Resistance Calcs Results from bow-right vs bow left...

This question may stir some indignation or excoriation, but I'm willing to stand by for heavy rolls.


How many of you feel that software resistance calcs engines should cater to the user's model orientation rather than the programmer's following of drawing lay down tradition?

I understant that in some programs, putting the bow to left renders all the calcs (at least resistance/flows) meaningless. Is that a correct understanding? Is this true in mainly older code bases (say, s/w written over 10 years ago)?

What amount work (or, tradition-rooted resistance) would need to be overcome for ALL devs in Nav Arch to ensure their products can actually and correctly generate useful values? Is it really tradition that is the obstacle? Or, is it that it's too expensive or under-demanded that the resistance calcs code be more flexible?

(An idea: suppose a vessel is in mixed seas/currents, or someone wants to calculate the power needed to avoid swirl effects/suction toward a bank or turning out of a sudden whirlpool (rare, and likely not something to worry much about, other than in Tsunamis (Sendai) or salt drilling site collapes (Lousiana, IIRC, in the 70's or 60's, or sailing too close to the edge of the Niagara Falls (anyone know the largest craft or vessel to ever go over the edge? Is it even deep enough for a large barge or ferry to go over ungrounded? I get vertigo and seasick feelings just seeing footage of rubber-ballers and kayakers treading close to or going over waterfalls...)


Also, how much effort is it really to have flows/resistance calcs start work from the keel being baseline rather than the bulb/dome/other appendages? Imagine finalizing (or nearly finalizing a hull, or working with an existing drawing) and having to work with two coordinates systems, especially when the optimal bulb is yet again changed. This seems to imply a very intricately-set up spreadsheet or database has to exist first with various offsets/correction factors input when the bulb start, top, end, and bottom move around to suit a different set of requirements.

Has this been a controversial point in a thesis? Would it "rock the boat" (be career-limiting for a writer/student) if a thesis strenuously called for more flexibility on the two items I bring up?

To be fair, I did take drafting (ages ago), and have used engineers' scales and tools, but aside from it being practical to have 6 scales on a 3-sided tool, I found (and still find it) frustrating (as in "crippling") and sometimes stressful (to put it lightly) to rotate my head 45 deg or turn papers over, or flip the device to suit my mind's eye, but then crane my neck over to read tick marks, subject to lighting condition. But, even with some tape measures, when the only way to measure is from the right of a feature or a room to the left (and, that is not even a Naval Architecture area), it can be frustrating. One Naval Architect I talked with about my frustration sort of agreed that the tradition shouldn't impact programming and that many younger people with too much L-R and not enough R-L measuring might feel more comfortable.

Also, in 3D, some features are really looked at from their prominent features, not through an invisible plane or such, and compartments and bulkhead-mounted items portrayed in a key plan sometimes must be shown bow-left.

I suppose that it probably shouldn't matter since if one really must have the bow aesthetically or for personal trademark reasons on the left, nothing prevents it but the engineer or hobbyist just has to "get over it".


What say ye?

 

There is no reason for the bow to be to the left or right. For that matter, it can be up or down. Many designers drew boats both ways.

 

For obvious reasons a program deriving information from a 3D model needs to know which end is the bow but other than that as far as I know the bow to the right "standard" is just tradition. As Gonzo said boats have been drawn both ways for ever. If anyone has a logical reason to draw them one way or the other it would be interesting to hear.

 

First off:
With the help of a little (or a lot of) software programming it is possible to draw the ship any which way you want: Multiple coordinate systems, upside down, mirrored, etc. Just as long as the software is able to recognise the ship and place the final set of coordinates (used for calculations) correctly to derive some vital coefficients!

The problem and possibly the answer to your question, lies in these vital coefficients! Currently the majority of resistance calculations, available through design software, are based on regression analyses. The regression analysis tries to predict the resistance of your vessel based on known resistance values of similar vessels. In order to correctly predict the resistance of your vessel, your vessel must be (very) similar to the set of vessels on which the regression calculation is based!

The coefficients used in the regression calculation are an attempt to describe the ships 3D shape in a set of simple and single values. Some well known coefficients are: block coefficient, prismatic coefficient, L/B; L/T, etc. However as you may understand, it is not possible to accurately capture the 3D shape of a vessel in a set coefficients. If we look at only the four mentioned coefficients: Cb, Cp, L/B and L/T. You can imagine that it is very easy to design two vastly different ships with exactly the same four coefficients Cb, Cp, L/B and L/T!

To arrive back at your question:
1) The (final) coordinate system used to calculate the resistance must be the same as used to describe all the vessels used in the regression calculation. This coordinate system can differ from that which is used to draw the ship.
2) It is not possible to calculate the reverse/sideway resistance by drawing the ship the other way round, as the ship will then no longer fit the regression parameters!

There are software programs available (for instance cfd based) that are able to predict resistance based on flow, and real 3D hull shapes. These are programs should be able to predict some form of reverse or sideway resistance, under the condition that the shape and flow conditions fall within the programming limits. Note that in the case of CFD calculations the programming limits are based on known flow variables/parameters etc. As of yet these flow variables and parameters still do not cover the real world flow parameters!

