Many of the 3D printed houses and structures we’ve seen use concrete and are — frankly — a little underwhelming. Making big squares out of concrete isn’t that hard and while we are sure there is some benefit, it isn’t overwhelming. [Andy Coward] apparently felt the same way and set out to find ways that 3D printing could offer unique benefits in building structures. The result: a beam that would be difficult to create with conventional techniques but is easy to make with a printer. The advantage is that it uses 78% less concrete than a conventional beam with the same properties.
The key is that in a normal beam, not much of the concrete is bearing a significant load. It is simply there because you need some concrete on one side of the beam and then some more on the other side. In the center, surprisingly little of the concrete actually supports anything. The new beam takes advantage of this along with a steel reinforcement at a strategic point. Still, it uses 70% less steel than a typical reinforced beam.
Reducing material has many benefits. Lower transportation costs and less carbon production, are among them. The beams also have built-in voids that can allow for services like electrical and water. You can see a good diagram of how it all goes together on Minimass’ website. Removing unneeded material is hardly a new technique, of course. We see it often in metal 3D printing, too.
“Removing unneeded material is hardly a new technique, of course. ”
The evolution of VCRs to near 100% plastic, and less materials. Or computers, and look at how long they last.
You are talking about two entirely different materials, markets, and use cases. These things are not equal.
How does it perform when the voids fill with water and it freezes?
Probably about the same as when you saturate wood and freeze it. I don’t think its intended for an environment like that.
The application is horizontal spanning beams, ie not at ground level. So if they’re filling with water and freezing, you’ve got bigger problems!
But also, no enclosed voids, so the ice would just grow out the holes.
Water leaks in buildings happen.
But the more important point is that this beam has no lateral reinforcement whatsoever. The conventional beam being compared has rebar on the sides as well, so if the building should distort from its original shape by the ground rising or falling, or something on top of the beam causing a twisting load, or you have things like earthquakes, the structure may just come down like a house of cards. If the beam should ever experience a compression load end-to-end, it has no resistance against buckling failure because all the steel is in the hanging span and nothing to keep the sides from crumbling.
That’s why this is not a general solution, but a case where the optimizations meet a very specific use case.
Also, because of its weird shape, the beam probably has to be transported upside down or on its side to the actual site and then be flipped over, which introduces the risk of breaking it. The conventional beam is also designed to survive handling it during transportation and construction.
Any fool can build a bridge. It takes an engineer to barely build a bridge.
Eng: So you want a bridge across the Tacoma Narrows huh? What kind of winds do you get round here?
Customer: Pffft, I dunno, 35mph maybe.
How do you rate a bridge?
1. Build bridge
2. Test to failure
3. Derate one ton and rebuild
I think I got that off my Statics prof in college :D
Least material required sounds like zero % extra for redundancy and unexpected load increases. The I-35W bridge was perfectly adequate when it was built, but it got overloaded with extra paving and increased traffic.
The city engineers had examined it and knew it had cracks, but rather than spend the $1M to $1.5M to repair and reinforce it they decided to just monitor it since a replacement was going to be built in a few years.
Then the decision was made to strip off old paving and repave it. Great idea to reduce the load on the old bridge IF they had repaired and reinforced the structure first. The heavy equipment and pounding of breaking up the paving was its undoing. Saving that $1.5M cost many millions more.
Now try that with a bridge built with “optimized” 3D printed concrete beams. How much damage will it sustain in 43 years? How much additional load from traffic? What if short sighted civil engineers overload it by layering new paving atop the original rather than stripping it off to put down the new?
It’s designing construction projects like how Airbus initially controlled the throttle on their fly by wire airliners. Only allow just enough thrust to take off, to reduce noise and save fuel. That proved disastrous when met with extreme adverse weather like microburst downdrafts. Pilots long knew the safest way to take off was to use maximum throttle so they’d already be at full power should any problems arise.
Not having “extra capacity” or “overhead” in building strength is just begging for a disaster to happen.
Engineers always design with some kind of safety factor. A real life beam of this type will be no different in that regard. You always have to know your materials and what they are capable of. That’s one of the things that separate engineers from casual designers.
https://www.youtube.com/watch?v=2hNe1DwDZNk
Counter example. Definitely an engineered structure… and mere 30mph winds took i out.
Modern engineers can’t even build proper ornamental arches without them falling…. and this totally would have had to have an engineer stamp it.
Not always.
100% optimization -> n=1
If you optimize the entire structure for a 2x safety factor (that is, any feature fails at 2x the anticipated load) you still have the same resiliency as a structure where the weakest point has a 2x safety factor but everything else has a 100x safety factor.
The question you raise is what safety factor is appropriate, not whether optimization to that requirement adds new risks.
For anything that may be occupied by people, safety factors begin from around 5-10x because the real world tends to not behave as designed and you sometimes get things like air bubbles in the concrete that throw you off the calculations. There’s too many cases where the engineers said “this should be enough” and then something happened in manufacturing and it wasn’t, and a building fell on top of someone, and then the legislators went “you have to use a factor of 30x instead of 3x next time”.
The case of having 100x safety factors somewhere is rather that it takes more money to optimize the material away than to just pour it full of concrete and call it a day. This is a place for the old adage, “Don’t fix what ain’t broken.” – because what does superficially save you something like concrete and CO2 output will then turn to cost more of the same because you need greater economic activity (which involves more infrastructure being built) to generate the GDP to pay the overall cost.
