Hull Design Software Based on NURBS

I have the following thoughts about using hull design software and would appreciate comments, suggestions or perhaps corrections from real designers.

Hull designers working with pencil and paper have in the past used battens and spline-weights or french curves for fairing purposes, and many still do; no doubt those that work in software still use pencil and paper for preliminary sketches. My concern is, how close to their design objective can they get in software. Let me explain.

I use FreeShip and Carenne. Carenne is limited but easy, FreeShip (also its cousin DelfShip) is powerful but somewhat counter-intuitive in use. Freeship create surfaces using NURBS (non-uniform rational B-spline). The math of NURBS is awesome and beyond the perview of this discussion, but briefly a number of control points controls a 3-D surface which presumably can be brought close to a specified surface given sufficient control points.

For new hull design purposes when starting the design in FreeShip, using the minimum number of points necessary yields a quick and easy solution and a fair hull shape, but not necessarily close enough to what is desired. Sometimes I have specific ideas on the shape of a new hull design, for one design I wanted hydrostatic and other data for a hull that would use plank developments mostly based on straight lines. Getting the precise shape into FreeShip was challenging and the results were never quite correct.

To obtain a precise design it seems time must be invested by calculating or drawing lines to generate accurate offsets, that can then be uploaded to the program to generate the hull shape required. Of course one then has access to all that computational power, but it creates an obscene number of control points and if a change is required, it's literally back to the drawing board.

Another challenge is when an old design needs to be uploaded into software, perhaps to generate a modification. If accurate offsets are unavailable a considerable part of the design cycle has to be undertaken to get them, and once uploaded into FreeShip anything but a simple parametric change involves tussling with an unwieldy number of control points, none of which do what is needed on its own.

 

You have come across the perennial problem that plaques professional NA/designers as opposed to those that dabble for fun or their own small design.

When designing a hull shape with a set/known in advance set of parameters, from sketched or wherever, to replicate this in a software program is very challenging. Amateurs do not come across this simply because their "parameters" is a simple goal, get a shape and faired for build.

For the past 20years we've been having this same debate, the software is driving the hull shape, not the NA. The "small" things that one wishes to have in the design, which can easily be shown in paper/sketch format, often takes endless hours to produce one single simple parameter, or often not at all. Thus ending up accepting a compromise.

There are 1 or 2 hull generation programs out there now that are amazing, Paramarine for example, but to get the 100% exact shape desired, from 1st principal calculations/sketches...requires significant software and computing power that is beyond the "day tripper".

Some software, well known, can't even produce a straight line!...may "look" straight, but is in fact a slight curve, no matter how slight. This, even if less than the thickness of the hull material, can still cause problems for build.

 

Thanks for the answer John.

- I've certainly noticed the problem, although I merely "dabble for fun" myself, and the phrase "software driving the hull shape" is also familiar. It hasn't been much of a problem for me so far. I often create a math model or use Presentations (similar to PowerPoint) to refine my requirements, or I can start the design off in Carenne then export the offsets to FreeShip where I can fix whatever Carenne cannot do.

In the case below I wished to model a stripper canoe with each strip shown so I could confirm that the planking lines would look right after stripping up from a bottom plank Adirondack Guideboat style, and the strips would all have the same developed width to simplify the build. Virtually impossible with NURBS. In Carenne2008 I was able to create all 15 planks very quickly with the precise bilge shape. Carenne insists on inserting a transom but that was no problem since my design was symmetrical and it got the bow right. I just exported the offsets from Carenne and mirror-imaged the bow data, then imported the result into FreeShip and bingo, I had my developments and pretty pictures! Maybe 15 minutes from start to finish, could be a record. However tricks like that aren't usually possible. Same problem arises when designing a lapstrake hull, where the planking lines are crucial to the appearance of the boat.

You mentioned Paramarine as an advanced program, one of a small number. Does it overcome the issue I described? It would be nice to know that professionals at least have access to tools free of these irritating issues.

 

Terry, there in is the problem. You can always find one software that does XX very well, but not YY and so on. We used to have 3 different software's depending upon the hull type and curvature we wanted. Tedious. None of them were exemplary at all.

