram plates. Each time I add one, you see the beam deflect downward just a tiny bit. After three plates, I added myself, bringing the total up to around 315 pounds or 143 kilos of weight. And actually, the deflection measured by the dial indicator came pretty close to the theoretical predictions made with the simple formula. Here they are on a graph, and there’s the point at my weight, with a deflection of around 2 thousandths of an inch or 60 microns, just like we said. But, we can’t always use dial indicators in the real world because they need a reference point, in this case, the floor. Up on the superstructure of a bridge, there’s no immovable reference point like that. So an alternative is to use the beam itself as a reference. That’s how a strain gauge works, and that’s the cylindrical device that I’m epoxying to the bottom flange of my beam.

A strain gauge works by measuring the tiny change in distance between two parts of the steel. You might know that when you apply a downward load to a beam, it creates internal stress. At the top, the beam feels compression, and at the bottom it feels tension. But it doesn’t just feel the stress, it also reacts to it by changing in shape. Let me show you what I mean. When I put one of the plates on top of the beam, we can see a change in the readout for the strain gauge. (Of course, I had the gauge set to the wrong unit, so let me overlay the proper one with the magic of video compositing.) For each plate I add to the beam, we see that the flange actually lengthens, in this case by about 3 microstrain. That’s probably not a unit of measure you’re familiar with, but it really just means the bottom of the beam increased in length by 0.0003%. When I add another weight, we make it 0.0003% longer again. Same with the third weight. And then when I stand on top of the whole stack, we get a total strain of about 0.002%, a completely imperceptible change in shape to the human eye, but the strain gauged picked it up no problem.

Imagine how valuable it would be to an engineer to have many of these gauges attached to the myriad of structural members in a complicated bridge or building and be able to see how each one responds to changes in loading conditions in real time. You could quickly and easily check your design calculations to make sure the structure is behaving the way you expected. In my simple example in the studio, the gauge is measuring pretty much exactly what the predictions would show, but consider a structure far more complicated than a steel beam across two blocks, in other words, any other structure. What factors get neglected in that simple equation I showed earlier?

We didn’t consider the weight of the beam itself; I’m not actually a one-dimensional single point load, like the equation assumes, but rather my weight is spread out unevenly across the area of my sneakers; Is the length exactly what we entered into the equation? And, what about three-dimensional effects? For example, I put another strain gauge on the top flange of the beam. If you just follow the calculations, you would assume this flange would undergo compression, getting a tiny bit shorter with increased load. But really what happens in this flange depends entirely on how I shift my weight. I can make the strain go up or down simply by adjusting the way I stand on top, creating a twisting effect in the beam, something that would be much more challenging for an engineer to predict with simple calculations. Putting instruments on a structure not only helps validate the original design, but provides an easy way to identify if a member is overloaded. So it’s not unusual for critical structures to be equipped with instruments just like this one, with engineers regularly reviewing the data to make sure everything is working correctly.

Of course, we don’t only use steel in infrastructure projects, but lots of concrete too. And just like steel, concrete structures undergo strain when loaded. So I took a gauge and cast it into some concrete to measure the internal strain of the material. This is just a typical concrete beam mold and some ready-mix concrete from the hardware store. And even before we applied any load, the gauge could measure internal strain of the concrete from the temperature changes and chemical reactions of the curing process. Shrinkage during curing is one of the reasons that concrete cracks, after all. Luckily my beam stayed in one piece. Once the beam had cured and hardened for a few weeks, I broke it free from the mold. Compared to steel, concrete is a really stiff material, meaning it takes a lot of stress to cause any kind of measurable strain. So I got out my trusty hydraulic press for this one. I slowly started adding force from the jack, then letting the beam sit so the data logger could take a few readings from the strain gauge inside. After the fourth step, at just over 50 microstrain, the beam completely broke. Hopefully you can see how useful it might be to have an embedded sensor inside a concrete slab or beam, tracking strain over time, and especially when you know about the amount of strain that corresponds to the strength of the material. This is information that would be impossible to know without that sensor cast into the concrete, and there’s something almost magical about that. It’s like the civil engineering equivalent of x-ray vision.