Hope this helps,

Tackwise

 

What about the full CFD analysis tools like Ansys, CFX, FLUENT, etc etc...

They take a hell of alot of resources to run the simulation so generally clients rent time on a supercomputer server to complete the analysis.

As far as i know, this is the most up to date and accurate method of predicting anything to do with fluid dynamics, im pretty sure US Navy uses it - not just for their ships, but also missiles, projectiles, aircraft and so on...

It may be unaffordable for weekend boat designers, certainly not for a professional NA, but if your looking for a better method then look no further... you will most likely need another university degree to use it effectively however, its a very complex line of work...

You can run a dynamic simulation of any shape object in a body of fluid, so for example, a ship travelling thru the ocean complete with wind and waves thru a specified timeframe, 1 minute would be about all youd need. You will get data and results for absolutely everything right down to how much extra drag that radio antenna created on the bridge, the temperature change on the hull surface from the water friction vs speed, accurately model the spray coming off the bow and a whole lot more, quite remarkable actually... Another use for it is to model a fluid thru a fixed object, such as air flowing thru an engine manifold etc... I managed to get a hold of ANSYS software and had a play around with it last year, spent a good few weeks trying to learn it but it was far too complicated to simply learn it quickly, really need to study it long term... Better idea is to simply send the 3D CAD drawing file in *.iges format to a professional in CFD analysis and pay them accordingly...

 

Actually the coordinate systems used don't matter at all for regression calculations as long as the coefficiencts and dimensions are calculated consistently.

 

Which is where the coordinate system matters again. Think of software which, for example, calculates the half-angle of entrance or the 1/4 beam buttock angle from the 3D model of the boat. These two angles would be very different if the hull is pointing backwards relative to what the programmer has intended.
Cheers

 

Nice!

Thanks for the various responses. This now is making me recall a student or research paper covering air wake around a naval destroyer (USN) and the impact(s) on various flight envelopes of various helicopters.

As for resistance calcs, I partly went off on my original questioning because I (probably) misunderstood in another forum that Delftship may or may not correctly interpret resistance if a hull is rotate 180 degrees. Since I'm using Hydronship (Freeship 3.26+ test), I began to wonder about those. I'm not an NA, and the values I have are mainly just to help me fair the hull properly at various stations, not that I could produce a model of a superior hull at a given LB/BT, etc.


(An aside:

BTW, I exported my fbm file (from Freeship 3.26) as dxf (which I temporarily imported to my install of Delftship Pro to use the "automatic" fairing (which, in the process re-jiggered my layers moving my file to and fro...). I THOUGHT I did a pretty good job at hand-fairing my hull, but still had a little more work to do when it got back into Freeship 3.24 (at the time).

I imported it into PolyCAD. Once I applied bsplines of least squares and entered edit mode, I found that I was able to manually the hull even further. It was nice to find out WHY my hull was so difficult to thicken in Punch! ViaCAD Pro. The stations were STILL horridly sending out crystilline-structure lattice-like patterns around the points, some seemingly going out to infinity, but most within a 100 or 200 meters of the model. using the bspline curves of least squares was a near-hidden gem. Now, when I export to DXF and import into ViaCAD and convert the meshes to surfaces, I for the most part can thicken many more areas of hull plating than previously if I had not gone into PolyCAD.)

 

In your example the input of the geometry in accordance with the coordinate system the software uses for the calculation of the coefficients is what is important to ensure the coefficients are properly calculated. That coordinate system is tied to the software, not the regression formula itself.

Other software may use a different coordinate system. If the coefficients are calculated using that software then the geometry must be in accordance with the coordinate system of the that software.

What matters for the regression formula is the the half-angle of entrance or the 1/4 beam buttock angle, not the coordinate system used to calculate them.

That statement directly tieing regression formula results to coordiate system is fundamentally wrong unless it is limited to calculations where the geometry is input into an integrated software package and the resistance results are output. Perhaps the use of integrated software has become so common that it is assumed to be the only way to do such resistance calculations.

But it is quite possible to calculate the required coefficients separately and then enter the coefficients into the regression formula(es). In such a case the coordinate system used to calculate the coefficients doesn't matter as long as the coefficients are determined consistent with the definitions used for the regression formula(es).

 

You are saying the same thing I have said by using different words. I was simply pointing out the necessity to take care of coordinate-system conventions used by programmers of the software. It was related to software like Delftship, Freeship etc. which are created to make your life easier by calculating automatically the necessary coefficients and angles. You just have to be careful and use the same conventions which the software developer has used (and to check the final results).

Agreed.

Cheers

 

I see I must weight my words more carefully....
I do sincerely apologise for the confusion!

You are of course correct: The regression calculations do not use a fixed coordinate system but merely the values of the coefficients, which must be consistently determined. I was merely trying to explain in a few words what daiquiri has correctly stated:

At least in the end we all agree!