Calculation of wave resistance and potential flow.

[yipster]

This is a visualization of a wave pattern of a catamaran. The free surface mesh has a specific structure, in order to follow the flow between the hulls. It is also posible to measure the wave heights in this area in order to optimize the distance among them.
After defining four posible ship hull forms for a fishing boat, varying different aspects - with or without bulb, different sizes for bulbs, etc - we calculate wave generation using Tdynlin. This calculation provides flow speed and pressure distribution on the hull and free surface deformations.

this Computer Fluid Dynamics program i found in search of wave and flow resistance and has a demo. now see if maxsurf has that in hullspeed, something i should have done in the first place
anyway, its compatable.

yipster

[CGN]

What is the price for this software?

[Andrew Mason]

We don't have full potential flow calculations in Hullspeed, however we have recently added a slender body method to the program. This is conceptually similar to the method used in Michlet, although details of the implementation are different.

A slender body method used is based on the work of: Couser, P., Wellicome J.F. and Molland, A.F., “An improved method for the theoretical prediction of the wave resistance of transom-stern hulls using a slender body approach”. International Shipbuilding Progress, vol. 45, No. 444, 1998.

This method uses the Maxsurf hull surfaces directly, so a full surface model is required - it is not sufficient to work from simple hull parameters.

[Leo Lazauskas]

Thanks for providing some beautiful wave patterns! I have been wondering when someone else was going to bring up the fact that most of the low cost software being rated on this board does not have the capability to estimate wave patterns and resistance using physics-based methods.

A note of caution though. Firstly, the program uses only 5,000 panels which is far too few to capture the fine structure of the diverging wave pattern. You are more likely to need 50,000 which will increase the execution time considerably.

At the Gothenburg 2000 conference many codes took over 200 hours to calculate the wave patterns at a single speed for a monohull, and even then their results were generally only barely acceptable for hulls with transom sterns. Also troubling is that for some of the hulls examined at Gothenburg, they could not agree well on simple quantities such as the wetted surface area. Variations of at least 5% from the true wetted area were reported. For catamarans with one hull immersed more than the other, 5000 panels will only barely describe the two hull surfaces adequately.

A more worrisome aspect was pointed out by Prof. Lawrence J. Doctors and Dr. Alexander Day in "Non-linear Free-surface Effects on the Resistance and Squat of High-Speed Vessels with a Transom Stern", 24th Symposium on Naval Hydrodynamics, Fukuoka, Japan, 2002: "... A further point is that more sophisticated computer codes ...do not always lead to more accurate or reliable predictions for numerically sensitive quantities, such as resistance. This is because resistance can be affected markedly by minor inaccuracies in the computed pressure distribution over the surface of the hull. ... the linearized approach gave predictions which were within 5% for most of the test cases, while the errors from the competing non-linear method were typically an order of magnitude greater."

Sorry for the long quotes, but I would rather show criticisms other than my own for competing software.

Do you know what method is used to predict skin friction? This is very important because the "standard" ITTC57 line (which is not a physics-based method) does not do as good a job as Grigson's algorithm in most cases, and viscous resistance makes up the bulk of drag for many vessels.

All the best,
Leo Lazauskas.

[Leo Lazauskas]

Originally Posted by Andrew Mason

We don't have full potential flow calculations in Hullspeed, however we have recently added a slender body method to the program. This is conceptually similar to the method used in Michlet, although details of the implementation are different.

A slender body method used is based on the work of: Couser, P., Wellicome J.F. and Molland, A.F., “An improved method for the theoretical prediction of the wave resistance of transom-stern hulls using a slender body approach”. International Shipbuilding Progress, vol. 45, No. 444, 1998.

This method uses the Maxsurf hull surfaces directly, so a full surface model is required - it is not sufficient to work from simple hull parameters.

Slender body theory is, as you say, conceptually similar to Michell's thin-ship theory, however there are some important differences. Prof. E.O. Tuck (one of the first hydrodynamicists to derive the slender body theory) made among other interesting observations, the following distinction in "Wave Resistance of Thin Ships and Catamarans":

"There are elements in the Michell theory (such as exponential decay with depth) that are absent in the slender body theory. ...the evidence of comparison with experiment seems to favour Michell."

The implementations of the hydrodynamic aspects of Michlet and HullSpeed are certainly different. Among the most important are:

1. Michlet (and other programs I have written or co-written, such as Flotilla, Flotsm and SWPE) use transom stern models that are quite different to the one presented in the paper by Couser et al. The "backward-facing step" approach presented in their paper *was* incorporated into early versions of Michlet (at least for the wet portion of the transom) but has been very much enhanced since Michlet 5.0.

2. Michlet includes the effect of boundary layers on wave resistance and wave elevations.

3. Michlet implements an eddy kinematic viscosity model in wave pattern calculations.

Best regards,
Leo.

[tspeer]

Originally Posted by CGN

What is the price for this software?

It looks like you have to have GiD @ E970, then contact them for the price of the Tdynlin module. What you pay will probably depend on what your traffic will bear...

BTW, Dawson's method is a bit dated these days. There are more modern codes, like Splash. If you're looking for a commercial application, the cost might be well worth it. But for a hobbyist like me, I'd love to get hold of a Dawson code because it should be able to handle leeway and hydrofoils (like keels and rudders).

Originally Posted by tspeer

It looks like you have to have GiD @ E970, then contact them for the price of the Tdynlin module. What you pay will probably depend on what your traffic will bear...

BTW, Dawson's method is a bit dated these days. There are more modern codes, like Splash. If you're looking for a commercial application, the cost might be well worth it. But for a hobbyist like me, I'd love to get hold of a Dawson code because it should be able to handle leeway and hydrofoils (like keels and rudders).