One of the most amazing things about these sensors is their ability to measure tiny distances. 1 microstrain means one millionth of the original length, which on the scale of most structures, is a practically impossible distance for a human to perceive. But in addition to tiny distances, they also are excellent in measuring changes that happen over a large period of time. A perfect example is a crack in a concrete structure. You can look at grass, but you probably can’t perceive it growing, and you can watch paint, but you won’t perceive it drying. And, you can watch a crack in a concrete slab, like this one in my garage, but you’ll probably never see it grow or shrink over time. So how do you know if it’s changing? You could use a crack meter like this one, and take readings manually over the course of a month or year or decade. But in many cases, that’s not a good use of any person’s time, especially when the crack is somewhere difficult or dangerous to access. So, just like strain gauges measure distance, you can also get crack meters that measure distance electronically. I put this one across the crack in my garage slab and recorded the changes over the course of a few months.

And, I know why this crack exists. It’s because the soil under the slab is expansive clay that shrinks and swells according to its moisture content. I thought it would be fun to use some soil moisture sensors to see if I could correlate the two, but my sensors weren’t quite sensitive enough. However, just looking at the rainfall in my city, you can get a decent idea about what might be driving changes in the width of this crack, which grew by about half a millimeter over the course of this demonstration. Cracking concrete isn’t always something to be concerned about, but if cracks increase in size over time, it can be a real issue. So, using sensors to track the movement of cracks over long durations can help engineers assess whether to take remedial measures.

And, there are a lot of parameters in engineering that change slowly over time. Dams are among the most dangerous civil structures because of what can happen when one fails. Because of that, they’re often equipped with all kinds of instruments as a way to monitor performance and make sure they are stable over the long term. One parameter I’ve talked about before is subsurface water pressure. When water seeps into the soil and rock below a dam, it can cause erosion that leads to sinkholes and voids, and it also causes uplift pressure that adds a destabilizing force to a dam. Instruments used to measure groundwater pressure are called piezometers. They often resemble a water well with a long casing and a screen at the bottom, but instead of taking water out, we just measure the depth to the water level. That’s made a lot easier with electronic sensors, like this one, but I don’t have a piezometer in my backyard. So, to show you how this works, I’m just hooking my pressure transducer to the tap so we can see how the city’s water pressure changes over time. I hooked this up to a laptop and let it run for about a day and a half, and here are the results.

The graph is a little messy because of the water use in my house throwing off the readings every so often, but you can see a clear trend. The pressure is lowest when water demands are high, especially during the evenings when people are watering lawns, cooking, and showering. In the middle of the night, the pumps fill up the water towers, increasing the local pressure in the pipes. This information isn’t that useful, except that it gives you a new perspective of thinking about real-world measurements. Recently I had a plumber at my house who took a pressure reading at the tap, which seemed like a totally normal thing at the time. But now, seeing that the pressure changes by around half a bar (or nearly 10 psi) over the course of a day, it seems kind of silly to just take a single measurement. And that’s the value of sensors, giving engineers more information to make important decisions and keep people safe after a structure is built.

By the way, the engineering of these instruments is pretty interesting on its own. Most of the sensors I’ve used in the demos were sent to us by our friends at Geokon, not as a sponsorship but just because they enjoy the channel and wanted to help out. These devices rely on a wire inside the case whose tension is related to the force or strain on the sensor. The readout device sends an electrical pulse that plucks the wire and then listens to the frequency that comes back. You can see the pluck and the return signal on my oscilloscope here. Just like plucking a guitar string, the wire inside the instrument will vibrate at a different frequency depending on the tension, and you can even hear the sound of the vibration if you get close enough. Of course civil engineers use lots of different kinds of sensors, but vibrating wire instruments are particularly useful in long-term applications because they are incredibly reliable and they don’t drift much over time. They’re also less vulnerable to interference and issues with long cables, since they work in the frequency domain. In fact, there are vibrating wire instruments that have been installed and functioning for decades with no issues or drift.

And the demos I’ve shown in this video just scratch the surface. We’ve come up with creative ways to measure all kinds of things in civil engineering that don’t necessarily lend themselves to garage experiments, but are still critical in performance monitoring of structures. Borehole extensometers are used to measure settlement and heave at excavations, dams, and tunnels. Load cells measure the force in anchors to make sure they don’t lose tension over time. Inclinometers detect subtle shifts in embankments or slopes by measuring the angle of tilt in a borehole along its length. Engineers keep an eye on vibrations, temperature, pressure, tilt, flow rate, and more to make sure that structures are behaving like they were designed and to keep people safe from disaster.

