Best Marine Design Software for Hull Design? (2001-2005)

[vacanti]

Tom -
This is a good time to rail against the overuse of the term FOIL to describe both a 2D SECTION that does not represent an entire wing, keel or rudder but just one chord shape and a PLANFORM that is composed of two or more FOILS to define an entire surface. You are obviously well versed in the subject of foil design and analysis so I am surprised to see you use the term FOIL so loosely.

It is not so simple as you might think. Yes a keel, rudder or wing planform is a solid surface that can be described with a NURB surface. However where you have gone astray in your assumptions is the bit where you assume all that is happening is there is a simple interpolation between offsets in a typical set of NACA data points. First, interpolation with what? A polynomial? Polynomial of what order? A NURB interpolation? What values of control vertices that lay off the foil shape are being used and how were they calculated? Polynomials and particularly interpolated b-spline curves will frequently exhibit "oscillations" (numerical instabilities)where the curve generated alternately moves inside and outside the desired curve. Clearly these methods will result in highly inaccurate lofting or CNC machining. Hence the need for better curve fitting methods.

FOIL 97 does two sets of things that are important 1) Allow a simple user interface for drawing new or copying old NACA foils into NURB curves-not surfaces (with automated methods), ready for use in a surface. 2) Calculate 2d Foil Section lift, drag, separation, turbulence and bubble formation, again with a simple user interface. XFOIL is only for the most dedicated user in my opinion. Its use is not obvious except to persons like yourself who is already familiar with the arcane DOS interface it uses (Version I have seen).

WINGS 32 takes the 2D Section shapes - properly called foils and combines them into a 3D NURB surface. Now we are talking about two different sets of calculations 1) FOIL 2D analysis and 2)PLANFORM 3D Analysis and design. (Hydrostatics, centers of effort, wetted surface etc etc -more than just lift and drag)

FOIL does the first and WINGS does the second in my case. I want to point out that while it is possible to place FOIL design and analysis into a 3D CAD program to design PLANFORMS such as keels and rudders, 2 things happen. 1) The program becomes very complex with two very different sets of functions vying for attention in very lengthy pull down menus, 2) The lengthy program becomes difficult to use and to debug and to price. WHy would I sell a 45,000 line program for $95 ?? I have to maintain the code, the web site, constantly add to it and respond to tech questions for free of course. In my mind separating 2D FOIL design and 3D PLANFORM design makes sense from the perspective of the user (who may not care one bit about foils) for cost and ease of use.

As you are aware a bulbed keel requires vortex lattice or other surface based methods to compute lift and drag. That would mean an automatic panelization method to permit the analysis. There are entire programs and services offered for this function alone. I wrote a short article on CFD for Boatbuilder Magazine and published it on my web site under Tutorials. You will also find two articles on keel design and foil design and development under the tutorial listing as well. I will take a look at your spreadsheet - thanks.

Respectfully -
Dave Vacanti
http://www.vacantisw.com

[tspeer]

Originally posted by vacanti
Tom -
This is a good time to rail against the overuse of the term FOIL ... and a PLANFORM that is composed of two or more FOILS ...

It's probably not worth quibling about semantics, but the term "foil" is used to refer to both the section and the 3D object. It's usually clear from the context which one is meant. The term "planform" is pretty much universally used to indicate the projected outline of the foil rather than the 3D object itself. When the aerodynamicists I know want to precisely refer to the 2D shapes, they call them "airfoil sections" or just "sections". For example, the term "hydrofoil" could refer to the 2D section, a planar surface, a whole assembly of surfaces and struts, or the entire boat itself. Likewise "aerofoil" can refer to the section or the whole wing.

It is not so simple as you might think. Yes a keel, rudder or wing planform is a solid surface that can be described with a NURB surface. However where you have gone astray in your assumptions is the bit where you assume all that is happening is there is a simple interpolation between offsets in a typical set of NACA data points. First, interpolation with what? A polynomial? Polynomial of what order? A NURB interpolation?

But don't you have this same problem with lofting anything? If you were given a number of hull sections as a table of offsets, wouldn't you also have an issue of how to loft the surface in between them? Common practice with wings, and presumably keels, is to use a smooth curve for the section and linear interpolation between widely spaced sections - often just root and tip. That's why the NACA 4- and 5-digit sections used analytical functions to define their shapes. For a CAD system, this is just another developed surface.

What values of control vertices that lay off the foil shape are being used and how were they calculated? Polynomials and particularly interpolated b-spline curves will frequently exhibit "oscillations" (numerical instabilities)where the curve generated alternately moves inside and outside the desired curve. Clearly these methods will result in highly inaccurate lofting or CNC machining. Hence the need for better curve fitting methods.

