Intriguing.....

[MilesH]

Intriguing.....

"Mr Milliken ended the visit by describing a new product that fits a niche not yet noticed by other software companies. You could consider it being the a form of... [Sorry! I can't tell you more for now.]" (sic.)

http://www.upfrontezine.com/current.htm

 

[cclark440]

I don't know about anyone else but, MAN I HATE IT WHEN THEY DO THAT!

I would just a soon not here anything about new products until I can get me hands on them.

 

[alexfranke]

Ooo, Ooo -- please let it be miniature bipedal robots that bring frosty beers and softly sing fancy songs while we're working! I often wonder why companies haven't seen the value of little helpers like this yet!

Cheers,
Alex

 

[Jimpulse]

I'd settle for the promised API updates/documentation and some improvements to the sheetmetal package.....but then..... thats so boring.

 

[indesign]

Niche not yet noticed........hmmmmm...

Is there something new left in CAD or is this just an unexploited area?

 

[alexfranke]

Yes, Jimpulse -- I'm very interested in seeing the updated API, too. I heard that it would add a lot of additional capabilities, so I'm naturally at the edge of my seat!

 

[swertel]

Well, there are lots of niches not yet noticed.

Take for example the fact that the other mid-range modelers are now realizing that small and medium businesses "require" a document management solution. Alibre has had the repository all along. Some may call that a niche.

Niche could mean anything from an interesting CAD function or feature. It may even refer to a niche industry that didn't realize they could utilize parametric CAD. Who knows? Ok, we know who knows. But for the rest of us, only time will tell.

 

[danbrinkman89]

cONCRETE

Didn't Greg ask a few weeks ago if anyone did concrete designs with Alibre? Is this "Niche" that now knows about?

 

[MilesH]

I'm not sure I'd describe concrete design as a niche, Dan

 

[swertel]

But solid modeling mixtures of materials is a niche - and concrete is a mixture of materials.

Concrete design is not knew or niche, but in an MCAD environment it is. Being able to model a "fluid" binder material around aggregate of various sizes would be amazing.

This would make composite design much easier and more realistic. Epoxy resins with fiber filler. Using MCAD in a chemistry lab for mixtures of solid particles. Using MCAD for defining crystaline structures of materials - down to the atom. This would be a huge benefit for nanotechnology design.

Oh yeah, there are several niches based on something that could be a marketing benefit right now.

 

[MilesH]

:shock: I thought we were talking about 'designing with concrete', not 'designing concrete'...

I'm not sure I follow you on the application of parametric solid-modelling to solid/binder stuff.. Crystalline, yes, I can see that.

 

[swertel]

Designing with concrete is no different than designing with aluminum. You create a protrusion and go. The only reason designing with concrete may be different is because you do not have a homogeneous-throughout material. There is a lot of cement with various sizes of aggregate. If I were to machine concrete, I have to know my speeds and feeds for cement, large stone aggregate, small stone aggregate, and where exactly those materials lie in relation to my cutting surface so I can adjust the mill or lathe as necessary.

Take it one step further, how do I easily model concrete with rebar? That is the same problem as doing a long fiber material in an epoxy matrix. If I need to easily place aggregate particles in a cement binder arbitrarily in a confined space, how is that different than placing atoms at specific location of a crystalline structure? Sure, the scale changes because of their relative sizes, but MCAD doesn't really care about size. Just zoom in and out of your screen to know what I mean. Change the units of a model to know what I mean. The same techniques will be used to model either, just the dimensions will change.

To give you a personal example of solid/binder stuff.
I have a warhead (my day job) filled with Octol. Octol is made of TNT (a wax) and HMX (a crystal). At varying concentrations of each, what dispersion of HMX crystals (solid) do I have in the TNT (binder)?
Sure, there are mathematical models out there for me to assume a average spherical size for the crystaline HMX and determine the number of crystals are within a certain volume based on a particular packing efficiency, but it would sure be nice to visualize that.

 

[MilesH]

Re:

Designing with concrete is no different than designing with aluminum. You create a protrusion and go.

............... :shock: ...............

The only reason designing with concrete may be different is because you do not have a homogeneous-throughout material. There is a lot of cement with various sizes of aggregate. If I were to machine concrete, I have to know my speeds and feeds for cement, large stone aggregate, small stone aggregate, and where exactly those materials lie in relation to my cutting surface so I can adjust the mill or lathe as necessary.

Take it one step further, how do I easily model concrete with rebar? That is the same problem as doing a long fiber material in an epoxy matrix.

