How to get hollow entry...

[Grant]

Hi,  I have been trying this for a week, but just do not seem to be able to get a feeling for what control points influence the waterline entry angle over two columns or so of control points.  I want an entry of 20 degrees, and for the waterline to come back in a shallow hollow for about a quarter or third of the Dwl.  I can get the hollow for about 15%, then it goes to a normal curve, but even that transition, and onward to the back, I can not get smooth.  I think part of the problem is that I have created a plumb bow and a pretty tight bilge, but still, I believe it should be possible.  I have moved points in out, columns back and forth, but with no really decent results.  Can someone out there give it a shot, so I can see where my ´feeling´ is missing?

Here is the create hull script of my base design (just copy and past into the command line).  The top one is for meter units, the lower for millimeter units.

-_OrcaCreateSailboat LengthOnDeck 5 BeamOnDeck 1.8 TransomBeamRatio 0.9 StemDeckHeight .600 TransomDeckHeight .450 TransomHeight -.050 Draft .200 StemRake 0 TransomRake 20 SheerHeightRatio 0.85 SheerHeightPos 0.25 DeckBeamPos 0.5 DraftPos 0.35 SectionFullnessFactor 0.1 BilgeTurnFactor 0.3 ForefootShape 0.5 NumberRows 6 NumberCols 7 enter

-_OrcaCreateSailboat LengthOnDeck 5000 BeamOnDeck 1800 TransomBeamRatio 0.9 StemDeckHeight 600 TransomDeckHeight 450 TransomHeight -50 Draft 200 StemRake 0 TransomRake 20 SheerHeightRatio 0.85 SheerHeightPos 0.25 DeckBeamPos 0.5 DraftPos 0.35 SectionFullnessFactor 0.1 BilgeTurnFactor 0.3 ForefootShape 0.5 NumberRows 6 NumberCols 7 enter

[lleibman]

Hi Grant,

Keep in mind when editing the control net that if you can't seem to generate the shape you want no matter how much you manipulate the control points, it may mean that you just don't have enough control. In general we preach keeping as simple a control net as possible, but sometimes you can't avoid adding control.

I've attached a hull based on the Hull Assistant input you provided. I've added two control net columns in the forward half of the hull to create the hollow waterlines forward. As you noted, you can get the hollow without adding these columns but to properly fair out the hollow aft you would have to carry the effect pretty far. By adding the additional control columns you can fair out the effect more quickly at the expense of a slightly more complex control net.

Larry

[Grant]

Excellent, a picture is worth a thousand words.

But, yes, there is always 'but': when I add waterlines, say one every 20 cm, and turn on the curve graphs, what I see is not really a fair line... is this my system, or am I mis-using the curve graph or what.  See attached picture.

In anycase, have a happy new year!!

Grant

[lleibman]

I think part of what you are seeing is that I didn't spend much time trying to fair the shape, I was just trying to introduce the waterline inflection. This could presumably be improved upon by some local fairing of the control points.

Another part has to do with how Rhino creates section cuts. We have seen cases where we are pretty confident from the surface geometry that a section should be fair but the sectioning algorithm (which internally fits a Nurbs curve to computed surface intersections) exhibits these apparent curvature jumps. Relaxing the Rhino tolerance may help alleviate some of this.

Also keep in mind that because curvature is derived from the second derivative of the curve, somewhat abrupt changes in curvature do not necessarily imply unfairness. That said, I would do a little more fairing on the shape and re-examine the curvature plots.

[Grant]

Hi Larry,
I have been working on the design on an off.  I have learned more about the curvaturegraph lines, and it does seem its a problem with how Rhino works, but regardless, I actually got the shape pretty fair, even the curvaturegraph lines, so there is hope for me yet!  ORCA3Ds Move Points is indeed handy.  However, please remove the .2 degrees default and change it to zero - I keep forgetting to do that when I am just trying to go horizontal or vertical... :-)

Questions:
1. When you inserted the columns, did you use insertcontrolpoint or insertknot?  I found that insertcontrolpoint changes the shape of the hull, but insertknot does not.  But given their names, I would have guessed the opposite would have happened. Anyways, is there any reason why I should not use insertknot?
2. Along the same line, are the control points these Rhino commands insert taken up as ORCA3D control points with all the same functionality, etc?  Hope so.

Cheers,
Grant

[lleibman]

Hi Grant,

Glad to hear you are making progress on the fairing. I guess the fact that it's not a trivial task means added job security for you. I will add the OrcaMove functionality request to our development list.

Regarding your questions:

1. I used insertcontrolpoint. Your observations about the effect of insertcontrolpoint and insertknot on the shape of the hull are right on the mark. Both functions add a column or row to the control net. Insertknot uses a methodology known as the "Oslo algorithm" (named after the person not the place) which determines the location of the newly inserted control net row/column as well as the locations of nearby rows and columns so as to maintain the same surface shape. In doing so, not only might it have to move adjacent rows/columns in the control net, but it will likely need to modify the NURBS knot vector in the associated parametric direction. The resulting knot vector will likely be "non-uniform" (the N part of NURBS). Insertcontrolpoint presumably retains the uniformity of knot vector if it was uniform to start with but as you found does not maintain surface shape in general.

The obvious question becomes, why do we care if the knot vector becomes non-uniform. We really don't that much. However, when editing the surface by directly manipulating control points, you will find that with a uniform knot vector, the influence of moving a control point is symmetrical with respect to the associated knot. In other words, the surface distorts somewhat equivalently on either side of the associated knot in the surface. When you have a non-uniform knot vector, the influence is asymmetric, the surface distorts more on one side than the other. I see the Rhino help file uses the phrase, "better point pulling behavior" for the uniform case. In reality, I think you would have to the knots quite nonuniformly distributed (ie two knots very close together) to really notice this. We generally use the rule of thumb that if we are pretty early in the fairing process and realize we need more control in the net, use insertcontrolpoint because you don't care too much that the surface shape is changing (in fact you want it to). If you are near the end of your fairing process and just can't seem to get the shape just right but don't want to make any significant changes, or you are adding changes to an existing design, then insertknot might be the way to go.

2. Yes, the control points added by these two commands are included when you use the OrcaPointsOn command.

[lleibman]

Grant,

I was looking at your request to modify the OrcaMove default nudge value more closely and want to make sure I understand. Are you referring to the default angle nudge of 0.2 degrees? That value should only be relevant when you are using "Polar coordinates". It should be the nudge step when clicking the spin button next to "Angle" in polar coordinates. Otherwise it should have no effect. If you wish to move only horizontal, you can click the "X" or "Y" spin buttons in cartesian coordinates and "Z" should remain unchanged. Are you finding this not to be the case or is there some other behavior you are referring to?

Larry

[Grant]

Hi Larry, thanks for the great reply, that helps a lot!
Regarding the move points, you are right, I was confused about how the nudge worked, so all is OK.
Cheers,
Grant