Reminds me to our blood transport system through the body where arteries get
typically smallest close to the skin. Not sure such are fractal though.
jd
From: arocket-bounce@xxxxxxxxxxxxx [mailto:arocket-bounce@xxxxxxxxxxxxx] On
Behalf Of Ian Woollard
Sent: vrijdag 18 september 2015 0:32
To: arocket@xxxxxxxxxxxxx
Subject: [AR] Re: DMLS Chambers WAS Re: Rocket Labs
Constructal law theory says that optimum cooling is probably fractal channels,
starting with large channels that divide into multiple smaller channels and so
on until you get down to the boundary layer thickness of the fluid and then
going back again the other way.
Doing that has lower pressure drop, so you can get better flow for any given
surface area; and you want high surface area, and it has *very* high surface
area. Of course you have to have the *right* scaling factor in your fractal to
give good performance.
On 16 September 2015 at 20:23, Dave McMillan <skyefire@xxxxxxxxxxxx> wrote:
On 9/16/2015 12:36 PM, Ben Brockert wrote:
I'm more curious if they're using standard channels or if they're
using a more modern design appropriate for a printed engine. I'm
always a bit frustrated when I see 3d printed engines with channels
that were designed for a milled chamber. At least spiral them enough
that differences in flow from channel to channel don't cause linear
hot spots and early failures.
That actually reminds me of something I've speculated about in the past: if
one assumes a "perfect" 3D-printing process for creating a regen engine
chamber/throat/nozzle, has anyone ever come up with an optimum design for the
cooling channels layout? While 3DP is hardly perfect, it certainly does make
previously-impossible complex interior geometries suddenly a lot less
impossible.
Hm... and it suddenly occurs to me that 3DP might make it possible to
iterate complex *injector* geometries much more rapidly than was previously
possible. Dunno if injectors might benefit from that as much as regen
channels, but anything that lowers the cost of trying different designs can't
hurt.
Going further afield... watching the various tech journals, it seems we
might not be so far from being able to print sensors, and their "wiring,"
directly into the structure of of the engine. I have to imagine that having a
sensor "carpet" every few tens of millimeters throughout the engine shell would
be a boon to engine designers.
--
-Ian Woollard
Sent from my Turing machine
