Heat exchanger header with fractal geometry | Hacker News Reader

Heat exchanger header with fractal geometry

patents.google.com

Related: "3D Printed Heat Exchanger Uses Gyroids for Better Cooling"

In short: 2x smaller, 2x more efficient

https://www.youtube.com/watch?v=1qifd3yn9S0

In general, algorithmic engineering and AI designs are where 3D printing can really shine.

1/2 the volume, which is 40% smaller diameter (Assuming the length is kept, which as a drop-in replacement it would have to be?)

The inventor talks about how the internal structure is based on minimizing surface area. That would of course reduce the material but in a heat exchanger don't you want to maximize surface area, not minimize it? Especially if your heat exchanger has a higher thermal transfer rate than the fluids. It makes me wonder if there's an answer to this design that's another 20% smaller.

Yeah I did this about 3 years back using thermally conducting filament, the goal is exactly that - maximize surface area while maximizing float through the manifold. It’s actually stupidly easy to do these days with cura. Make your heat sink shape, say a cube, and slice it with zero wall thickness and gyroid infill. Then attach to a copper plate, and add a fan to the top. The problem is materials. Most people don’t have the ability to print in copper or aluminum. The thermally conducive filaments are absurdly expensive (like $120/kg) and the conductivity isn’t that great, so you can generally do better with a fin based thing in copper or aluminum. I think it would be possible to make a filament with aligned graphene in the filament that does really well, but I’m not that hard core. Another idea I have is building these out of graphene aligned in a super critical bath of silver. But that requires some serious equipment for the pressures and temperatures.

For some geometries, you can also 3D print plastic and then cast in sand.

That won’t get you the full power of 3D printing, but you can get many geometries you couldn’t otherwise.

Yeah I’ve thought of doing this. I probably will at some point. A gyroid should be castable

I believe 'minimal surface' is a technical term for these types of shapes:

https://en.wikipedia.org/wiki/Minimal_surface

>> The inventor talks about how the internal structure is based on minimizing surface area.

That gives something like maximal strength for the amount of material. I suspect that's kind of related to reducing resistance to flow for a given atea, or something like that. They can increase surface area by shinking the gyroid pattern. But yeah, I thought the same thing when he said that - dude you want the maximum surface area!

Like any field there is a lot more complexity they didn't cover in the video, but I would expect the designers did, at least in part. Particularly around pressure drops (especially in lube oil and fuel systems) and fouling (again, in a system like oil) Also I'm surprised they use S&T exchangers on a helicopter. I'd have expected plate or PCHEs, both are much smaller and more efficient. If they're using a S&T there may be a reason for it, but I'm not an aviation engineer.

I wish they'd cut it open or otherwise show the cross section(s) of the interior in detail.

It's for a military helicopter? It looks more vulnerable to shrapnel than what I imagine it's replacing. Is that so, and is that an important consideration?

This heat exchanger works by being surrounded by fuel. Thus if shrapnel is entering the vicinity of this part, the vehicle and it's occupants are probably already toast.

Caveat / plot twist: If it's a kerosene or diesel chopper (common with Jet powered helicoptes), the fuel is much less susceptible to igniting due to a spark.

Even still, how often does the outcome land on "If only the heat exchanger was more robust and impenetrable, everything would've turned out significantly better!" ?

Thanks.

Presumably if the size of the heat exchanger can be reduced, the risk of getting hit by a bullet is reduced, and also it can more easily be placed behind protection.

I feel like this is an obvious idea, but we simply haven't had the 3D printing technology to make things like this before. Casting this using traditional methods would be extremely difficult. Nobody bothered making stuff like this, not because they never thought of it, but because it was always too expensive.

It's essentially how any mammal cools itself - so yeah, pretty obvious

The header is not actually meant for cooling, just to distribute the fluid evenly across the heat exchanger. 3D printed heat exchangers use shapes like gyroids.

Seems pretty obvious: maximum surface area with minimum backpressure, the design becomes like "lungs" or "placenta" branching from "artery" down in steps to "blood vessels". It looks like some early jet turbine combustors.

"Kidneys" and reverse osmosis cartridges OTOH appear to function for optimizing for different constraints involving liquids and pressure.

Reminds me of the fractal vice grip:

https://www.youtube.com/watch?v=QBeOgGt_oWU

I'm sure there are going to be many applications for fractal geometry found as computer aided design becomes more AI supported.

It seems like it would massively complicate heat exchanger repair..

I learned of an obscure tool recently- a roller tube expander. You use this to tight-fit a tube in a boiler's tube-sheet. So you can install or replace tubes in pressure vessels such as a boilers or heat exchangers. Check it out:

https://www.youtube.com/watch?v=u7UQKNapTBE

Failure:

https://youtu.be/HlXVP2HHTfo?t=696

Repair:

https://www.youtube.com/watch?v=7znd-hpy2Vc&t=708s

I think this is why this application is a military helicopter part where weight savings are at a premium and budgets are not.

This is for aerospace heat exchangers, where weight is the primary concern. This would not be used along-side tube-base heat exchangers, but rather with 3D printed heat exchangers.

Replacing tubes in aerospace heat exchangers seems to be a thing:

https://www.meggitt.com/products-services/heat-exchangers/

"Its innovative design means that single tubes can be easily replaced during overhaul periods, minimising costs to operators.

