Modelling Stringers - Solidworks

Discussion in 'Software' started by SLM, May 24, 2014.

  1. SLM
    Joined: May 2014
    Posts: 16
    Likes: 0, Points: 1, Legacy Rep: 10
    Location: Sydney, Australia

    SLM Junior Member

    Hello,

    I have designed a hull shape in Freeship and then exported this to Solidworks using an IGES file. This hull is to be plywood skinned over frames and stringers.

    This is my first attempt at modelling a boat in Solidworks (although I have used Solidworks for years). I have quickly come up against a problem modelling the stringers. The stringers are wood say 40 x 25mm and wrap around notches in the frames and need to sit against the surface defining the skin. They will twist as they wrap around the hull. I have a line on the hull surface that represents path the stringer should follow. What is the best way to model these stringers and generate the notches in the frames. Any tips ??

    Some of the stringers are running down the center of plywood panels and others follow the chines and need to be chamfered to match the two intersecting panels.

    Thanks,

    Steve
     
  2. TANSL
    Joined: Sep 2011
    Posts: 7,369
    Likes: 699, Points: 123, Legacy Rep: 300
    Location: Spain

    TANSL Senior Member

    You could make an extrusion of a rectangular shape (stringer section) along a path (the curve on the hull).
    You may considrerar the chine as equivalent to a stringer. I mean, you do not need stringers near the chine.
     
  3. SukiSolo
    Joined: Dec 2012
    Posts: 1,269
    Likes: 27, Points: 0, Legacy Rep: 271
    Location: Hampshire UK

    SukiSolo Senior Member

    This is briefly how I would approach the modelling problem.

    1. Create a new curve in SolidWorks along the chine line.

    2. Extend the curve at both ends.

    3. Create sketching plane normal to the curve, most likely the transom end especially if it is a vertical face.

    4. Sketch stringer, and extrude (maybe sweep) along curve. It should twist correctly so check any options for normals.

    5. Trim (or cut) the stringer with the offset (ie inner face of the hull). This creates the end result stringer.

    This is a very simple model in Rhino btw, and yes I have used SolidWorks but not for a while. In the real world when you build the stringers you may find that it is better to laminate them to eliminate stress.
     
  4. John Perry
    Joined: Nov 2003
    Posts: 308
    Likes: 53, Points: 28, Legacy Rep: 129
    Location: South West UK

    John Perry Senior Member

    I think there are many ways to do this. The first idea that comes to mind is to use the 'sweep' feature together with a 'guide curve'. The sweep feature basically requires a cross section shape and a sweep path, guide curves are optional extras.

    There are various ways to specify the path, depending on how you want the stringer to run along the length of the boat. For example you could make the path the intersection between the inner skin surface and some plane you have specified, or you could make it the intersecton between the inner skin surface and some other wavy surface that you have set up. Then you can make a guide curve as the intersection of the inner skin surface and another plane spaced 25mm from the first plane, or a wavy surface spaced 25mm from the first one.

    You could make your cross section a quadrilateral in a sketch and specify right angle corners at the two inside edges and also specify your 40 and 25mm dimensions on two adjacent sides of this quadrilateral. Another corner of would lie on the end of the path and a the fourth corner on the end of the guide curve. I havent actually done it that way myself, but it is perhaps the most obvious way.

    When making such features I would nearly always make sure the 'merge part' box is not ticked, so that the stringer becomes a separate body to the hull skin. I do like the 'multibody parts' in Solidworks, indeed I cant see why they really need assemblies of parts distinct from multibody parts - couldnt everything just be a multibody part and when you want one bit to move relative to another you just adjust the constraints between them as required.

    Another way would be to make the stringer as a loft, with two guide curves.

    Another way would be to make the stringer either as a loft or a sweep but with a rectangular section 40x25 mm right along its length, then use the surface cut feature to make the inner hull skin cut away the unwanted material from the stringer. This follows the procedure that you will use when you build the boat, you will fit the stringer with its rectangular section intact then apply an angle grinder, sorry, carpenter's plane, to shape the outer surface to fit the skin. Disadvantage is that the outer surface becomes a complex 3D shape, may slow down your rebuilds but if you have a powerful computer you may not care.