I never see SPLASH validation with public release experiment for hulls. Say for wave drags or for total. SHIPFLOW does and many others. Why not SPLASH?

[burakreis]

here is some results for wigley etc. against shipflow;

http://www.panix.com/%7Ebrosen/catamarans.html

Originally Posted by burakreis

here is some results for wigley etc. against shipflow;

http://www.panix.com/%7Ebrosen/catamarans.html

Thank you for URL, but which is better program? These nice pictures do not show experiments!

[burakreis]

i can't say which is better unless to test them personally so have no exact idea about that matter.

by the way these potential flow codes (there are many more indeed) have limitations because of the lack of viscosity effects. if i remember right shipflow uses some rans calculations for the aft part of the ship and potential for the rest.

i think, for the future full rans-urans calculations (with les -large eddy simulation- turbulence model in near future and may be dns - direct numerical simulation- for the later) will be used widely. but with free surface effects these solutions are beyond the power of today's pc.

for my point of view the most interesting cfd code for ship design is comet;

http://www.iccm.de/

[Leo Lazauskas]

Originally Posted by burakreis

i can't say which is better unless to test them personally so have no exact idea about that matter.

by the way these potential flow codes (there are many more indeed) have limitations because of the lack of viscosity effects. if i remember right shipflow uses some rans calculations for the aft part of the ship and potential for the rest.

i think, for the future full rans-urans calculations (with les -large eddy simulation- turbulence model in near future and may be dns - direct numerical simulation- for the later) will be used widely. but with free surface effects these solutions are beyond the power of today's pc.

for my point of view the most interesting cfd code for ship design is comet;

http://www.iccm.de/

You might find some experimental results for the Wigley hull used in that comparison, but it won't be easy. I think some work was done at NPL by Dr. Millward who is now the the University of Liverpool. There have been many dozens of papers comparing predictions with experiments for Wigley hulls having L/B=10 and B/T=1.6, but the comparison of Splash and Shipflow uses L/B=10 and B/T=2.0.

In any case, I wouldn't put too much emphasis on comparisons with experiments when they are presented in publicity brochures. For example, some Splash "brochures" quote a statement made by the organisers of the famous "Wake-Off" where the performance of several ship hydrodynamics codes was compared:

"FLOPAN (Splash) gave the best predictions of all for the details of the free-surface disturbance in the region of one ship's beam from the model."

Splash publicity omits adverse criticism made in the same paragraph
such as:

"... FLOPAN (Splash) wave patterns died out much too rapidly."

Or

"At the extreme, FLOPAN (Splash) predictions showed no waves at the edge of the computational boundary."

In another thread, Tom Speer mentioned that he would like to get hold of a Dawson code. I'm not sure that a crude version of something like XYZFS will be particulary useful for prediction of resistance and wave patterns, although it might give rough estimates of leeway effects. XYZFS performance was considered at best marginal during the Wake-Off. On the other hand, enhanced Dawson codes such as Hoyte Raven's RAPID might be a better, albeit considerably more expensive, alternative.

Burakreis mentioned that the code COMET seemed interesting. At the recent Gothenburg 2000 Workshop on Numerical Ship Hydrodynamics many of the latest prediction methods were compared. COMET's peformance was at best marginal (as were most codes at that workshop). According to the organisers of the workshop":

"...COMET suffers from instabilities upstream of the Kelvin wedge."

If anyone is interested in looking more deeply at the performance of the latest hydrodynamic codes I highly recommend the Proceedings of the Gothenburg workshop. The following codes were examined:

CFDSHIP, CFX, COMET, FINFLO, FLOWPACK, FLUENT, HORUS, ICARE, MGSHIP,
NEPTUN, PARNASSOS, SURF, UNCLE, UVW, WAVIS.

Burakreis also mentioned that the execution time for some recent programs is likely to be very long. That is a bit of an understatement. According to the Gothenburg organisers:

"The total CPU time now varies between 2 hours and 240 hours with
an average of 78 hours".

That, by the way, is for a single hull, at a single speed!

In CFD, I guess you just have to pay your money and take your chances.

Leo Lazauskas.

[redcoopers]

I thought I should chime in on the remark about SPLASH. I have made an exhaustive study on the performance prediction on the new Naval Academy's 44' sloop. The new design by David Pedrick still retains most of the heavy disp/length ratio as of the current boats, but fairs the underwater body as well as adds a small bulb.

Using a 10:1 scale model, I completed ~ 500 tank tests in a 120' tank. These tests were made at all upright conditions as well as many different rudder, yaw, trim, and heel angles. (as well as multiple turbulence simulation trials)

I then used a code from Carderock NSWC, which is based on Michell's eqn. The upright conditions matched surprisingly well for a ship with such a large B/L ratio.

Finally, I used SPLASH. SPLASH was able to make good predictions both upright and in the heeled & yawed conditions.

If you would like to see some of my validations, go to:
http://www.redcoopers.net

For some reason, my pictures do not load too well, so if you want the complete version, I recommend downloading the pdf at the top link on the page.

Anyway, if you can afford it, SPLASH is very good for potential flow calculations.

-Jon

[brosen]

I'm pleased to announce that our SPLASH 1995-vintage
fancy color brochure is now available electronically at:
http://www.panix.com/~brosen/splash-1995-brochure

This may help shed light on some of the questions
raised about SPLASH. There are a good number of
comparisons available, both steady and unsteady.
Many of the results are still as good as any available
elsewhere, even after ten years or more.

Also, the papers listed at our website:
http://www.panix.com/~brosen/publications.html
have additional comparisons, especially the three
papers from the Chesapeake symposium.

Regarding public release experiment for hulls,
to which might this refer?

Bruce Rosen
President, South Bay Simulations, Inc.

P.S. I have just discovered this forum.