East Palestine Train Derailment Explained

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

On the evening of Friday, February 3, 2023, 38 of 149 cars of a Norfolk Southern Railway freight train derailed in East Palestine, Ohio. Five of the derailed cars were carrying vinyl chloride, a hazardous material that built up pressure in the resulting fires, eventually leading Norfolk Southern to vent and burn it in a bid to prevent an explosion. The ensuing fireball and cloud brought the normally unseen process of hazardous cargo transportation into a single chilling view, and the event became a lightning rod of controversy over rail industry regulations, federal involvement in chemical spills, and much more. I don’t know about you, but in the flurry of political headlines and finger pointing, I kind of lost the story of what actually happened. Freight trains, like the one that derailed in East Palestine, are fascinating feats of engineering, and the National Transportation Safety Board (or NTSB) and others have released preliminary reports that contain some really interesting details. I’m not the train kind of engineer, but I think I can help give some context and clarity to the story, now that some of the dust has settled. I’m Grady, and this is Practical Engineering. In today’s episode, we’re talking about the East Palestine Train Derailment.

Modern freight trains are integral to daily life for pretty much everybody. Look around you, and chances are nearly every human-made object you see has, either as bulk raw materials or even as finished goods, spent time on the high iron. One of the reasons trains are so integral to our lives is because there’s nothing else that comes even close to their efficiency in moving cargo over land at such a scale. Steel wheels on steel rails waste little energy to friction (especially compared to rubber tires on asphalt). Locomotives may look huge, but their engines are almost trivial compared to the enormous weight they move. If a car were so efficient, its engine could practically fit in your pocket.  And yet, the trains those locomotives pull are not so much a just a vehicle as they are a moving location, larger and heavier than most buildings.

With this scale in mind, you can see why the crew in a locomotive can’t monitor the condition of all the cars behind them without some help. A rear-view mirror doesn’t do you much good when part of your vehicle is a half hour’s walk behind you. There was a time not too long ago when every freight train had a caboose. Part of their purpose was to have a crew at the end of the train who could help keep a lookout for problems with the equipment. Now modern railways have replaced that crew with wayside defect detectors. These are computerized systems that can monitor passing trains and transmit an automated message to the crew over the radio letting them know the condition of their train in real time. Defect detectors look for lots of issues that can lead to derailment or damage, including dragging equipment, over height or over width cars (a hazard if the train will be passing through tunnels or under bridges), and, important in this case, overheating axles and bearings. Depending on the railway operator and line, these detectors are often spaced every 10 or 20 miles (or 15 to 30 kilometers).

The freight train that derailed in East Palestine, designated 32N, passed several defect detectors along its way, and NTSB collected the data from each one. The suspected wheel bearing responsible for the crash was located on the 23rd car of the train. At mile post 79.9, it registered a temperature of 38 degrees Fahrenheit above the ambient temperature. Ten miles later, the bearing’s recorded temperature was 103 degrees above ambient. That might seem kind of high, but it is still well below the threshold set by Norfolk Southern that would trigger the train to stop and inspect the bearing. Twenty miles later, the train passed another defect detector that recorded the bearing’s temperature at 253 degrees above ambient (greater than the 200-degree threshold), triggering an alarm instructing the crew to stop the train. But, it was too late.

Freight trains are equipped with a fail-safe braking system powered by compressed air. There are two main connections between cars on a train: one is the coupler that mechanically joins each car, and the other is the air line that transmits braking control pressure. As long as this line is pressurized, the brakes are released, and the cars are free to move. But if one of these air lines is severed, like it would be during a derailment, the loss of pressure triggers the brakes to engage on every single car of the train. That’s what happened shortly after that defect detector recorded the over-temperature bearing. When the defect detector notified the crew of an issue, they immediately applied the brakes to slow the train. But before they could reach a controlled stop, the train’s emergency braking system activated.  A security camera nearby caught this footage showing significant sparks from what is presumably the failing car moments before the derailment. Understanding the severity of the situation, the crew immediately notified their dispatcher of the possible derailment. They applied handbrakes to the two railcars at the head of the train, uncoupled, and moved the two locomotives at the head end (and themselves) about a mile down the line away from the fire and damage, not knowing the events that would quickly follow.