Again, this sounds like a generic CAD problem with over-fitting of the data. B-splines are a big help in this area, since they can have only local support (depending on the order of the basis functions). Isn't fitting a fair NURBS surface to a given set of mold lines a common CAD function? It might be possible to use something like the Akaike Criterion to balance the order of the fit with the number of data points available so that the fitted surface is the best representation in between the given data points (http://www-ssc.igpp.ucla.edu/personn...eg/node15.html).

FOIL 97 does two sets of things that are important 1) Allow a simple user interface for drawing new or copying old NACA foils into NURB curves-not surfaces (with automated methods), ready for use in a surface. 2) Calculate 2d Foil Section lift, drag, separation, turbulence and bubble formation, again with a simple user interface.

Putting a simple user interface on the Eppler code is definitely a useful thing! It's a shame to lose the inverse design capability though - that's the best part of the code.

Eppler's handling of separation is very crude. He calculates the separation point based on a fully-attached pressure distribution, then applies a simple empirical reduction to the lift that amounts to neglecting the surface behind the separation point. I don't believe he makes any changes to the drag.

Again, I'd encourage you to add the capability to combine the 2D section data with the planform effects to calculate the 3D lift and drag of the keel/board/rudder/wing/foil with an intuitive user interface. Otherwise, the user is still missing big pieces to the puzzle.

XFOIL is only for the most dedicated user in my opinion. Its use is not obvious except to persons like yourself who is already familiar with the arcane DOS interface it uses (Version I have seen).

I admit there's a definite learning curve with XFOIL - but nothing like the native Eppler interface!! Still, the power of the analysis and design tools makes it worth it. None of the other 2D section programs even comes close.

I believe John Roncz wrote an article a number of years ago where he pointed out some of the short-comings of the Eppler code's boundary layer results for convex pressure recoveries and heavily aft-loaded sections (admitedly not a problem with symmetrical keel & rudder sections). Eppler's code was a real break-through 40 years ago, but it's a bit dated now. Still, it's probably received more wind tunnel and flight validation than any other, and that counts for a lot.

WINGS 32 takes the 2D Section shapes - properly called foils and combines them into a 3D NURB surface. Now we are talking about two different sets of calculations 1) FOIL 2D analysis and 2)PLANFORM 3D Analysis and design. (Hydrostatics, centers of effort, wetted surface etc etc -more than just lift and drag)

Ah, now I see. I guess I should check out your website to see the difference. From you post, somehow I had the impression that the capability of both FOIL and WINGS were combined into one program. Does FOIL help the user select from a large number of existing sections like http://www.davincitechnologies.com/AirfoilOptimizer.htm, or use an inverse design to refine the shape?

FOIL does the first and WINGS does the second in my case. I want to point out that while it is possible to place FOIL design and analysis into a 3D CAD program to design PLANFORMS such as keels and rudders, 2 things happen. 1) The program becomes very complex with two very different sets of functions vying for attention in very lengthy pull down menus,

Yes, I agree. I wonder if one could combine them together by showing a wireframe of the wing with the sections in place, and clicking on a section would take you to the menus for section design and analysis. Then you'd return to designing the planform and analyzing the 3D wing with the new section characteristics included in the analysis.

2) The lengthy program becomes difficult to use and to debug and to price. WHy would I sell a 45,000 line program for $95 ?? I have to maintain the code, the web site, constantly add to it and respond to tech questions for free of course.

Whoa - I apologize if I sounded like I was running down your program or questioning the cost. I didn't mean to do that. I was confusing WINGS and FOIL, and perhaps I should have wondered about the difference between WINGS and the rest of the CAD system. Again, I'm not trying to throw rocks - I'm genuinely interested in understanding the pro's and con's of the different programs out there.

I realize there's a world of difference between a commercial software product and an academic one. And I expect you are paying royalties to Somers and Maughmer for part of those 45,000 lines, since they were given the US distribution rights even though there was an older version of the Eppler source code published in a NASA document. So that has to add to your costs, too. I don't think what you're asking is at all out of line.

...As you are aware a bulbed keel requires vortex lattice or other surface based methods to compute lift and drag. That would mean an automatic panelization method to permit the analysis. There are entire programs and services offered for this function alone.

For sure a bulbed and winged keel is a different animal. Do you think you could do an adequate job with just a single vortex lattice along the quarter-chord line and control points at 3/4 chord? The VLM has proven to be surprisingly accurate for most purposes, and about all you'd have to do for panelization would be to determine the spanwise spacing. With today's PC's, you could take a brute force approach and include far more elements than necessary instead of trying to finesse their distribution.

I wrote a short article on CFD for Boatbuilder Magazine and published it on my web site under Tutorials. ...

I saw it - good article.

Cheers,

[Leo Lazauskas]

Originally posted by vacanti
[b]Leo-
Thanks for the question - We don't have plans just yet for Physics based drag models. Do you have a source where I might find means to implement such analysis methods? Keep in mind the methods need to be reasonable and compliment the use of NURB surfaces.

