Developed Surfaces????

Discussion in 'Boat Design' started by ChrisGibbs, Aug 20, 2003.

  1. Chris Krumm
    Joined: Aug 2003
    Posts: 92
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: St. Paul, MN

    Chris Krumm Junior Member

    Morgig -

    I think we're in 100% agreement. This discussion has ranged over the process and mathematics of checking for developability along with a number of references to current software that will do the job for you from as little as $200 into the thousands-of-dollars range.

    programs by New Wave Systems, for example, give you gaussian curvature maps to check for developability along with plate expansions. It'll even force developability on patches of the hull if you like, provided you can live with it adjusting your hull lines.

    Understand how you don't want to mess with trimming thick plates on big boats. Prosurf and ProBasic have been used to build some large tugs, brges, and who knows what else, along with the plywood projects of us amateurs.

    The software is a tool for most of us. Some earn their bread with the tools, others play with the tools, and yet others play with making the tools.
     
  2. Ian
    Joined: Apr 2004
    Posts: 22
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Camden, Maine, USA

    Ian Junior Member

    developable surfaces

    I know that you have received a great number of very interesting and very technical replies--but on the off-chance that you just want to get going with something functional--I suggest that you look into a program called Plyboats. It is very cheap, very easy to use, more flexible than it at first appears--and the results are buildable.
     
  3. carlg
    Joined: Mar 2003
    Posts: 7
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Fremont, CA

    carlg Junior Member

    I'll second Plyboats. In spite of its low cost, which is now around $40 to $50, it is extremely useful for plywood hulls. I've built five boats using Plyboats and all the panels have fit perfectly.

    It's unfortunate that the program is passed over in forum discussions.
     
  4. Andrew Mason
    Joined: Mar 2003
    Posts: 397
    Likes: 18, Points: 18, Legacy Rep: 206
    Location: Perth, Western Australia

    Andrew Mason Senior Member

    I think Morgig makes a valid point that is not widely appreciated by many designers. The perfectly developable surface is usually not the surface that is most desirable from the point of view of the designer. Typically on a powerboat hull with twist in the bottom panel forward, the truly developable panels have too much convex curvature to be ideal.

    However, by modifying the sections toward a more ideal shape, it is often possible to come up with a surface that is just as buildable as the truly developable surface. The reason for this is that we are not dealing with inelastic materials, even steel plate has some elasticity that allows it to be persuaded into a slightly non developable shape.

    This is one reason that we take the approach we have in Maxsurf. We show, using marker points, where the true developable surface will lie, and the designer chooses whether to have his surface conform to those points or whether to deviate from them.

    This approach also allows another important design function - the ability to intentionally make a portion of a surface non-developable. For example, on the topsides of a typical powerboat hull, the aft 2/3 of the surface may be perfectly developable. Up forward, however, the designer may choose to sacrifice developability to create a significant amount of topside flare. Maxsurf allows the designer to do this within a single surface, and to measure the amount of strain (i.e. compound curvature) in the resulting plates. Apart from small boats, this is a typical requirement of steel and aluminium workboats - make as much of the boat as developable as possible, but sacrifice developability in those areas that have a critical influence on seaworthiness and performance.

    Remarkably, the use of Maxsurf to design surfaces that are practically developable has resulted in examples that were easier to build than truly developable design. The example I can give was one builder in the N.W. U.S. who was building a workboat with a deep forefoot at the bow. This is a hard shape to get developable but it can be done with fairly full sections. One plate in the bow, although developable, needed to be formed around a radius of curvature of about 1 meter. When the design was modified to give straighter sections and less developable plates, the minimum radius of curvature increased to around 2 meters, resulting in less force being required during the plating process.

    We recommend with Maxsurf and Workshop that if a plate is developed and has a maximum internal strain of less that 0.25%, the plate can be considered to be practically developable. We have had builders who have relaxed this recommendation to 0.4% with good results. As a result I have concluded that the enforcement of perfect developability by many programs is an artificially strict requirement that results in the unnecessary compromise of seakeeping and resistance characteristics for many designs.

    As an additional point, I have not used Plyboats, only seen printouts, but the impression I got from these was that it only did straight section designs. This is fine for untwisted panels, but results in large discrepancies from the developable surface when a panel is twisted. Ironically, I think that the reason that Plyboats has produced many reasonable boats is for exactly the reasons described above - the inherent elasticity of plywood allows a panel to be tortured into a shape that is not developable, allowing the straight sectioned designed to be panelled effectively while retaining straight sections.
     
