I need help by building a heat exchanger with special honeycomb-geometry

Replies to This Discussion

Permalink Reply by Joseph Oster on March 26, 2017 at 2:12pm

It looks like you didn't understand what Hyungsoo Kim did for you in the other thread.  You reintroduced 'SDiff' for no reason and worse, didn't notice that doing so created 56 copies of your extrusion!!  Drop the 'SDiff' and extrude only the 'Boundary' surface connected to the blue preview.

I don't think you'll get your twisted shape down this path though...  Instead of extrude, you'll need to isolate the triangles for your pattern, copy the pattern multiple times alone a line, and rotate each copy incrementally for sweep or loft.

You have to walk before you can run!  What good is your thesis if someone else does all the work?

Permalink Reply by Joseph Oster on March 26, 2017 at 2:48pm

Something more in this direction:

Permalink Reply by Carina on March 27, 2017 at 8:27am

Could be that I didn't understand everything Hyungsoo Kim did but I try to understand it. GH ist really new for me so I try my best to get a geometry I have in my head.

My thesis is not about what is the right code for this part, my thesis is about to develop guidelines for additive manufacturing for heat exchanger. I discussed a few possible geometries and decided to work with GH on this honeycomb-design. I try to understand GH by reading and watching tutorials, but that is not really easy.

"Wouldn't it be MUCH EASIER and cheaper to build if you get that extra surface area by just making a straight extrusion a little bit longer?" Could be, that you're right, but this is what we want to find out in simulations. But for the simulations I need a model, but so far I have no. I'm working on it.

Until now, I worked with the CAD software Solid Works, so there is a big difference to GH. And because of this I am really thankful for help.

Permalink Reply by Joseph Oster on March 27, 2017 at 9:47am

Could be that I didn't understand everything Hyungsoo Kim did

It was the entire point of the other thread.  The surface you want to extrude in this thread is the same one he provided previously, connected to the blue preview:

Try to understand what is coming out of each GH component.  Connecting a 'Panel' to the output is a good way to see it as text, which reveals the tree structure (one or more items, how many "branches") and tells you what it is.  Preview is great too, of course, but not as obvious when there are duplicate geometries in the same location.

Permalink Reply by franco ferreiro on April 12, 2017 at 9:21am

Hi Carina, one of the great things about GH is its ability to work parametrically and in a component by component (meaning - and helping to maintain (very good thing to have: healty) structure!) as you might already noticed.

Second great thing is that is quite fast, precise and versatile (for this kind of things); also is way OPEN (meaning you can attach and or interface it with almost anything you can imagine, meaning hardware, and other sw components, etc (like a CNC machine (additive manufacturing toys..) or any sw like C# component)) making a GREAT HUGE difference with almost any other CAD (and CAM sw i must say) 

i made a simple fully functional CAM component - highly powerful ! - in a couple of days...

also tested an arduino interface (meaning control over almost any elctronic device out there)... in a matter of hours...

and saw and can easily think about lots and lots of extremely cool usages of this great tool in almost any area ...

So that's why i would suggest - and will do something about for - it (or similar tools) to be teached at first stages of education !

But power comes with responsability. and is far better exploited when your are smart ;)

I think people that uses GH will be n-times as good when they don`t forget manufacturing.

This includes teachers btw....

Interesting thing to account is that all things that GH is great at (a LOT) means reducing dramatically the time spent to model almost anything...

But usually the purpose (unless the objective is just learning or doing some kind of virtual art (both legal stuff btw...;) but guess it might not be your case now and after graduating..)) is to end up by actually building some real 3D stuff...

So what Joseph is poining is key...

If you have a good teacher.. i guess it should pay more and more attention not just at your gh skills but rather the way in which you use the power, versatility and extra time gh (and additive manufacturing tech) saves,  to think about how to design the stuff focusing on the ultimately relevant stuff...

optimisation...

So..

I would say that any heat interchanger like the one involved in your thesis, has to deal with fluids.. have to account for some sort of life span (involving cheaper an ideally no maintenance needed along its life...), and of course also critical the costs of manufacturing.

so... be the best one...

use GH smartly ! ie...

account for different profile paths for oil and water.. they're different fluids meaning they have different specific heat, viscosity, blah... and so... they might not even traverse the interchanger at same flow ratio, etc.

So... maybe you want to start by reshaping the grid... (parametrically...!) so you can arbitrarily and dynamically modify and get to see interactively in your definition the areas ratio of sections so as to finaly get to set the "ideal" (meainng optimum) relative areas (sections) ratio of oil to water paths... (or whatever other fluids could be !), and the material also...

Secondly you might also consider that triangles might not be well suited for the conduit sections because are not the best shape to carry most fluids... (hoses are of circular sections...worst case are kinda rectangular with rounded corners..;) not only because the're easy to manufacture but also because they minimise (optimize) flowing energy losses AND are less prone to (ie salt or debree deposits in the interior) ). so think about rounded shapes, of if you want some regular polygons stuff but 5 or more faces...kinda circular...got it ?

I love bees by the way..

and if you happen to need more interchange area (obviously another (and probably the #1 key one) figure you should be displaying interactively in your definition ) you can always add some more extrusion length...

third... the twisting stuff is cool... (artistically ;)) but i 100% agree with Joseph is far likely to involve higer costs for manufacturing with no clear benefit on surface maximization... and most probably some other losses in added friction to the flow of fluids (meaning higher costs for pumping, etc...)...

fourth...

consider the area, (then the volume!) of the "building material"... you should optimise that too ! so this could be another one of your interactive displays...

in this case... you not only can see optimisation by reducing the amount of materials to build your interchanger...

but you can also notice that if the "building tech" involves the well and common additive manufacturing process of extrusion deposits... that surface area, and that extrusion length, meaning volume and cost of raw material, also mean TIME to manufacture... and i guess you teacher will find good for you to consider and mention that one too...

fifth...

finally (for now hehe), and globally most important in the short term :)

 if the objective of yor teacher is for you just to learn GH and impress him and the rest of the world then, ok, do the twist the swirl and imagine all kind of sea star and or ondulated conduit sections (maybe some recursive forms (fractals...) like snowflakes... or any n-arms (mutant !) starfishes shapes) but make sure you first get to know and validate what it will be the objectives of your evaluator...

.. in the near end this is all about passing your thesis while learning GH while having fun.. isn't it ?

go for it and best of luck !

ps: for the mid and long term.. some day take a look at linear optimisaton if you already didn't.

i think GH  is a great tool to try out some linear optimisation stuff directly linking geometry related figures (areas, volumes...) along with costs figures !...

I haven't seen anything like that yet (but since i'm only a few months old in gh, i think is likely to already be something or this stuff out there.  )

If not... well you can always be the first !

(and this particular case of your thesis is a great example (few key variables) to try out "automatic optimisation")

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

that... by the way...has lots to do with spatial geometry...

 

Permalink Reply by Joseph Oster on March 26, 2017 at 3:39pm

By the way, have you considered how to actually manufacture this thing?  How much extra surface area is added with the twist and rippled interior surfaces?  Wouldn't it be MUCH EASIER and cheaper to build if you get that extra surface area by just making a straight extrusion a little bit longer?