How much airflow required to fill the wake?

[IamIan]

A smaller version of:

Jet Beetle

With , just enough thrust to overcome some % of the drag.

[Cycle]

Originally Posted by freebeard

How does it work? Constructing a model with programming statements. I can see (per RBF) everything is amplitudes and angles; but how do you construct a data structure of some arbitrary constellation of component parts? Start with some high-level description?

For instance, the main member of my frame:

module MainMember(){
// Main member: 2100mm length, 50.8mm diameter, 3mm wall, Part#:1
  color([0.9,0.9,0.9]){
    rotate([0,90,0]){
      translate([-50.8,0,0]){
        difference(){
        cylinder(h=2100,r=50.8,center=false);
        cylinder(h=2100,r=44.8,center=false);
        }
      }
    }
  }
}
MainMember();

"color" is the standard RGB notation, but rather than going from 0 to 255, it goes from 0.0 to 1.0.

"rotate", of course, rotates the described object.

"translate" moves it.

"difference" allows you to cut out parts of the described object, in this instance, it creates a tube of 50.8 mm diameter with 3 mm walls.

The next thing I've got to figure out is how to create triangles... it uses the "polyhedron" command, and that's a bit complicated. I need that to create the frame gussets.

[freebeard]

I took a look at some Youtube tutorials. I see you can do Boolean operations on primitives. I don't see any ability to touch individual edges or vertexes. For instance, how would you bevel an edge?

But if it works for you that's great. I looked at a number of programs but the only one that fits my understanding is Wings 3D.

[seifrob]

I know that the topic of this particular thread shifted a bit, but i want to post this for future reference to show others what has been done:



In early 50's there was a racing car called Tatra 607/2 monopost built to F1 specifications of that time. Its air-cooled V8 engine vas first cooled by two fans, but later they applied ejector technology, what is using stream of exhaust gases to such the air through cooling. That saved approximately 20 hp of this 200 hp engine. You can find reference here:


TATRA 607 F1,Formule 1, construct, okruhy, závody, grand prix | Constructors F1,

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

Pet motoru z ceskoslovenska (czech only, sorry).

Now why I am still sceptical:

- they used exhaust gases to replace cooling fans only. so applying to the part of aistream that already was "draggy", not adding another ducting to move more air to fill the wake (and Ledwinka knew enough about wake, as he worked with Jaray)

- although i don't doubt the exhaust gases were able to transport large volumes of air, I suppose they didn't produce much of "sucking pressure". Air cooled engines have cooling fins much less dense than common water-cooled radiator is. Therefore i worry that it would severely reduce performace in common water-cooled radiator cooling

- the idea you proposed is not new in its essence . It has been done - the T607 is living example - and it is well documented. But it was - despite its remarkable results (T 607 were unbeatable cars of its time) - abandoned. IIRC it was because it works properly only in narrow set of conditions. Have too low rpm and your engine will overheat, have revs too high and you will overcool.

- From the above stated reasons I must conlude, that it is probably idea not worth to pursue - otherwise bike disegners would already applied it.

I dont want to be a spoilfun. You asked us our opinion and each of us tried to
answer as honestly as he could. It is definitely doable, but it is too much effort with too uncertain result. But, feel free to go ahead.

BTW - Anybody knows how to solve Navier-Stokes equations? If they solved it in the fifties...

[freebeard]

Thanks for that.

I saw a short article in Popular Science/Mechanix years ago about a Porsche race car that had a similar installation. Sorry, no linky.

That didn't mention cooling problems, it said it was abandoned because it was LOUD.

Edit: My thinking was to retain the cooling fan and reuse the normally dumped hot air. I don't know if it would increase or decrease the backpressure in the exhaust. A Porsche-style fan would allow tuning because they are available with five to eleven fan blades.

A rear or mid-rear engine or radiator should have the advantage here.

[skyking]

Originally Posted by seifrob

I know that the topic of this particular thread shifted a bit, but i want to post this for future reference to show others what has been done:



In early 50's there was a racing car called Tatra 607/2 monopost built to F1 specifications of that time. Its air-cooled V8 engine vas first cooled by two fans, but later they applied ejector technology, what is using stream of exhaust gases to such the air through cooling. That saved approximately 20 hp of this 200 hp engine. You can find reference here:


TATRA 607 F1,Formule 1, construct, okruhy, závody, grand prix | Constructors F1,

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

Pet motoru z ceskoslovenska (czech only, sorry).