    When it comes to the notches in the bulkheads I would set a plane at one side of the bulkhead and sketch the bulkhead with the notches. To include the corners of each stringer in this sketch a good way is to make these corners pierce points with the edges of the stringers. If you do the alternative and intersect the bulkhead plane with the surfaces of the stingers you will get lots of complex splines in your sketch, which makes your files unnecessarily complex but you may not care. When you extrude to make the bulkheads the direction of extrusion will depend on whether this is wood or metal construction - for metal extrude towards the larger side so as to leave a gap that weld metal will fill, for wood extrude towards the thinner smaller side so as to leave a bit to plane off, unless perhaps you also want a small gap to fill with epoxy. Of course, if you really want detail, you can accurately include the weld prep for metal construction.

    Yet another way for the bulkheads - make each bulkhead a rectangle larger than the largest cross section of the hull then make a knitted surface out of the inner hull skin and all the stringers then use this surface to cut away the outer parts of all the bulkheads. I have done this and it is a quick way to make the bulkheads but the disadvantage is that it is too good - every edge of every bulkhead becomes a complex 3D shape which perfectly fits but that is more detail than you ever need for building and will probably slow down your computer.

    Somewhere in a thread called 'post your own design ideas' there is a catamaran I drew a few years back. I added more detail and refinement after posting these pictures but then left it, I keep drawing these things then never build them (perhaps fortunately). The stringers in that boat were easy since I wanted them to run horrizontally with level top and bottom surfaces so that they could easily be used as supports for bunks, cabin table, shelves etc. So I only had to set horizontal planes at the right heights, intersect with the inner hull skin, complete a sketch and extrude verticaly - never mind an imperfect fit with the hull side, its only fractions of a millimeter and this would be built with epoxy!
     
  5. SLM
    Joined: May 2014
    Posts: 16
    Likes: 0, Points: 1, Legacy Rep: 10
    Location: Sydney, Australia

    SLM Junior Member

    SukiSolo:

    I have tried the approach you suggested. Yes I could probably live with this but it is not ideal. I guess I was hoping to find a method that would allow the face of the stringer (the 40mm dimension) to be constrained tangent to the outer loft line. I will fiddle with this a bit and see how well it works in some real examples.

    Re Rhino. I have used Rhino a lot in the past. However I have decided that I really need to use Solidworks going forward because it allows the generation of working drawings much more easily. Solidworks does have a number of significant issues though. Another being generating flat patterns of plywood skins. It has some capability in this area via the sheet metal functions but it will not flatten the panels on some sample multichine plywood hulls I have played with. I think Rhino would do a better job there .....
     
  6. SLM
    Joined: May 2014
    Posts: 16
    Likes: 0, Points: 1, Legacy Rep: 10
    Location: Sydney, Australia

    SLM Junior Member

    John Perry:

    Thanks for taking the time to write such a long response.

    There are some good thoughts there re the use of guide curves. I need to have a play with these ideas on my hulls and see if they are practical. As per my post above to Sukisolo, the challenge is to ensure that the stringer is tangent to the hull (round bilge) or makes an equal angle to each of the two panels intersecting at a chine (hard chine hull). However I can live with some inaccuracy.

    Great guidance on generating the notches in the bulkheads as well. I will play around this week in my spare time with these ideas and see how it goes.

    Since you seems to have some experience modelling hulls in Solidworks - do you have any thoughts on how I can flatten the plywood panels in a multichine hull to make the flat patterns ? I work a lot with the sheet metal function in sheet metal parts I design (not for boats) but I have not had any luck getting this functionality to flatten a hull panel. The shape seems to be too complex for it so work with (even if it was a developable surface in Freeship).
     