I can't recall. I used to know how to use if fairly well, as I have the program. But I haven't used it in anger as such for about 5years...i simply don't have time to melded with nuances in lines any more. I just scribble down on paper the basic parameters and angles/shape of lines, scan it, and send it to a friend who does this 24/7 to draw up for me. Leaves me free to do other more important work..

 

Well said Ad Hoc. In other design fields the same problem exists and I have always been of the opinion that the design is more important than the software. At least the last twenty years have allowed 3D design on computers to become widespread and affordable. However the aim of good design is to deliver that, and minimise any compromise from software ability, compatibility etc. Not forgetting that some clients insist you use the same software as they have in house which is a very limiting factor......

Using a high end parametric package on one project we were able to do something similar to the modification of a single design into 5 deliberate iterations. It was not easy but by establishing enough rules in the software, the desired outcome was achieved. The price of that software is beyond the amateur or small shop.

There are other ways of getting complexities into 3D formats from scale models etc but at least if you don't forget the basics of 2D drafting you are in a better place. Certainly, it has been a while since I put pencil to tracing or film but can still do it....

 

Hi Ad Hoc,

I am replying to your:

"You have come across the perennial problem that plaques professional NA/designers as opposed to those that dabble for fun or their own small design."

I by no means am any NA. But, I'm doing more than dabbling. A few years ago, I embarked upon a needlessly complicated path of hull design when I started with Deltfship and later Freeship Plus/Hydronship. My original method was shaping the hull in DS/FS-HS and spending countless hours nudging control points along. Once I needed naval combatant-relevant hydros and powering info - something I was never going to get out of Delftship - I switched to FS/HS and tried to continue on. I would export the curves and meshes, and spent countless hours in 3D CAD trying to make order out of chaos where facets would not line up correctly in CAD, and spent inordinate amounts of time stitching bad meshes as well as using the FS-HS curves as a basis for recreating the sideshell.

I then dabbbled with PolyCAD, and learned a number of things, but found I could not operate as smoothly as in CAD. Still, I provided some feedback to Marcus B. I also provided feedback to Victor T., and both kindly incorporated a few things here and there, or fixed bugs, or enlightement me about things. It was I guess positive and productive feedback looping for all involved.

Now, my technique is to pretend that in 3D CAD I am selecting a "proven" or "parent" model and its hull-making splines. One of my Freeship Plus models that produced hydros I really liked became the "parent" model and I traced the hull with splines, faired them, and then created a surface of the halfbreadth. I even managed to get the DDG's sonar dome bow created with the 12 or so hull-shaping splines. To get the transom, I just mirror the opposite side of the ship and then create a surface between the transom edges. If I want some special shaping, I use curves to control two surfaces and use those surfaces to "split" associative material from the remaining transom. This also works for dealing with a "stepped" hull/deck. If the step and transom angles need adjusting, just modify the cutting surfaces.

So, I've been "plagued", too.

If I had listened to some and done better experimenting, I would have saved myself a couple or 3 years forging ahead on my own path. Now, I can take one good hull, and then for expediency toss the hull and modify the curves. I toss the hull because the associative surface is CPU intensive to adjust, and speed-editing is torturously slow.

Next, I export that hull from CAD to a VRML model. Import that VRML model into Freeship Plus and use the Hydronship reports and calcs to influence re-editing the CAD model. This now reduces my efforts to a few hours instead of weeks of tweaking and exploring. It is much more straightforward (for me) to create the hull in CAD and import it into FSP-HS than to use FSPs GUI to edit curves. I honestly have to thank Marcus Bole because by exploring PolyCAD, I realized that I could to some extent mimic PolyCAD in Punch! ViaCAD and Punch! Shark. These baby steps even helped me in 2012 create a vastly better propeller of 5 blades and hub than my horrid attempts in 2010. I even can export the props and shafts as VRML and bring them into Freeship Plust just to obtain better presentation than using a mere disc to represent a propeller. I also bring in fins, stabilizers, and ruddders this way, too. And, no I can even get hydros on my models that have the 2nd, UK/ROK-style triangulation-type second sonar abaft the main head. Trying to in FSP create the 2nd sonar (even the main sonar) was and excruciatingly painful process for me, given my informal training in drafting. My mind could see clearly what I needed/wanted, but FSP was too intricate a dance. PolyCAD and surface modeling in it cleared my mind. Learning the CAD app more also helped me a lot, too. Now, I can create the fuel and ballast tanks in CAD and get them into Freeship when I chaffed that I could not afford as a hobbyist to get the pipe and tank-making professional add-on for Delftship.