  5. Tim_Hastie
    Joined: May 2004
    Posts: 55
    Likes: 1, Points: 0, Legacy Rep: 14
    Location: Canada

    Tim_Hastie Junior Member

    I currently use AUTOSHIP and have so for many years. I build metal boats from 20' to 65' . I have found AUOTSHIP to be perfect every time. Keeping in mind "Garbage in---garbage out"
     
  6. maximillian
    Joined: Jul 2004
    Posts: 4
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: m'sia

    maximillian New Member

    i'm currently doing a final year project on creating a database on 'development of developable surface area'. since knowledge in this topic is very shallow, and i've to start from scratch. i need some advice where needed to write a program, using parameters such as common elements, directrixes etc, to determine whether a surface is developable.
     
  7. Chris Krumm
    Joined: Aug 2003
    Posts: 92
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: St. Paul, MN

    Chris Krumm Junior Member

    Maximillian -

    Read the earlier posts in this thread, particularly those by tspeers (Tom Speers). Also look for my references and links to the Brian Konesky papers. I think that points you in the right direction regarding the math and background on methods that could be implemented in computer algorithms.

    Chris Krumm
     
  8. Andrew Mason
    Joined: Mar 2003
    Posts: 397
    Likes: 18, Points: 18, Legacy Rep: 206
    Location: Perth, Western Australia

    Andrew Mason Senior Member

    maximillian

    If you mean that you wish to analyse and existing surface for developability, rather than creating a developable surface from scratch, the procedure is straighforward but does require some math.

    The definitive test for developability is whether the Gaussian curvature of all points on the surface is zero. Gaussian curvature is simply the product of the two principal curvatures at a given point of the surface. The principle curvatures are the maximum curvature and the minimum curvature, by definition the directions that these two curvatures run in are always exactly 90 degrees to one another (a perfectly flat surface is a special case, the principal curvatures are both zero, the Gaussian curvature is also zero, and the principal curvature directions are undefined)

    For a developable surface, the minimum curvature is zero, in other words the surface is straight in one direction. This is the familiar directrix or ruling of the standard development methods. The curvature in the perpendicular direction may be positive, negative or zero, but the product of the two curvatures (i.e. the Gaussian curvature) will always be zero.

    To test for developability analytically, you need to know the principal curvatures for any point on the surface. To calculate the principal curvatures you need to know the first and second derivatives of the surface in the u and w parametric directions.

    The calculation of the principal curvatures from the partial derivatives is covered in the book "Geometric Modeling" by Michael E Mortensen (pp 280-285).

    The calculation of the partial derivatives of a NURBS surface is covered in several books, including "Mathematical Elements for Computer Aided Design" by David F. Rogers, "An Introduction to NURBS" bt David F. Rogers, "The NURBS book" by Les Piegl and Wayne Tiller.
     
  9. maximillian
    Joined: Jul 2004
    Posts: 4
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: m'sia

    maximillian New Member

    Thanks so much, Andrew and Chris.

    Chris, i see that you've played with the surface using Visual basic. i'm now doing the same too, but not much progress. i've yet to define a specific objective for my project. but i'm planning to proceed with developable for all the hull, most probably using fixed apex for all the connicle. and i've yet to manage to access to the Konesky's paper. Basically, what do you find in your program(Visual basic), i means what's the inputs (parameter) and outputs?
     
  10. BIG MAC
    Joined: Aug 2004
    Posts: 12
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: TEXAS

    BIG MAC Junior Member

    easy developing in 3d cad

    if you have a hard time following this, it's because i can't explain it - not because it's difficult.

