ICEM SURF ii

Moving on to the next segment. I just repeated all the steps up to the creating of the patch from curves.

And after making the patch from curves, I noticed I was not able to create G2 continuity.  Highlight analysis reveals the flaws in the surface.

These are just the sort of surface flaws that I'm trying to avoid. So what next?

Here is what I'm thinking. I'm using "too advanced" features of Catia to create this surface? As an analogy, I'm using Windows 7 when what I really need is DOS?

How about I just create a point to point line?

And using the "control points" command make this a line of the order of 2.

And repeat and repeat

So now I've got four connected lines of an order of 2.

And my surface would look like this:

Then after making the vertical lines symmetrical, created a blended curve with G3 continuity.

Convert this to a NURBS line using the control points command

Repeat

Patch from curves

So far so good. Connect checker shows surface-to-surface connections as G3. Only disadvantage is not being able to use the uppermost guideline, we'll see what that means later I think.

 

On to the next segment

hmmm, nothing to show because that clearly doesn't work...still lots to learn

ICEM SURF

So continuing on an idea from last year! I'm starting again this time mainly using ICEM SURF. The idea is to use as few control points as possible to create as simple surfaces as possible.

Also picked up a neat screen capture tool: Snagit

I've just copied forward all the reference images and lines I already had

Also copied in the front wing design I've already got

Based on my reference pictures, lets say this part is perfectly straight

I'm also going to say that the side is fundamentally flat vertically, so I've created a verticle line here, and I'll build a surface based on this.

So here's the swept surface

Now I'm looking at creating the top profile and I can see that the reflection lines start to diverge from the top of the profile, so I'll make an adjustment

I'm making this adjustment by moving the highlighted point closer to the front.

Now I'll make another plane below the reflection guide line to cut the vertical plane with

Split the surfaces, the vertical surface that remains will be used as the G3 reference for the curved top of the nose.

Make that symmetrical across the zx plane

Top profile lines

Extract the geometry

Create the profiles

Patch from curves

Make symmetrical

Check on nose

So I can see this will require some adjustment

Ok, some minor adjustments on the height of the points and this looks good to me

So let's say that this part is done and move on to the next segment

It's been a long time

Time to come back to this project.

I've been thinking about this a lot lately, and it just seems to me that I'm doing something to make this too complicated. It doesn't seem like this design phase should be so difficult. I can understand that making decisions about the shape and optimizing in a windtunnel should be difficult, but just copying an existing design? Especially one from the early 90s?

Simple plan

I've got an idea. Just an idea. I'm going to step by step mold the top of the monocoque in the ICEM Aero Expert workbench making everything as simple as possible (read: the fewest control points necessary). I don't know if this leads anywhere, just an idea.

So here is the starting point

A couple of points and a line

A couple of points and a line

A mirrored line.

There are our guide lines for the coming loft. Now we need a profile. I'm going to connect the dots in the front with a 3d curve using "near points"

Note this says N5. So there are 5 control points, or an order of 5. (I think). So if I click on control points, it create a new line and shows me the 5 weighted control points (circled).

Doing something similar on the straight lines shows 2 control points.

In order to make a loft we'll need a second profile. Otherwise I think it would be a sweep. Same theory as the first profile, N5.

Loft. Make sure to select the top profile first. This may be important later.

So, fiddling around with the Loft command, there is nothing I can do to get the loft to have the same structure as the profiles, in other words a surface with a v order of 5, and u order of 2. I can get it down to a minimum of 10 bezier surfaces.

Worse yet when you try to modify by control points you'll find that each of these Bezier surfaces has a v order of 5.

So I think bottom line with the lofts is this: if you have the greatest curves in the world and plan to make no edits to the surfaces go for it. It seems like the loft is not connecting the control points of the profiles, but projecting the control points from the first profile to the second, and then the control points from the second profile back to the first.

So what if this was a patch from curves instead of a loft?

The great part about this is that now we have a single bezier surface with a v order of 5 and a u order of 2.

We can even remove some control points and maintain the shape by left clicking on Nv:5 and selecting "3". Selecting and order of 2 sort of defeats the purpose of what we are doing because a degree 1, order 2 surface is by definition perfectly flat.

Now that I think of it, our profiles don't need to be N5, they could be as little as N3.

Double click on the front profile, right click on N:5 and change to N:3. In other words 3 control points.

Now the Bezier surface changes to a v order of 3.

Great. So how about we make this symmetrical across the zx plane.

I want to do a highlight analysis on this, to see if it's G3 continuous across the centerline

No surprises, looks good. How about a connection checker analysis?

Ok, it wasn't that bad. I've got two adjacent Bezier surfaces with a v order of 3, u order of 2 which are G3 continous. The simplier the shapes, I can see the easier it is to control them. But these shapes are so simple, they don't remotely resemble my intended object. Let's say that the sides of the monocoque are perfectly flat. I'll have to create a flat surface, and then make these curves G3 continuous with the sides of the monocoque.

I'm going to raise the mid guides

And I'll add a couple more points at y=0, and create a order of 2 line between the points

Combine them

Make a pair of N:2 profiles connecting these guides

Alright, so let's make a patch. We'll see that we have the desired structure, that is u:2 v:2