3D Printed Aerospike Was Designed By AI

A 3D printed copper aerospike engine cutaway showing the intricate, organic-looking channels inside. It is vaguely reminiscent of a human torso and lungs.

We’re still in the early days of generatively-designed objects, but when combined with the capabilities of 3D printing, we’re already seeing some interesting results. One example is this new copper aerospike engine. [via Fabbaloo]

A collaboration between startups Hyperganic (generative AI CAD) and AMCM (additive manufacturing), this 800 mm long aerospike engine may be the most complicated 3D print yet. It continues the exciting work being done with 3D printing for aerospace applications. The complicated geometries of rocket nozzles of any type let additive manufacturing really shine, so the combination of generative algorithms and 3D printed nozzles could result in some big leaps in coming years.

Aerospikes are interesting as their geometry isn’t pressure dependent like more typical bell-shaped rocket nozzles meaning you only need one engine for your entire flight profile instead of the traditional switching mid-flight. A linear aerospike engine was one of the main selling points for the cancelled VentureStar Space Shuttle replacement.

This isn’t the only generative design headed to space, and we’ve covered a few projects if you’re interested in building your own 3D printed rocket nozzles or aerospike engines. Just make sure you get clearance from your local aviation regulator before your project goes to space!

36 thoughts on “3D Printed Aerospike Was Designed By AI

  1. The article was from 2022. I would like to know if it worked or not? Did they fire it?? What happened!! It seems to have vanished after the press release.

    1. It was an artpiece rather than a piece of functional hardware. Note the total lack of any turbomachinery, lack of a combustion chamber, and absurdly excessive size for what would essentially just be the bell and throat portion of an engine. Worse, the part of the truncated plug nozzle that gets the hottest, the very tip of the central core, has no cooling channels at all, and just an empty chamber instead.

      1. So… it’s not really an engine at all, just an engine-shaped art piece? What a convenient detail for Hyperganic to have omitted. Kinda emblematic of AI as a whole, it can make some cool looking stuff but it’s useless when the rubber hits the road.

      2. Isn’t the inside of that nub the preheater for the fuel and so the cooling for the tip?
        The combustion camber being rhe anular ring shapoed chamber around it?

        I though the fuel came down the center though that massive heat exchanger?

    2. I was thinking the same—any plans for testing, or is it just a publicity stunt for a CAD shop and a 3d printing shop? Obviously it was the latter.
      Ironic that they mention the VentureStar concept, which was also entirely fake and verged on the definition of outright fraud. I’m actually surprised people didn’t get arrested over that obvious budget siphon which never had any realistic plans for production and they surely knew it. My dad still has a poster from that project on his wall.

    1. “Algorithmic Variations – these Aerospike variations were automatically generated over the course of a weekend. Each of these geometries is roughly 30 cm tall and took about 10 minutes of program runtime. All of them were generated using the same construction logic. Yet, by applying a different set of input parameters every time, the resulting geometries look notably different.”

      Essentially surfaces and infill that it supposed to make sense in terms of associated physics. I’m not sure I’d like to want to wait a weekend for a slicer to come up with optimized structural mechanics, but it is an exciting way moving forward (and past gradient infill, please).

    2. Also, it’s pretty clear the vast majority of comments making criticisms or outright blasting the concept haven’t read any of this page (which is actually an interesting read). It also explains the purpose behind the whole thing.

      I suggest that before making your expert commentary on “AI” (yes, I also hate AI branded things that are actually RPA, ML, procedural, algorithmic or automation) generative design, propulsion or nozzle design that you read the above link. It answers and questions the vast majority of the things people seem to be commenting on.

    3. Well if they are based in USA or allied countries, a ground-breaking high efficiency aerospike engine design would be extremely ITAR-controlled, and just to look at the cutaway gives huge hints to the design, so scrubbed from website makes perfect sense… if it really works, it’d be really nice in a hypersonic missile, no?

    1. They say generative design was used in this article, I can’t see any reason it couldn’t be combined with some form of machine learning to speed up the iterative process.

      1. That’s because every technique that was sold as AI wasn’t intelligent to begin with – just another mechanical turk or the same old dumb thing ™ done faster and bigger.

      2. “Intelligence” is the ability to solve previously unencountered problems. By that definition, generative design is an “intelligence” limited to a very narrow field, that falls closer to blind optimization than to actual problem solving. Compare that to a tokenizing LLM transformer, which can split any kind of knowledge into tokens, then transform the context it’s given into any number of solutions, including driving a generative design process.

  2. It’s not quite a real part meant to be fired, just a showpiece for Hyperganic/AMCM, and the designer for the part has left to make her own company (LEAP 71), which might be why some info has been scrubbed.

  3. It’s a pity that aerospikes like this will never deliver on their promises. Looking at the image of this beautiful art piece, it’s maximum expansion ratio is probably less that most sea level engines, and is undoubtedly way less than the SSME expansion ratio.

      1. Eh, once you start insisting on expertise and planning for human reproduction your civilization is basically a walking corpse. Spengler illustrated this rather well. Which is exactly what happened to us, so we dodged the idiocracy future but ended up with something that’s probably a lot sadder

          1. A civilization of narcissists that will wake up in their 40s to realize they wasted their entire lives chasing debt based materialism and never know the love of raising a child, that the population is in decline and it will take society with it, and some might even realize that they were motivated to take this course because of a malthusian fraud that could’ve been disproven with preliminary education in basic economics.

          2. That’s a funny way of describing the prevailing economic theory, Dude. I wouldn’t say it takes center stage in my identity, but yes, there is no such thing as peak oil, and if there was then it would be driven by a revolution in chemical/material science that provides a more cost effective alternative for production than oil. I don’t want to do you the disservice of straw-manning your reasoning for believing in peak oil, I’ll outline a couple points and respond to whatever you say I am missing.

            -There is no hard limit to oil reserves (at least not until you start talking about building a dyson sphere), stated oil reserves are only the oil that has been found from exploratory surveys. Year after year oil companies launch these expeditions to find more oil reserves that they can tap when the economics line up, they have generally targeted around having 10 years of reserves surveyed. There is little economic incentive to go beyond that and discover more oil that isn’t being used. This is why its essentially always the case that “theres only 10 years of oil reserves left”.

            -Oil production has factors that affect the price of production; as easier to reach deposits dry up it requires drillers to target more expensive sources, this causes an upward pressure in prices; as technology gets better it gets easier to survey and extract oil, this causes a downward pressure in prices. The history of price coordinated markets shown that the net effect of these pressures, is a decrease in prices. If you look at the time-price of anything from steel to smartphones, the only items that increase in
            “real price” over time are things the government subsidizes/heavily regulates. Whereas the general increase in “nominal prices” of all goods is driven by an inflationary monetary system, the US dollar.

  4. So um, maybe dont infer that the company is saying this is an AI project when they aren’t. Generative design /= “AI”
    They literally say it’s “Algorithmic Engineering” and an experiment in fast CAD iterations using generative/algorithms for an engineering design that makes iterating quickly almost impossible.

    A whole lot of assumptions being made because of a misleading article title.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.