Something else that ate up about 2 years of my time was my effort to make Freeship Plus give me hydros and powering information on my hull with decks, platforms, superstructure, keel fins, stabilizer fins, rudders, and sonar dome, and one misplaced control point/node would wreck entry angle information or KM or other info. I nowadays have almost zero problems now in FSP/HS when I create the hull in CAD first, then clean it up using finer-precision CAD, then importing its VRML counterpart into FSP.

All this has led me to thinking that if I had deep-pocketed funding, I could pay/hire someone to create *my* ultimate ship design tool based on my head-banging work-arounds in 3D CAD modeling. It would be a combination of database, CAD, and Freeship, taking some cues from PolyCAD, and possibly Cafe, tweaked by some leaps in my imagination, exploiting BOM tables and Shark's/ViaCAD's solid modeling. Having played with around 6 different ship design apps and perused tutorials of others', I think I have some ideas that are unlike any of the others. Too bad for me I don't have money to pay a private team so that its public disclosure would be on my timetable, not that of another party.

Unfortunately, Shark/VCP have some curves/entities issues that I need resolved, and I will get the developer to look at (but not necessarily address) my observation of something I won't go into detail here. Once that issue is nixed, I can resume working on my tutorials to help others reproduce my methods and then improve as they see fit. I've basically set upon a fast-track/ad hoc, solid-modeling hull modeling prototyping tool that is not ShipConstructor nor AutoShip. But, I think I've tweaked my approach to aim for making the ID (Initial Design) stage a lot more than just a sketch-stage model, but not as simplistic as SketchUp.

I'm aiming for "unplauging" hull design, but I will not achieve much unless I give up more info and distinguish my approach enough to convince someone to take it to a development level.

 

John you noted that you now send parameters for a friend to draw up; such parametric design is by its very nature open to interpretation, so you must rely on your friend to understand your views and preferences I guess. Would it be correct to say that, in essence, you hand off the details of the lines for him to add? Of course as a NA you still need significant control over the construction details.

SukiSolo, the cost of high end software and the hardware needed to run it is the problem for many of us, as is the learning time.

CD: you've hit the problem we all have when there are excessive control points by the sound of it. As you noted, when there is a "parent" design that requires minimal changes which can be done parametrically life gets a great deal easier in FreeShip and related programs. I am not familiar with VRML - I assume you don't work directly into the VRML file but just use the file format as a tranfer vehicle between programs. Your approach is similar to my less sophisticated use of Carenne and FreeShip with an offset file used to transfer data. Some of your discussion went over my head, but I'm with you on de-plaguing hull design software.

I'm starting to understand why a number of NA's to this day do much of their work on a drawing board rather than a computer.

 

I wonder if some of the problems could be avoided by using more general purpose CAD software rather than software specifically intended for hull design. This may seem an odd thing to say, but if a CAD program is capable of very precisely defining a complex curvy shape such as a car body shell or suchlike, why should it not define a somewhat simpler shape such as a boat hull?

Being employed as an engineer in my last job I was not supposed to use CAD during working hours - only draughtsmen could do that(!), but having a bit of an interest in the subject I did get to play with CAD programs during the lunch hours, especially Autodesk Inventor. Since taking early retirement I have also played with Solidworks as one of my favorite computer games. I have drawn boat hulls with both Solidworks and Inventor and then continued to draw the internal structure and do FEA structural analysis all within the same suite of software. 'Plug ins' are available to extend the capabilities of these programs for drawing really intricate curved surfaces but so far I have managed with just the standard features of the programs. I haven't yet tried Freeship or Careene, but I do wonder whether these days this kind of specialist program is being superseded by more general purpose design software.