A train’s “consist” defines the collection of cars that make it up. 32N’s consist included 2 locomotives at the head, a locomotive near the center of the train called distributed power, and 149 railcars. 38 of those 151 cars had come off the tracks, forming a burning pile of steel and cargo. Of those 38 cars that derailed, 11 were carrying hazardous materials including isobutylene, benzene, and vinyl chloride. Local fire crews and emergency responders worked to put out the fires and address the immediate threats resulting from the derailed cars. But despite the firefighting efforts, five of the derailed cars transporting vinyl chloride continued to worry authorities due to rising temperatures. Norfolk Southern suspected that the chemical was undergoing a reaction that would continue to increase in temperature and pressure within the tanks, eventually leading to an uncontrolled explosion and making an already bad situation much worse.

The cars carrying vinyl chloride were DOT-105 tank cars. These are not just steel cylinders on wheels. The US Department of Transportation actually has very specific requirements for tank cars that carry hazardous materials. DOT105 cars have puncture-resistant systems at either end to keep adjacent cars from punching a hole through the tank. They have a thermal protection system with insulation and an outer steel jacket to protect against fires. They are tested to pressures much higher than they would normally see, and they include pressure relief devices, or PRDs, that automatically open to keep the tank from reaching its bursting pressure. The PRDs on some of the vinyl chloride cars did operate to limit the pressure inside the tanks, but the temperature continued to increase.

As fires continued to burn, state and federal officials noted the temperature in one of the vinyl chloride cars was reaching a critical level. Rather than trust the PRDs to keep the tanks safe from bursting, they decided to perform a controlled release of the chemical to prevent an explosion. While they were still making the decision, the Ohio National Guard and the Federal Emergency Management Agency were running atmospheric models to estimate the extent of the resulting plume. Local emergency managers used these models to evacuate the area most likely to be affected by the release. On February 6, crews dug a large trench in the ground, vented the five vinyl chloride tanks into the trench and set the chemical on fire to burn it off. Despite being done on purpose to reduce the danger of the situation, the resulting fireball and pillar of smoke have become symbolic of the disaster itself.

You might be wondering, like I did, why the controlled burn was necessary if the tank cars were fitted with PRDs. While the NTSB’s full report hasn’t been released yet, they have released some details about their inspections of the vinyl chloride cars. Three of the cars were manufactured in the 1990s with aluminum hatches that cover the valves (as opposed to the more updated standard steel hatches). During the initial fires and “energetic pressure reliefs”, it seems that the aluminum may have melted and obstructed the relief valves, impacting their ability to reduce the building pressure.

You might also be wondering why a train passing through a populated area would be carrying so much vinyl chloride in the first place. Vinyl chloride might sound familiar to some of you as it is the ‘VC’ in PVC. This channel makes a lot of use of PVC demonstrations. It’s a material used in a lot of applications, so we produce it in vast quantities, and railways are usually how we move vast quantities of bulk materials and chemicals. But, vinyl chloride is a toxic, volatile, and flammable liquid, not something you want a big pool of near your city, so officials decided to burn it off. Flaring or burning chemicals is a pretty common practice for dealing with dangerous gases or liquids that can’t easily be stored. It’s essentially a lesser evil, a way to quickly convert a hazardous material to something less hazardous or at the very least, easier to dilute. 

While the byproducts of burning vinyl chloride are far from ideal, combusting it into the atmosphere was intended to be a way to quickly address the concern of it harming people on the ground or polluting a larger area. In fact, the US Environmental Protection Agency flew a specially-equipped airplane after the burnoff to measure chemical constituents of the resulting plume. They found low detections of any chemicals of concern and concluded in their report that the controlled burn of the railcars was a success.

But “success” is a strong word for an event like this, and I might have chosen a different word. While there were no immediate fatalities resulting from the crash, the impacts are far-reaching. Chemical pollutants were not only released into the air, but also washed into local waterways during the firefighting efforts. Hazardous substances reached all the way to the Ohio River, and the Ohio Department of Natural Resources estimated that roughly 40,000 small fish and other aquatic life were killed in the local creek that flows away from East Palestine. Between the contamination of water and soil, it’s impossible to say what the long term impact on the local ecology will be.

As for the residents, both the state and federal EPAs have been heavily involved in all aspects of the cleanup, monitoring air quality and water samples from wells and the city’s fresh water supply. So far, they haven’t detected any air quality levels of health concern after the derailment. As for the area’s groundwater, out of 126 wells tested, none have shown evidence of significant contamination. But as you’ve seen in some of my previous videos, it can take a while for contamination to move through groundwater.