    1. take your lines and create ruled surfaces. this divides the curves into triangles with vetices lying on the curves. the more divisions you make with the rulesurf variable, the more points you will have on the curves.
    2. get in a 3d view that is convenient, exlplode the surface to get individual lines.
    3. copy the first triangle, bow or stern, and move it a little away from the curves
    4. reset the user coordinate system to align with the single triangle you just copied.
    you now will only use 3 commands: distance, circle, and line.
    5. here is where a very simple procedure may get hard to explain. you have your 3d curves divided into 3d lines forming triangles and you have the one copied tringle, now parallel to the coordinate system. on the curves, you are going to measure the length (distance command) of the next lines next to the first triangle. any units are fine so set units to decimal and accuracy to thousandths to save some keystrokes. the distance comand will give you several distances - the actual line length, length in the x axis (as though it were in the corodinate plane, which it isn't), lenth on the y axis, etc. just use the actual line length. pick a line of the next adjacent triangle on the curve and measure its length and then draw a circle on the single triangle with the same diameter, from the same vertex. on the curves, measure the distance to the next line's endpoint form the other vertiex and draw a circle of that radius from the corresponding vertex of the single triangle. where the circles intersect is the endoint of that next line of the adjacent triangle - so draw a line from there back to the single triangle. this is now that next line projected into the current xy plane. keep doing this for all lines of the triangles of the curve and you have your projection. the more divisions you create with the rulesurf command, the more lines you have to process and the smoother the reult will be.
     
  11. maximillian
    Joined: Jul 2004
    Posts: 4
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: m'sia

    maximillian New Member

    developable surface

    i knew that to obtain the developablility, the Gaussian curve must be zero. the problem however, is that i'm trying to develop the surface from scratch. The most important part now is the math algorithms to represent the surface. i'm currently trying to use spline, such as B-spline or bezier to define the space curves between the surface, say, the chine and keel. I've also obtained some constraints from the Konesky's paper for developable surface. Is it possible to determine the algorithms for the surface, under these constraints? or do i need to obtain it through the sweep surface?
     
  12. Jeff D.

    Jeff D. Guest

    Working Backwards

    Wonderful discussion on developable surfaces, but I wonder if you'll indulge me trying to work backwards for a moment.

    For the sake of arguement, let's say that I have the lines drawings for a vessel not necessarily designed to be built from sheet material. Let's further assume that the drawings are in AutoCad, or some other 2-D drafting program, so I can do conical projections the old fashioned way, but with a bit better accuracy. Mathematically, I could also determine the equations of the lines, as they are all simple radiused curves. No splines, higher order curves, etc.

    As opposed to using cones to generate lines, which would insure a developable surface, I have been trying to find cone locations which correspond to the existing lines, thereby insuring that the existing surface is buildable from sheet material. Trial and error isn't working. At least I don't think it is, but there is always the possibility that I am trying to prove the developability of an undevelopable surface!

    Does anyone have any suggested techniques, mathematical or graphical, to verify that a surface is developable if given the lines drawing?

    Thanks in advance for your help,

    Jeff
     
  13. maximillian
    Joined: Jul 2004
    Posts: 4
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: m'sia

    maximillian New Member

    mathematical solutions

    it is always good to prove the developability of all the surface, or at least make it developable.
    But i do think that working backward (Kilgore's method) is very tedious and inaccurate. Furthermore, all you are doing is actually multiconic development method.
    What i am trying to develop is a set of simple algorithm to represent the surfaces of the hull, using spline function as a beginning, and its derivatives for conditions. this is relatively simpler and more accurate compare to graphical methodology.
    so there's no need to going back on the Kilgore's method.
    Even so, your comment has been very helpful.
    Thanks.
     
  14. Gilbert
    Joined: Aug 2004
    Posts: 525
    Likes: 5, Points: 28, Legacy Rep: 28
    Location: Cathlamet, WA

    Gilbert Senior Member

    Kilgore's method is not a multiconic development method. It is the rotation of two 3d lines about the axis of of a tangent point on one of the 3d lines to determine a ruling line which is tangent to a point on the other 3d line. I am not aware of a more accurate method. Most of the computer software I've seen that claim to generate ruling lines are a joke. And it's obvious when they are a joke that they are not using Kilgore's method. What use would you have for software that does not work "backward" as you put it when you want to know if the surface is developable?
    My 2c.
    Gilbert
     

  15. bholtermann
    Joined: Nov 2004
    Posts: 5
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Richmond, VA

    bholtermann Junior Member

    Surfaces in 3D

    Hi Guys,
    I am new to the site just today and am trying to reply to a question I saw about CAD CAM and a router to make some small models... I have a CNC machine in the garage and am willing to help if I can.
    In my experience pretty boats are mostly elipses, one or more together. I have used Design CAD to prepare section views. This is the way boats are ofter presented with cross cuts running from sem to stern. BobCAD can surface it from there.
    I would need DXF files so if you can get that far, or if I can help let just me know? I've been around boats all my life and do not expect that to change any time soon.
    Thanks,
    Bruce
    bholtermann@sealeze.com
     
Loading...
Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.