[snip]

There are literally dozens of mathematical models and implementations. A good source to start with are the Symposia on Naval Hydrodynamics. The latest was the 24th Symposium held in Fukuoka, Japan during July 2002. Earlier symposia are well worth exploring.

I believe that Newman and colleagues were using B-Splines in sort of their hydrodynamic codes.

Leo.

[Ezhov]

Dear Sirs,
On my point of view the best soft is Sea Solution
See www.seasolution.com
I use this soft more then for years.
Dounload and try to use.

Hi Ezhov,

Could you clairify - In this thread http://forums.boatdesign.net/showthr...?threadid=1369

It seems the developer said SeaSolution was brand new and thus a price had not been decided on or released yet:

"We only announce the system.
The marketed version and prices will be ready in the May of 2003."

How have you been using it for 4 years if it's brand new? Or have you been using an older commercial version of the same software that was released 4 years ago? Or a demo version for all that time???

P.S. and SeaSolution looks interesting - I'll definitely give it a look after prices have been released. Before then, since it might be $1,000 for a license or it might be $10,000 for a license there's no sense in looking at something that might not be in my budget...

I'm also very confused by how long it's been around. Is it brand new, or an incremental version upgrade of a program that is established?

[vacanti]

I began the development of PROLINES as a hull design program back in 1985. Over the last 18 years I have become very familiar with virtually all of the competing programs. I have never heard of SeaSoultions before now. I found it very curious that no where on the Sea Soultions web site was there an address, fax number or other information giving the source of the programs. We have seen several programs from France, England, the Netherlands, Australia and a few other countries and all disclose their origins.

You will note that it is not listed here in Boat Design (http://boatdesign.net/software/index.htm) on the recommended software lists so it must be new or at least not previously known. There are no limits to who can develop software and a new developer may be just fine.

I recommend that you download demos from Sea Soultions and other competitors and try each one. Then write to the companies and evaluate the customer support that you get. Then make a choice.

Dave Vacanti
Vacanti Yacht Design LLC
http://www.vacantisw.com

The poll seem to place the software on hull form generation with calculation capacity, in the order of performance and cost.

Cad system can produce the geometry in 3d. Some function in getting area of surfaces and some can get to the CG. But it is more difficult to map out inertia etc. This manual process is not preferable. Unless the model is in solid 3d, that one will measure the weight for displacement and CG or inertia if you are CAD pro only. Cadkey is an alternative to integrate bodies with varying sections. It is compatible to Acad. There are other lower price software that can do similar, but are not Cad oriented. The higher cost software like ProEngineer can do more. For these and other reason, I think the AutoShip set compatible to GHS is attractive in dealing from start to finish, i.e. setting form and able to calculate a full set of calculations.

Peter

[terhohalme]

Rhino and Phaser hydrotatics is a great combination. I have used AutoCAD and Mechanical Desktop for a very long time (from Autocad 2.6). But two years ago I tried Rhino and now I am on that way...

I also have a Prosuf 3, but the use of it's 3D is far too difficult compared to Rhino.

What I still want is a possibility to use Rhino geometry in a windows version of Michlet for dynamics. That would be a very good reason to purchase Michlet.

Leo! Are you there somewhere?

Terho

[StreamWork]

On my point of view the best soft is Tribon M2+ Rhino 3.0, and then use for fotorealistic images 3DMax 5.1+Brazil

[ochaves]

Solidworks with surface works

[Paul B]

Hi Osvaldo.

When using Solidworks to loft a hull surface do you have a problem with the bow waterline becoming hollow? How do you solve this?

See the sample in my gallery.

[duluthboats]

Paul,

Welcome to the forum.

This happens when lofting in most software. In Rhino it helps to extend the shear line past center, copy the stem profile to the end of the line, loft, then trim the surface back to the center line.

Gary

[Paul B]

Thanks for the suggestion. I'm sure something like that can be done to make it look better. I guess you're talking about lofting one side of the hull, trimming back, then mirroring along the CL.

I don't think you are really controlling the shape properly if you do it like that. If the "spline" is bending concave to make the hollow it is surely popping out a bit on the convex side.

I was hoping that the SW user could point me to a way to toggle the spline type, or some other solution. I'm not sure if the mode of loft would make a difference either, base loft or surface loft.

[ErikG]

In PBB no 79 Eric Sponberg mentions these problems with computers creating hollow waterlines at the bow. As far as I've come on testing, it seems that this is a problem that has to do with the math that controls the code in the programs.

Max Surf has the same problem. You have to adjust, adjust and adjust again until you get the shape you want. I use a large number of wl and stations and buttockswhen creating the design just to be able to look at the curvature in many different areas.
It's easy to fool yourself thinking hat the wl looks good, but just a little bit above or below the dwl the hollow is back!

When looking at the control points/lines it'll look a bit strange but the wl will be ok.

ErikG