The EPA has ordered Norfolk Southern to conduct all cleanup actions associated with the East Palestine train derailment. The company itself has pledged to “meet or exceed” regulatory requirements in regards to the cleanup. Cleaning up after such a disaster is no easy feat, from air, water, and soil testing, to disposal of huge volumes of contaminated water and soil, the whole thing is a mess, literally. The cleanup is still underway as I’m releasing this video, but so far they’ve removed over 5,000 tons of contaminated soil and collected about 7 million gallons or 26 million liters of contaminated water from rain falling on the site and washing off trucks working on the cleanup. The response has been robust, but we know how these cleanups can go. The EPA’s list of almost 1,800 hazardous waste sites of highest priority only has 450 examples of sites cleaned up enough to be taken off of the list!

The whole situation has also sparked policy discussions among several agencies. The NTSB is opening a special investigation into the safety culture and practices of Norfolk Southern. From congressional testimonials, to public statements from the Department of Transportation, to political posturing from a huge variety of public officials, one thing seems clear to me: this disaster will have an impact on the way railroading is conducted in America for years to come.

The residents of East Palestine have a long road ahead of them. While all the preliminary testing so far paints a relatively safe and healthy picture of the town after the event, many have reported symptoms and effects. Even if there really are no residual compounds present at dangerous levels, the anxiety and unease of living near a high profile chemical spill is hard to escape. The economic impact of just the perception of contamination is also very real, and things like home values and local agricultural businesses have already taken a direct hit. I live really close to a freight line myself, something that is a unique joy for my two-year-old. But now, when I see those tanker cars roll by, I can’t stop myself from just wondering what’s inside them and what might happen if they came off the rail in my neighborhood.

But I also recognize that much of the lifestyle I enjoy depends on those trains rolling by my house, and despite the tragedy of events like East Palestine, the DOT recognizes rail transportation to be the safest overland method of moving hazardous materials. Even with the bulk of hazardous materials being transported over rails, highway hazmat accidents result in more than 8 times as many fatalities! So, freight rail isn’t going away anytime soon. It’s the only feasible way to move the mountains of materials required for all of the industries in the US, and really, the world. And the fact that we rarely have to consider the incredible engineering details of tanker cars, defect detectors, and hazardous material cleanup operations is a testament to the hard work that goes into regulating and operating these lines.

But freight rail in the US is unlike any other industry. There are only seven companies that operate Class I railroads that make up the vast majority of rail transportation in the country. The US rail market essentially consists of two duopolies: CSX and Norfolk Southern in the east and Union Pacific and BNSF in the west. That gives these companies enormous political power, as we’ve seen in recent news. So, we have to ask ourselves, are accidents like East Palestine, however rare they may be, just a part of doing business, or is there more that can be done? And I think the answer in this case is clear. I expect we’ll see some changes to safety regulations in the future to make sure something like this never happens again. And hopefully the next Practical Engineering video on railway engineering will have a more positive light.

Why Engineers Can't Control Rivers

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

This is the Old River Control Structure, a relatively innocuous complex of floodgates and levees off the Mississippi River in central Louisiana. It was built in the 1950s to solve a serious problem. Typically rivers only converge; tributaries combine and coalesce as they move downstream. But the Mississippi River is not a typical river. It actually has one place where it diverges into a second channel, a distributary, named the Atchafalaya. And in the early 1950s, more and more water from the Mississippi River was flowing not downstream to New Orleans in the main channel, but instead cutting over and into this alternate channel. 

The Army Corps of Engineers knew that if they didn’t act fast, a huge portion of America’s most significant river might change its path entirely. So they built the Old River Control Structure, which is basically a dam between the Mississippi and Atchafalaya Rivers with gates that control how much water flows into each channel on the way to the Gulf of Mexico. It was certainly an impressive feat, and now millions of people and billions of economic dollars rely on the stability created by the project, the now static nature of the Mississippi River that once meandered widely across the landscape. That’s why Dr. Jeff Masters called it America’s Achilles’ Heel in his excellent 3-part blog on the structure.