But there is specialist software out there to do this... If you want to design R.C. structures, use something like this: http://www.tekla.com/user_nf/default.as ... 519&site=1

If I need to easily place aggregate particles in a cement binder arbitrarily in a confined space, how is that different than placing atoms at specific location of a crystalline structure? Sure, the scale changes because of their relative sizes, but MCAD doesn't really care about size. Just zoom in and out of your screen to know what I mean. Change the units of a model to know what I mean. The same techniques will be used to model either, just the dimensions will change.

Sure, but I still don't see why 'Parametric solid-modelling' is needed?

To give you a personal example of solid/binder stuff.
I have a warhead (my day job) filled with Octol. Octol is made of TNT (a wax) and HMX (a crystal). At varying concentrations of each, what dispersion of HMX crystals (solid) do I have in the TNT (binder)?
Sure, there are mathematical models out there for me to assume a average spherical size for the crystaline HMX and determine the number of crystals are within a certain volume based on a particular packing efficiency, but it would sure be nice to visualize that.

Sounds more like a task for Next Limit than Alibre... http://www.nextlimit.com/xflow/index.htm

 

[swertel]

Sounds more like a task for Next Limit than Alibre... http://www.nextlimit.com/xflow/index.htm

But I don't want an analysis, I can do that with plenty of CFD code. I just want a pretty picture that looks realistic, is quick, easy, and doesn't require me to determine material equations of state.

Sure, but I still don't see why 'Parametric solid-modelling' is needed?

Just thinking outside the box. Why does "parametric solid modeling" always imply prismatic shapes? Just because we're so used to sketching in 2D and extruding the third dimension doesn't mean there aren't better ways to parametrically construct the virtual representation of a part.

But there is specialist software out there to do this... If you want to design R.C. structures, use something like this: http://www.tekla.com/user_nf/default.as ... 519&site=1

Agreed. And if I was designing structures I'd be using Revit or some other structural/architectural program and not an MCAD program. Just wanted to illustrate that assembling particles in a matrix binder is no different than assembling atoms in a crystalling structure. To further illustrate, consider each part you create to be represented only by its bounding box - i.e. a particle, and you mate it into a rectangular welded frame - i.e. the crystalline structure. In simplest terms, there is no difference.

 

[MilesH]

I realised you were after a nice graphic, didn't you see the videos :wink:

"Why does "parametric solid modeling" always imply prismatic shapes?"

Well, you were talking about using MCAD, before.... if you want some sort of parametric particle modeller, that's something else...

Maybe you could get Alex to whip up a module to create variable density, structured or chaotic states of AD "Reference Points (particles)" He likes a challenge

 

[swertel]

I would consider that a worthy challenge. Does he like to get paid for his efforts or does he do it just for the thrill?

 

[MilesH]

I would feel obliged to make a substantial donation to his son's college fund to set this one in motion

 

[alexfranke]

Get paid? What a neat concept! :wink: Right now it's more for the thrill than the college fund contributions, but I'm hoping the contributions will pick up over time.

You're right, Miles -- I do indeed like a challenge. I'm not even sure I understand this one, though. I think 3d would make things easier to visualize, especially if the particles are of different relative sizes, and if you're moving things around for best fits -- but I don't think I have my head around the problem because it doesn't seem too diffrent than modeling parts for all the particles and throwing them all together in an assembly.

So what exactly are "variable density, structured or chaotic states of AD Reference Points (particles)"?

-Alex

 

[MilesH]

Re:

Hi Alex,

So what exactly are "variable density, structured or chaotic states of AD Reference Points (particles)"?
-Alex

Haven't a clue! It was something I wrote last night as I was falling asleep at the keyboard.. at the time, it seemed to be what Scott was after.

Seriously, isn't it the "throwing them all together" that's the difficult bit?

Parametrically controlled crystalline structures, I can envisage. How would you model 'random' distributions in a defined 3D space, though?

I think I'll leave this to Scott, it's his proposal, I'm only a woodworker :wink:

Did you look at the X-Flow videos?

 

[swertel]

Yeah. I looked at the videos. We use ANSYS with LS-Dyna at my day job and we have several videos just like that. Problem being, those took a long time to define and run.

For the random particle distribution with a certain packing efficiency problem, we were able to calculate that with Excel. I need to rerun it in MathCAD to get more accurate results and better curves, but the point is that the math is actually pretty simple. Displaying it is difficult.

If MCAD can do the math to project a planar curve onto a curved surface, then they can do the math to pack particles. Getting it to display correctly, now there's the challenge.