 

John: I've seen those car body (and aircraft) design programs and they are indeed awesome. However, I wonder if they have the same problem as the hull programs do, where "the software is driving the hull shape" in John's words of post #2. It sounds like some of our respondents want to get away from that.

I have over 120 study plans and have been looking at them lately, attempting to classify the geometric curves and the drawing tools that designers actually use. I've only got as far as the plan-view sheerline at this point, but they seem to fall into a quite small number of categories:



Simple circular arc. Often found in canoes.

Curve defined by batten and three spline weights. This would be a conic of course and seems common in kayaks.

Curve defined batten and four spline weights. This would be a circular arc connected to the stems or transoms by conics. Often found in sailboat territory and I've also spotted it in rowing shells.

Other combinations of conics and arcs: for example a number of powerboats seem to have an arc from the stem followed by a conic, perhaps with a change to another conic before the transom.



I would need more precision before pronouncing a theory and I need to look at the other views. I will take a look at the few larger scale drawings and plans that I have. It does seem like an interesting theory, and if it holds up, all of the above can be expressed by a simple computer program. I actually have an old spreadsheet which generate the first three cases above which I did as part of a canoe design exercise (yes - it got built) a few years ago.



This may be a basis for a new design tool, a rather simple one I suspect, that would define hull shape in a manner mimicking the use of the traditional tools of the design - compass, battern and spline weights - and generate the offsets to be imported into the existing heavy-duty programs to do the crunch work of hydro-statics/dynamics, plank development and so forth. I rather like the idea of a virtual batten . . .

 

No, and that's the point. I provide guidance/instructions what is required. It is then up to him how he achieves it, in what ever software program he wishes to use, but because he uses his program of choice 24/7 he is very experienced with its nuances can get pretty close to what's needed. The time he takes is significantly less than I ever could do the work in..thus why waste my time when I can get someone else to do this for me, quicker and better than I can. It is all about time/resource management to get the job done.

Remember, I'm not dabbling away..i have a contract and a deadline to meet. The only deadline many dabblers with such software have is can they finish learning page 5 of the manual before going to the pub or dinner. Thus their amount of time spent playing is disproportionate and at variance to the amount of time I have available to complete.

I should also point out..a hull lines...is just that, a set of lines. The nuances I describe, are for production/build purposes, not NA calculations per se. Since whether the lines are off by 10mm does not affect any of my naval architecture calculations one bit; I'm more than happy with such accuracy. However, it makes a significant difference to the fabricator and welder when the plates don't meet up. Thus the production faired lines requires significant skill, for production accuracy, far greater than I can afford as there are many other 'design' function that must run in parallel. Hence, a totally different set of requirements from the dabblers and their expectations of what they wish to achieve and their time scales.

This is where the true skill of those old boy Loft's man come into their own. Their trained and expert eye could pick up a wonky line from a thousand yards in the dark, software simply can't do that! And this very very valuable skill/function has been completely and utterly over looked by all dabblers/day tripper with hull lines programs, because it has pretty colour plots or their needs are just simply very different. All because what we professional NAs require from a set of lines is very different, in terms of accuracy and times scales, than that required by an amateur.

There some very goods ones. But they all have the same issues. The only ones which I know of which is very powerful and falls into the wow factor, but mega ££££/$$$/€€€ is Siemens' NX.

 

There's a new 3D CAD engine being developed by the guys that did the original Pro/E engine and then went on to do Solidworks. The idea is to combine the flexibility of NURBS with Solids modeling right out of the box instead of doing patching. I know the guys doing this - a smart group of folks

http://www.onshape.com/

 

John, OK, I understand, you two work the way I used to with the design department of my former employers, they do the drawings and the boat gets built from those but you do the design. Been there - different technology. Gotcha!

I have to say that FreeShip is capable of producing wonky lines with the best of them, but I didn't take long to realise there were precautions to be taken that avoided that situation. Those precautions limit what I can achieve in a design, and although it hasn't been a problem in the past, with what I am trying to achieve these days it will likely be one.

 

Yes and no

If the project is small enough, I do the lot; I just don't have time to waste getting the difference from 10mm down to 0.1mm accuracy for the hull lines that is required for production. If its too large then the amount of drawing work I actually do reduces accordingly, but I still control/do the principal dwgs regardless how much I allow the DO to do.