  7. SukiSolo
    Joined: Dec 2012
    Posts: 1,269
    Likes: 27, Points: 0, Legacy Rep: 271
    Location: Hampshire UK

    SukiSolo Senior Member

    SLM, you may well be able to constrain the sweep by using another offset curve inboard. It should at least allow the correct Normal.

    I have before now unfolded aircraft wings in Rhino for SolidWorks users. It worked fine and showed where the aluminium was 'stretching'. My gut instinct is to unfold the panels in Rhino and import them into SolidWorks if it cannot develop them. Yes Rhino is rubbish for 2D drawing and I use a 2D package to clean up and dimension the output. Much quicker for dimensioning, bulkheads, transoms, c/board case et etc. So I fully understand how the excellent 2D drawing generation of the 3D model is so useful.

    It surprises me that SolidWorks has not quite caught up, but then Rhino is a Solid Gold partner. I have also used Pro-E quite a bit and even with that I have occassionally resorted to some of the capabilities of Rhino and imported the result. The required addon software for Pro-E for the type of modelling was simply not worth it.
     
  8. SLM
    Joined: May 2014
    Posts: 16
    Likes: 0, Points: 1, Legacy Rep: 10
    Location: Sydney, Australia

    SLM Junior Member

    Thanks SukiSolo - great idea to import the SW model into Rhino and use that to develop the skins. I will give that a go later today when I get a moment.

    I have used Pro-E as well - also Catia. However I like Solidworks for its ease of use - not to mention that that is what I use in my day job. I find Rhino (at least V4) to be very powerful but not the easiest to use. I have not used V5.
     
  9. John Perry
    Joined: Nov 2003
    Posts: 308
    Likes: 53, Points: 28, Legacy Rep: 129
    Location: South West UK

    John Perry Senior Member

    I realise from your recent post that at least some of the longitudinals that you referred to as stringers are what I think many boat builders would call chine logs - i.e. lengths of timber used to join panels at a chine. That being the case, I think you may find it easier to model them as lofts rather than sweeps, although both would probably work. Using lofts gives you close control over the cross sectional shape at as many positions along the length as you may feel necessary. As with real boat building, programs like SW give you many different methods to achieve a similar end result and you have to experiment a bit to find the methods that suit you best.

    Yes, I have had problems flattening curved panels with Solidworks using the options under the sheet metal menu- sometimes it works sometimes it doesnt! Considering the colossal amount of work that must be required to produce software such as SW (I should know since I was at one time involved in writing bits of code for CAD programs) it seems a pity that they didn't go the extra mile to include a robust algorithm to do this. I understand that there are at least two third party plugins available to provide enhanced surface modelling for SW, so I perhaps these do provide that functionality.

    If we consider a hull panel spanning between two chines in a multichine hull, I think the requirement for the SW panel flattening to work properly is that two adjacent corners of the panel (e.g. the two corners at the transom in the case of a transom stern hull) must lie on a plane that is tangental to both the chines. A shorter way to say that is that one edge of the curved panel must be a straight line.

    For many hull shapes the above is not a limitation, you may well be able to achieve the requirement by generating the hull as a loft with the transom end sketch drawn with straight lines and the next cross section in the loft also drawn with straight lines, these being parallel to the corresponding lines in the transom sketch. Alternatively, you might be able to identify some section part way along the hull where you can do similar, then loft the two halves of the hull separately and join up the pairs of flattened panels later (not something I have tried though).

    The picture below shows a little row boat for which I developed flat panels for the chined hull using only the standard sheet metal flatten in SW. In this case I initially drew the transom with straight lines as described above, producing a lofted hull shape with a planar transom. However, I actually wanted to have a pretty curved and sloped transom, so after doing the loft to make the basic hull shape I made a curved surface to cut the end off the hull on a slant to produce the transom I liked the look of. Rather to my surprise, all the panels in the hull still flattened by pretending they were sheet metal! The second picture shows one of the panels peeling away from the hull to make a flat part (purple). The third picture shows all the main hull panels (but not the deck) flattened and converted to a Solidworks 2D drawing file. I actually printed this on a sheet of stiff paper, cut out the panels with scissors and made a little model - all fitted perfectly as far as I could tell at model scale. This is yet another boat I have drawn but not made, perhaps this one might actually get made someday, its not such a big project and we would love to have a lightweight row boat to use from our local beach when the sea is not too rough.