You see, the Atchafalaya River is both a shorter distance to the gulf and steeper too. That means, if the structure were to fail (and it nearly did during a flood in 1973), a major portion of the mighty Mississippi would be completely diverted, grinding freight traffic to a halt, robbing New Orleans and other populated areas of their water supply, and likely creating an economic crisis that would make the Suez Canal obstruction seem like a drop in the bucket. Mark Twain famously said that "ten thousand river commissions, with all the mines of the world at their back, cannot tame that lawless stream, cannot curb it or confine it, cannot say to it, Go here, or Go there, and make it obey;" And engineers have spent the better part of the last 140 years trying to prove him wrong.

In my previous video on rivers, we talked about the natural processes that cause them to shift and meander over time. Now I want to show you some examples of where humans try to control mother nature’s rivers and why those attempts often fail or at least cause some unanticipated consequences. We’ve teamed up with Emriver, maker of these awesome stream tables, to show you how this works in real life. And we’re here on location at their headquarters. I’m Grady, and this is Practical Engineering. On today’s episode, we’re talking about the intersection between engineering and rivers.

One of the most disruptive things that humans do to rivers is build dams across them, creating reservoirs that can be kept empty in anticipation of a flood or be used to store water for irrigation and municipal supplies. But rivers don’t just move water. They move sediment as well, and just like an impoundment across a river stores water, it also becomes a reservoir for the silt, sand, and gravel that a river carries along. That’s pretty easy to see in this flume model of a dam. Fast flowing water can carry more sediment suspended in it than slow water. The flow of water rapidly slows as it enters the pool, allowing sediment to fall out of suspension. Over time, the sediment in the reservoir builds and builds. This causes some major issues. First, the reservoir loses capacity over time as it fills up with silt and sand, making it less useful. Next, water leaving on the other side of the dam, whether through a spillway or outlet works, is mostly sediment-free, giving it more capability to cause erosion to the channel downstream. But there’s a third impact, maybe more important than the other two, that happens well away from the reservoir itself. Can you guess what it is? 

In the previous video of this series, we talked about the framework that engineers and the scientists who study rivers (called fluvial geomorphologists) use to understand the relationship between the flow of water and sediment in rivers. This diagram, called Lane’s Balance, simplifies the behavior of rivers into four parameters: sediment volume, sediment size, channel flow, and channel slope. You can see when we reduced the volume of sediment in a stream, like we would by building a dam, Lane’s Balance tips out of equilibrium into an erosive condition. In fact, according to Lane’s Balance, any time we change any of these four factors, it has a consequence on the rest of the river as the other three factors adjust to bring the stream back into equilibrium through erosion or deposition of sediments. And we humans make a lot of changes to rivers. We want them to stay in one place to allow for transportation and avoid encroaching on property; we want them to drain efficiently so that we don’t get floods; we want them to be straight so that the land on either side has a clean border; we want to cross over them with embankments, utilities, electrical lines, and bridges; we want to use them for power and for water supply. Oh and rivers and streams also serve as critical habitat for wildlife that we both depend on and want to preserve. All those goals are important and worthwhile, but, as we’ll see (with the help of this awesome demonstration that can simulate river responses), they often come at a cost. And sometimes that cost is borne by someone or someplace much further upstream or downstream than from where the changes actually take place.

One of the classic examples of this is channel straightening. In cities, we often disentangle streams to get water out faster, reduce the impacts of floods, and force the curvy lines of natural rivers to be neater so that we can make better use of valuable space. I can show it in the stream table by cutting a straight line that bypasses the river’s natural meanders.

The impact of straightening a river is a reduction in a channel’s length, necessarily creating an increase in its slope. Water flows faster in a steeper channel, making it more erosive, so the practical result of straightening a channel is that it scours and cuts down over time. It’s easy to see the results in the model. This is compounded by the fact that cities have lots of impermeable surfaces that send greater volumes of runoff into streams and rivers. That’s why you often see channels covered in concrete in urban areas - to protect against the erosion brought on by faster flows. And this works in the short term. But, making channels straight, steep, and concrete-covered ruins the stream or river as a habitat for fish, amphibians, birds, mammals and plants. It also has the potential to exacerbate flooding downstream, because instead of floodwaters being stored and released slowly from the floodplain, it all comes rushing as a torrent at once instead. And it’s not just cities. Channels are straightened in rural areas too to reduce flooding impacts to crops and make fields more contiguous and easy to farm. But over the long term, channelizing streams reduces the influx of nutrients to the soils in the floodplain by reducing the frequency of a stream coming out of its banks, slowly making the farmland less productive.