Least material required can also mean you double the number of beams for double the load, and still come out spending less. Pretty good deal.
Anything is a ‘pretty good deal’ if you just pull assumptions from a dark smelly place.
Hint: ‘crete is cheap.
The steel in the minimass beam is exposed, and much more susceptible to tampering and damage than an old-style beams. Cutting the steel would guarantee instant and catastrophic beam failure, so there are some places and exposures where you cannot use this beam.
The steel in suspension bridges is exposed, too.So it works in general in real life.
Steel in suspension bridges requires constant painting…. but since its exposed at least you can see and monitor it. The downside with unexposed still is it can rust away out of sight.
Well, unlike the curmudgeons replying before me, I think this is a great idea to explore. Keep up the good work.
I’m curious, why is this hard to build without 3d printing?
It’s not hard. You can make a mold and you’re done. But it’s much harder to find investors for a classical concrete mold maker idea than a hype “3d printing concrete” talk, IMHO.
I’m guessing that it’s expensive to make one-off molds? Not sure why though.
I’m more surprised that printing concrete doesn’t mess up the properties due to small voids, lines on the edge, etc.
Also pumped ‘crete is generally weaker. You want the stuff that barely flows. See also ‘slump test’.
In most applications it’s fine, you just do the math and set the forms appropriately.
I get that you can make things with 3D printing that can’t easily be made, but in this case, to my untrained eye, assuming all the beams are the same size, it looks like the sort of thing that could more easily be made with a reusable 1 or two-piece mould.
(Though if that was how it had been done, we wouldn’t be reading about it here, because that’s not what Hackaday is about.)
I’d hope it still gets a mention here, as its still an interesting engineering method we can perhaps hack into our projects.
Not that you are wrong having the buzz word in just makes it so much more likely to propagate on the internet..
“I’d hope it still gets a mention here,”
The link showed a similar bridge (in France?) that used steel struts instead of concrete.
Yeah, that seems like something worth covering in HaD.
It sounds like there are internal voids for wiring or plumbing that would be difficult or impossible to cast.
You have to have steel in such a beam anyway… no reason that couldn’t provide said internal voids.
Plenty of ways to do it. If you can’t think of at least two, you’re being deliberately dense.
I think it’s the assumption you stated that makes this interesting. A well-set-up shop could produce a dozen beams, all of different sizes, with one of these machines in one day.
I think a very large draw would also be the decreased amount of specialized labor. You don’t have to have form makers and steel workers used to working off of engineering drawings, you instead have a simple CNC rebar bender for the end baskets and a 3D printer, a good maintenance crew and some lightly trained machine operators. I know it may not seem like a big deal, but that’s what floats the entire aerospace and automotive automation markets.
This, of course, won’t be the best solution for every project, but I can see it being correct for a significant subset.
or… the could produce hundreds with molds and the same manpower…
Sure, and then store hundreds of molds, repair them from storage issues (inevitably happens with tooling), retain the skilled labor required to manufacture replacements or special orders, etc. etc.
Again, I’m not saying that this is the solution to everything. There will inevitably be projects where they’ll need 200+ of one kind of precast beam, or a standard beam used in countless examples of a specific kind of construction, and molds will be the better form of manufacture. Molded manufacture of precast beams is probably the better method for the vast majority of projects that use them. That doesn’t mean that this method won’t be great for a subset of projects that require precast beams, and with how much is constructed every year even 0.1% of the precast beam market would be more than enough to support regional specialist shops.
There’s room for more than one method of manufacture with almost anything, which is why 3D printing, injection molding, pressing, vacuforming, etc, are used in plastic part manufacture. It’s why milling, 3D printing, turning, forging, welding, etc, are used in billet part manufacture. It’s why pressing, bending, creasing, punching, riveting, welding, etc, are used in sheet part manufacture.
I think this could work out really well if the steel sections can be prefabricated and placed using existing machinery. Alternatively, if it’s easy to (partially) fabricate the beam near the site and move it into place using existing machinery then that could work too.
To get something implemented, your first enemy must be cost. The fact that it saves on material is great but if it ultimately costs more (e.g. requiring highly specialized labor and machinery) then it’s a no-go.
What looks like huge chunks of steel on the ends in that model is really just cast concrete, there’s more detail at the link. There’s a rebar basket and then the cable tensioning mechanism inside of this chunk of concrete. I imagine that these end-pieces would be pre-cast in a series of family sizes and matched to the size of printed beam in the middle. Post-tensioned cables like the one in this are already extremely common in precast beams, so the cable and fittings would likely be supplied by existing manufacturers. Similarly, the shipment and installation of prefabricated beams is a solved problem, with the only difference would be how the beam is made before it’s shipped.
Don’t forget that there’s plenty of specialized labor and capital requirements in precast beam manufacture as it is. Automation doesn’t always require more skilled labor, especially if there’s enough demand for the machine to be run on a shift basis.
The same arguments have been made in every heavily automated manufacturing segment. It’ll fall on the MBAs to decide which side wins out with this one, but i heavily suspect that there will be a (likely relatively small, but by the nature of the industry large in absolute terms) portion of the prefab beam market that will be better served by more tailored fab methods such as this.
Building something just as perishable with less material is green, but not necessarily sustainable. This “useless, move backwards” solution to climate problems seems short sighted sometimes.