    If you get really stuck with flattening surfaces I think there is a way to do it that must work, although I haven't tried it. Take two points fairly close together on one chine and a third nearby point on a second chine. Make a plane through the three points then intersect this plane with the second chine to get four points all lying in a plane, 2 on one chine and 2 on the adjacent chine, then make a planar quadrilateral surface from these four points. Now pick another point a little way further along the first chine, use this to make another plane from a triangle, then another planar quadrilateral and so on all the way along the hull panel. Then plot all the quadrilaterals joined together as a 2D plot - tedious but it should give a close enough approximation for developable panels, even those that are twisted along the length. I have seen it suggested that something similar could be done with a series of triangles rather than with planar quadrilaterals but that wont work accurately when there is twist in the panels.

    I note that you use both Solidworks and Rhino. I have never tried Rhino but I have seen plenty of impressive models made with it. My understanding is that Rhino has limited parametric capability compared to SW, is that so? To my mind the parametric capability of programs like SW and Inventor, also I think Catia and Siemens NX, is invaluable and I would not want to go back to software that does not have this capability. I used to use Inventor quite a lot, although not for drawing boats. Inventor is also good software and I expect their are other good options too.
     

    Attached Files:

  10. SukiSolo
    Joined: Dec 2012
    Posts: 1,269
    Likes: 27, Points: 0, Legacy Rep: 271
    Location: Hampshire UK

    SukiSolo Senior Member

    Hopefully the image below shows something like a rectangular block similar in size to that suggested (40 X 20mm) swept into a stringer shape. Note it twists correctly using the inherent normals from the chine curve. If I have interpreted the original request, this should be close to what is desired.

    John, you are quite correct, there is limited parametric ability in Rhino, it uses a Record History device which allows you to say change a loft curve section (or two) and reloft automatically. Fairly useful time saver but not truly parametric. It was new in V 4. Maya has or had much more of a parametric capability for a surface modeller.
     

    Attached Files:

  11. DCockey
    Joined: Oct 2009
    Posts: 5,229
    Likes: 634, Points: 113, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    Rhino also has Grasshopper, which is a parametric system.
     
  12. SLM
    Joined: May 2014
    Posts: 16
    Likes: 0, Points: 1, Legacy Rep: 10
    Location: Sydney, Australia

    SLM Junior Member

    SukiSolo:

    I tried as you have suggested. The hull is a round bilge design from Freeship (one of the default hulls when generating a new model).

    (a) On the center plane I sketched a stringer profile.
    (b) I extruded this to create a surface that passed through the hull surface.
    (c) I then found the intersection line between this plane and the hull (sweep path)
    (d) At the transom I generated a plane normal of the line of intersection.
    (e) Sketched the rectangular stringer on this plane so that it was tangent to the hull profile at the intersection (sweep) line.
    (f) Generated the stringer by sweeping the rectangle along the intersection line.

    The face of the resulting string is not tangent to the surface. See attached screen grabs. Close enough would be OK but this is a fair way out. It would be OK for a rough model but not good enough for notching frames.

    I may have to try the loft approach but this is a lot more work and then I have the issue that the lofted shape will not necessarily follow the hull form between frames were the stringer cross sections are defined.

    I also need to trial various options for the chine stringers (yes they are chine logs).
     

    Attached Files:

  13. SLM
    Joined: May 2014
    Posts: 16
    Likes: 0, Points: 1, Legacy Rep: 10
    Location: Sydney, Australia

    SLM Junior Member

    John:

    Thanks again for the long reply. Your pictures show a nice model and design.