 

Dabbler/Daytripper

Hi,

I snipped out a chuck for readability and hope that I have not created a misquoting situation.

I am sort of a "daytripper/dabble", one who definitely is not a naval architect, but also definitely not one who toils away trying to impress people in ShipBucket.

I am not much of a drafter, but when I use Shark and ViaCAD for general arrangements and hull shaping, I am trying to keep DFP (Design for Production) in mind. It drives me nuts when after I have the CAD app thicken the hull sideshell/surface the sideshell plating at top deck surfacing is inclined according to the curve and apparently has no ability to keep the top deck hull area parallel to baseline. That has vexed me for years, and forces me to explore different ways to obtain sideshell that won't require me to engage in convoluted, tricky ways to do it with just curves and simple sweeps.

Also, akin to an NA, I use FreeShip Plus/Hydronship to obtain initial stability info, entry angle, Cb, Cp, and sundry other information to make sure my basic hull is not just within Victor's parameters of his library of information, but that my hull would fit a range of varioius nations' ideas of acceptability, not just my own internal muse. I also put in a lot of time looking at powering, turning prediction, acceleration, and cavitation risk based on draft setttings, displacement, and so on. It forces me to think like or consider only a fraction of what real NAs get paid to do based on their many years of learning, interning, working, and managing. I could never be a real NA, but I strive to be more than a mere dabbler/daytripper, too. I try to make sure that I keep to practicable clear deck heights, set up space for structural members and place proxy shapes of generators, switchboards, pumps, ladders, deck cutouts, and more in the model. I also put in shafts, hubs, and prop blades, line them up, and try to make sure the exit path is suitable to the hull, that the engines will fit in the hull on their foundations, and so on. I also have to run paths for air intake/GTM exhaust, plan the fueling piping, firemain piping, working walk paths and emergency paths, airlocks, CHT, ventilation, and myriad other things.

Most daytrippers, especially from what I've seen on ShipBucket, do not come anywhere near that level of consideration, except maybe those exploring and using the program to design WWII era warships and toil over the number of boilers, cannons, and such, or the armor plating on the hull.

BTW, in Japan in 2004/2005, people who saw my (poor line quality, hand drawn GAs of gas turbine-powered combatants/destroyers) and INSISTED that I was lying to them that I have no college degree and that I am not at all a Naval Architect. It defied their logic (themselves multi-degree holders) that I could in that much detail design a destroyer. I later introduced myself to Mitsubishi, got invited into a room with 3 officials there, and then told "Nashunnal Seeekrets", with the senior of the 3 blocking his eyes with his hands, and turning away from my drawings. Being polite by culture, in person, they did not summarily eject me, but the next-senior official said, "I recommend you do not display your "art" in Japan again." The third, most junior, gave me his business card recommended that I pay a visit to Funenokagakukkan. I did, and because of him, I shot maybe 200 photos in what I consider to be the best maritime museum of old and contemporary human maritime and naval history on Earth.

Unfortunately for me, I have a concentration/focus issue and make no headway in turning this stuff into a profit vehicle for myself, and to spin off other activities that could be useful in the entertainment world. I have no illusions about being any honorary NA, but I won't stop at the cartoon stage, either. This is too compelling, fun, and interesting stuff, and I am always in awe of anyone who can endure all the math and mental processing it takes to conceive of, build, and manage dynamic structures such as all the ships of steel.

 

What you're describing is ostensibly a draughting issue, not a design issue per se. As Terry notes above, from my comments, once the 'design' is done, it is then completed (where required) by the drawing office. The function of the DO is pretty much what you are describing..details, endless details.

Design and the point here, Lines generation, whilst can become "details", is ostensibly a single design function. That being to create a 3D shape/hull, to a set of given parameters. Anything else..is simply that..something else, from a NA point of view.

Stability etc...i don't use any "freeware" at all, it is all commercially robust software. It may not have all the bells and whistles that many wish to see, but it does the job and very well. Powering etc...one doesn't need a perfect 3D lines plan for estimating powering...just experience!