Now why I am still sceptical:

- they used exhaust gases to replace cooling fans only. so applying to the part of aistream that already was "draggy", not adding another ducting to move more air to fill the wake (and Ledwinka knew enough about wake, as he worked with Jaray)

- although i don't doubt the exhaust gases were able to transport large volumes of air, I suppose they didn't produce much of "sucking pressure". Air cooled engines have cooling fins much less dense than common water-cooled radiator is. Therefore i worry that it would severely reduce performace in common water-cooled radiator cooling

- the idea you proposed is not new in its essence . It has been done - the T607 is living example - and it is well documented. But it was - despite its remarkable results (T 607 were unbeatable cars of its time) - abandoned. IIRC it was because it works properly only in narrow set of conditions. Have too low rpm and your engine will overheat, have revs too high and you will overcool.

- From the above stated reasons I must conlude, that it is probably idea not worth to pursue - otherwise bike disegners would already applied it.

I dont want to be a spoilfun. You asked us our opinion and each of us tried to
answer as honestly as he could. It is definitely doable, but it is too much effort with too uncertain result. But, feel free to go ahead.

BTW - Anybody knows how to solve Navier-Stokes equations? If they solved it in the fifties...

Go back and read my links about the Cessna 310 and others. They effectively used the venturi effect and the exhaust gasses to vary the cooling flow, eliminating cowl flaps.
The big downfall of the early models was:
A) it was noisier with the exhaust over the wing, drumming on the aft fuselage. This was immediately known.
B) corrosion of the rear wing spar. That took a while to rear its ugly head, but the gasses leaked out of the muffler and got into the skin laps on the rear spar and caused a lot of damage.

[Cycle]

Originally Posted by freebeard

I took a look at some Youtube tutorials. I see you can do Boolean operations on primitives. I don't see any ability to touch individual edges or vertexes. For instance, how would you bevel an edge?

But if it works for you that's great. I looked at a number of programs but the only one that fits my understanding is Wings 3D.

You essentially use difference to subtract the edges:
I Heart Robotics: OpenSCAD Tip: Round 1 of 3 - Basic Rounding
I Heart Robotics: OpenSCAD Tip: Round 2 of 3 - Advanced Rounding

There are also libraries that let you create these sorts of complicated shapes much more easily, but the one I tried for triangles used deprecated commands, so it was outdated.

Not one to give up easily, I sat down last night at work and went through the polyhedron section of the user manual. In doing so, I learned how to create pretty much any shape I want. It's complicated, but very precise. You have to divide up any shape you want to create into a series of faces (the surfaces of the shape), then divide those faces into triangles that are delineated by the corner points. So for instance, the front face of a cube would be denoted by two triangles to make a square. Once you create the shape in the size you want, you can translate and rotate it into position.

That Wings 3D looks good... if only I had even an iota of artistic ability. My clockwork brain thinks in code, not artistically.

[freebeard]

seifrob -- TFA makes it sound like a compound system, rather than a replacement of the axial fans.

There was a unique system of air-cooling, ensured by two axial fans, driven by cone belts from the crankshaft. The fans are supported by an ejection effect of exhaust gas, because the exhaust piping leads to air channels with air heated by the engine, entrained by the motional energy of exhaust gas. This was a unique layout of combustion engine cooling system

Not saying you're wrong, though.

Cycle -- I've done 'arty' things and 3D modelling doesn't remind me of that. In Wings you are constructing a string of operations on primitives, you just choose from right-click menus instead of writing the code out directly. It's still sensitive to the order of operations. You find yourself combining, rotating and separating objects.

Artistic comes in at the point of texturing and lighting for renders.

[Cycle]

My first polyhedron!

polyhedron(points =[[0,0,0],[0,10,0],[20,0,10],[20,10,10],[20,0,20],[20,10,20],[0,0,40],[0,10,40]], faces=[[0,4,2],[0,6,4],[1,3,5],[1,5,7],[0,7,6],[0,1,7],[0,3,1],[0,2,3],[2,4,5],[2,5,3],[4,6,7],[4,7,5]]);

In OpenSCAD, if you go to the 'View' menu and select 'Thrown Together', then click the 'Preview' button on the toolbar, any badly oriented polygons in the faces will be highlighted. It helps to draw on paper the shape you want, then label each corner, starting with zero. That makes it easy to order the face polygons such that the vertex ordering is correct.

Using this technique, I can create the odd shape I need for the frame gussets... imagine looking at a triangle with one corner lopped off flat. That creates a small space between the gusset and the tube joint. That way you're not welding all the way into the corner where the joint is, which is difficult to do and doesn't do much to transmit the stress in the frame anyway.

[seifrob]

Originally Posted by freebeard

seifrob -- TFA makes it sound like a compound system, rather than a replacement of the axial fans.

Not saying you're wrong, though.

First models, in later revisions was system tuned to eliminate fans and save approx. 20 hp.(the source is thr last reference link i provided.)