    A lot of ideas there on developing the plywood skins. It is going to take me a few days to work my way through them all and test them on my model. I have a fairly good grasp of the maths and underlying theory of generating developments. Like you, I don't understand why Solidworks have provided such a half baked solution. The panels in my test model have zero gaussian curvature in Freeship so they should be developable but Solidworks refuses. Even if they were not quite developable, most software these days flattens them anyway and then indicates areas where the skin has been shrunk / stretched so you can work on the shape or otherwise decide that it is close enough and you can force it into shape (with plywood anyway - perhaps not sheet metal). Like you, I have found Solidworks to be a bit cantankerous - sometimes it works and other times not.
     
  14. SukiSolo
    Joined: Dec 2012
    Posts: 1,269
    Likes: 27, Points: 0, Legacy Rep: 271
    Location: Hampshire UK

    SukiSolo Senior Member

    Thanks for the images. I may have got the wrong end of the stick, I thought you were devloping a multi chine (with ply panels) boat not a strip plank one.
    I've not tried developing a strip planker - yet. Seems like an awful lot of lofts to do, with each one requiring a guide curve. I say lofts because they will taper in width to get the hull skin, and possibly even then require 'let ins' to get it to go together. If I was using Rhino I would probably think about using project curve to surface to generate the desired surface spline curves, although they would have to be simplified to get a really good result. It would unfold each strip OK, into a 2D profile. Other ways would be using a Polar array of curves to generate surfaces to 'chop' up the skin of the hull.

    The image I posted earlier was one off a multichine hull and is of a rectangular section (stringer or chine log) that could be trimmed by either the outer or inner hull skin. I extended the desired chine curve, created a line perpendicular to it and workplane (Sketch plane), drew rectangle (on that workplane) and used Sweep 1 Rail with rebuild within tolerance option in Rhino. It is a perfectly useable surface for other operations.

    I am not sure how well SolidWorks interfaces with Freeship. I have had quite a few tolerance issues on hulls for example, from Prolines using Iges V4.x as the translator. Rhino surfaces are good into SW and they remain parametric operations so you can change the input file and subsequent operations still remain valid. May require minor rebuilding but generally pretty good, I have built a few like that, not hulls though. I have had surfaces that needed rebuilding to be fair, up to 1.5mm out between port and starboard sheerline. You may need to rebuild your Freeship surface or edit it, or use as a guide only to create a new surface.

    Only other comment would be, how tight is the tesselation for the graphic render of the stringer? Certainly in Rhino you can set fast render and it will give angular graphics or slow render with more tesselations which gives lovely smooth rendering but a bit slower to regenerate. I often use custom meshes to ensure an accurate render, especially for the surface analysis tools.
     

  15. SLM
    Joined: May 2014
    Posts: 16
    Likes: 0, Points: 1, Legacy Rep: 10
    Location: Sydney, Australia

    SLM Junior Member

    SukiSolo:

    Sorry for the confusion. Yes I am designing with a multichine hull. However in attempting to prove that the swept cross section does not necessarily stay tangent to the surface I resorted to a round bilge hull not thinking about the confusion this might cause. I get the same thing happening when I sweep along a chine but it was not so obvious in the graphics as on the rounded hull shape so I posted the pictures of that. You are right ... perhaps the results will depend on how the line on the surface of the hull has been generated. I need to experiment some more. Time consuming stuff this experimentation.

    Solidworks does do a reasonable job of importing the IGES geometry that has been exported from Freeship. However I generate a new surface from this in Solidworks to work with.

    Unfortunately it is necessary to use software such as Freeship (and I may buy Delftship if I can get all this working) for the initial design of the hull to facilitate all the hydrodynamics calculations and allow the hull to be rapidly adjusted to meet the hydrodynamic requirements. If this were a composite hull or even a strip planked hull there would be no issue. The export to Solidworks would be painless and the detailed design stage of the project would carry on seamlessly in Solidworks. Unfortunately the same is not true of a multichine hull with chine logs and stringers. I have a lot of work to do to develop workable modelling techniques.

    If modelling a strip planked hull I would not try to model the individual strips.
     
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.