Stream restoration is big business right now as we have begun to recognize these long-term impacts that straightening and deepening natural channels has and reap the consequences of the mistakes of yesteryear. In the US alone, communities and governments spend billions of dollars per year undoing the damage that channelization projects have caused. Even the most famous of the concrete channels, the Los Angeles River, is in the process of being restored to something more like its original state. The LA River Ecosystem Restoration project plans to improve 11 miles (18 km) of the well-known concrete behemoth featured in popular films like Grease and Dark Knight Rises. The project will involve removing concrete structures to establish a soft-bottom channel, daylighting streams that currently run in underground culverts, terracing banks with native plants, and restoring the floodplain areas, giving the river space to overbank during floods. Thanks to fluvial geomorphologists, projects like this are happening all around the world. But, straightening channels isn’t the only way humans impact rivers and streams.

Another impactful place is at road crossings. Bridges are often supported on intermediate piers or columns that extend up from a foundation in the river bed. Water flows faster around the obstruction created by these piers, making them susceptible to erosion and scour. Engineers have to estimate the magnitude of this scour to make sure the piers can handle it. You don’t have to scour the internet very hard to find examples where bridges met their demise because of the erosion that they brought on themselves. In fact, the majority of bridges that fail in the United States don’t collapse from structural problems or deterioration; they fail from scour and erosion of the river below.

But, it’s not just piers that create erosion. Both bridges and embankments equipped with culverts often create a constriction in the channel as well. Bridge abutments encroach on the channel, reducing the area through which water can flow, especially during a flood, causing it to contract on the upstream side and expand on the downstream side. Changes in the velocity of water flow lead to changes in how much sediment it can carry. Often you’ll see impacts on both sides of an improperly designed bridge or culvert; Sediment accumulates on the upstream side, just like for a dam, and the area downstream is eroded and scoured. Modern roadway designs consider the impacts that bridges and culverts might have on a stream to avoid disrupting the equilibrium of the sediment balance and reduce the negative effects on habitat too. Usually that means bridges with wider spans so that the abutments don’t intrude into the channel and culverts that are larger and set further down into the stream bed.

Just like bridges or culvert road crossings, dams slow down the flow of water upstream, allowing sediment to fall out of suspension as we saw in the flume earlier in the video. The consequences include sediment accumulation in the reservoir and potential erosion in the downstream channel, but there’s one more consequence. All that silt, sand, and gravel that a dam robs from the river has a natural destination: the delta. When a river terminates in an ocean, sea, estuary, or lake, it normally deposits all that sediment. Let’s watch that process happen in the river table. River deltas are incredibly important landscape features because they enable agricultural production, provide habitat for essential species, and they feed the sand engines to create beaches that act as a defensive buffer for coastal areas. Wind and waves create nearly constant erosion along the coastlines, and if that erosion is not balanced with a steady supply of sediment, beaches scour away, landscapes are claimed by the sea, habitat is degraded, and coastal areas have less protection against storms.

And hopefully you’re seeing now why it’s so difficult, and some might even say impossible, to control rivers. Because any change you make upsets the dynamic equilibrium between water and sediment. And even if you armor the areas subject to erosion and continually dredge out the areas subject to deposition, there’s always a bigger flood around the corner ready to unravel it all over again. So many human activities disrupt the natural equilibrium of streams and rivers, causing them to either erode or aggrade, or both, and often the impacts extend far upstream or downstream. It’s not just dams, bridges, and channel realignment projects either. We build levees and revetments, dredge channels deeper, mine gravel from banks, clear cut watersheds, and more. Historically we haven’t fully grasped the impacts those activities will have on the river in 10, 50, or 100 years.

In fact, the first iteration of the stream tables we’ve been filming were built by Emriver’s late founder, Steve Gough (goff) in the 1980s. At the time, he was working with the state of Missouri trying to teach miners, loggers, and farmers about the impacts they could have on rivers by removing sediment or straightening channels. These people who had observed the behavior of rivers their entire lives were understandably reluctant to accept new ideas. But, seeing a model that could convey the complicated processes and responses of rivers was often enough to convince those landowners to be better stewards of the environment. Huge thanks to Steve’s wife, Katherine, and the whole team here at Emriver who continue his incredible legacy of using physical models to shrink down the enormous scale of river systems and the lengthy time scales over which they respond to changes down to something anyone can understand to help people around the world learn more about the confluence of engineering and natural systems. Thank you for watching, and let me know what you think!