The operational lifetime of a tubular unit is 3 million flying hours and 50 million Serck Aviation bonded interfaces in the field have never failed."

Even the patent mentions:

"In addition to aerospace applications, the disclosed headers can be used generally in other transportation industries, as well as industrial applications."

But actually I think the spreader could be contained within some kind of bell-shaped header, and not actually attached to the heat exchanger tubes, just immersed in the same fluid. So replacement might be easy.

But any kind of spreader is adding weight compared with an empty header, so not sure if the easing of the fluid flow is worth the weight gain. No doubt this depends on the application.

1. This design would have to be 3D printed because there is not a good way to manufacture it using traditional methods, so it would probably be paired with a 3D printed heat exchanger. 3D printed heat exchangers do not typically have tubes. Rather, they employ more complex geometry such as gyroids. edit: actually I'm not sure this is true. I guess you could come up with a similar design that could be IM'd.

2. Just because you could use the design in "other transportation industries" doesn't mean anyone is actually planning to do it. With the technology as expensive as it is right now, it would not make economic sense.

3. Having a spreader is not primarily about "easing the fluid flow," it's about ensuring that the entire heat exchanger gets fluid flowing through it. Typically this is worth it in applications where weight matters (because you can't afford the dead weight of part of the heat exchanger which is not actually getting much fluid flowing through it), but it is not worth it in industrial or HVAC applications (because it's cheaper to just build a bit larger of a heat exchanger). I do like the idea of a "floating spreader" - I will have to see if that has been done before.

Good example of when obvious patents harm innovation.

Can you use nature (i.e. lungs or penguin's blood vessels) as prior art?

Maybe? I think the problem is does a small entity like a person or a start up have the resources to make that fight with a big entity? Even if it’s a really obvious and winnable case?

Is that really fractal in the sense of fractional dimension? Or is it just self-similar? I watched a video on this for 3b1b and it undid 30 years of thinking I knew what "fractal" meant and now I know nothing.

If you really want something to be fractal, there needs to be a limiting process. You can't refine something infinitely in physical space, because of fundamental physical limits. So, you could say it is "only" self-similar. But this is kind of missing the point. The engineering benefits of something "being" fractal are realized in this case even though the limiting process terminates below a certain point.

No material object can be "fractal"--- it is just being used as a buzzword here.

Fractaloid would be a more accurate description.

True fractals are self-similar regardless of scale.

> True fractals are self-similar regardless of scale.

That's not true: the coastline of England is a well-known fractal and it has zero self-similarity. Self-similarity can be a property of some fractals, but it is not a requirement to be a fractal.

FYI, this is not something that people are looking to use outside of aerospace. The heat exchangers in homes, factories, power plants, etc. are designed to be energy efficient and cheap. This is designed to be extremely light weight and 3D printed. It's much, much more expensive than anything used outside of aerospace.

While this isn't relevant to my line of work, so looking at it isn't going to be problematic, I don't think that posting direct links to patents is a good idea.

Just one employee reading a patent can have implications on a business

Why patents in particular?

Intentional patent infringement (if you knew that something was already patented) has triple damages compared to accidental infringement.

I 3d printed this well before 2019... Made in plastic as a prototype heat exchanger designed in openscad.

Is there anything I can do, or must I now shelve my project because I didn't see this patent be filed?

Did you publicly publish your work? E.g. blog post, youtube video, etc. Could be as simple as a single photo or a brief written description of the design.

If so, you're good to continue your work. You could even try to help the patent examiner find your work by publishing it and using keywords related to the patent.

A simple Google Image Search for "fractal 3d print heat exchanger" yields ample prior art.

It really should be denied simply on the requirement that it not be obvious to a practitioner in the relevant field.

Fractal branching increasing surface area while reducing sub unit cross sections for increased heat exchange efficiency was obvious to practitioners in the field in the 1980s.

That's when the Mandelbrot book(s) hit the mainstream in mathematics and engineering following his coining of the word in 1975.

As an undergraduate at the time I can attest such ideas were extensively discussed, there are many people with notebooks from the time sketching out notions of practical and impractical applications and further study in multiple dimensions.

Relationships between fractal branching, leafs and fluid exchange (be the fluid heat or CO2 or reactants) have been sketched out diagrammatically and analytically ad nauseam in the time since.

IANAL but if you published anything, now would be a great time to maybe make a YouTube video showing off any posts you made, just to help build a case. Maybe you don’t need to file a lawsuit but if you use the keywords from the patent in a new video about your previous work then someone else trying to commercialize such a thing could find what you’ve done and use that to file a lawsuit.

Realistically (not legally) you can continue for personal use. Good luck commercializing though. Even if you published the work previously you'd most likely have an expensive legal battle ahead of you.

Good luck building and cleaning that. Dorks don’t make good engineers.

Didn't the Reason Hypervelocity Railgun in Snow Crash use one of these?

Yes, that’s also what I was thinking of when I read the headline.

“I’m sure they’ll listen to reason”

2011: Fractal heat transfer device https://patents.google.com/patent/US9228785B2

Should we allow patents for something AI designed?

Would like to see how it handles thermal cycles

Reply on news.ycombinator.com