Resistance Calc

Hi, We are trying to do res calc for a barge with ship shaped foreend with four protrusions (almost semi-circular) on sides (two each side. These semi-circles are approx 5 m radius. The vessel itself is 40 m beam. Draft is 4.5 m.
Our FLUENT CFD man tells me a 3x increase in resistance due to these protrusions which I believe is totally overestimated. And FLUENT is giving very high resistance even without protrusions.
Any advises on which CFD software to use to solve this issue?
Thanks.

 

FLUENT is an excellent CFD software and if you look for something better, you will hardly find it.
But it is just a software, and the reliability of the results will be proportional to the competence of the guy who uses it. Same thing is valid for any other software you will find around.
So, don't expect to obtain more accurate results by simply changing the software.

If you think that the resistance is overestimated (based on what consideration?) then you need to check the input ship model, flow parameters, boundary conditions, output mesh, calculation parameters and mathematical models used.
Also, check the water depth of your model. You might have expected to see the drag numbers for deep-water conditions and have been give the results for shallow water, which can be 2-3 times higher.

Just in order to obtain the magnitude of the clean-hull (no protrusions) drag you should obtain, you can calculate manually the frictional resistance of your hull and multiply it by a factor of 2.
You might also check out some published (and verified) drag data for existing barges and see if you can scale or extrapolate (with caution!) the results to your case.

 

I think you will need to build a model, CFD is good for fine tuning a hullform but there can be large errors in CFD prediction when developing new classes of vessel . Probably better to run the bare hull without the protrusions and add the drag from them separately to the predicted resistance.

Did you run through one of the manual resistance predictions for the unmodified hull to compare with the fluent prediction already?

 

That is definitely true, particularily for a ship of these dimensions and costs.
I'm understanding that a part of the problem is that CFD gives a high resistance even for a conventional, clean-hull configuration.
Then, assuming that this is a preliminary design stage and depending on the degree of accuracy one wants to obtain, a less expensive way could consist in finding a published drag data for a barge with a known hull shape and then using them for fine-tuning of the CFD parameters.
If the hull shapes and operating conditions are sufficiently similar, the CFD results for clean-hull configuration can be expected to be reasonably accurate. You can then play with adding protrusions etc, and evaluating the results with caution.

 

Dear Daiquiri and MikeJohns,
First, thank you very much for your replies.
Noted your response on FLUENT.
The vessel is sailing at 10 kts.
I did run a holtrop on clean hull (no protrusions) and got 750 kN resistance without protrusions.
But FLUENT model gives me 1800 kN. I have already asked our CFD team to check again (which they had done already).

Mike, I agree with you that I should make a model and do the tank test.

But however, do you really feel that adding four 5 m radius semi circular protrusions (like a spud leg of a barge) will increase the resistance by 2 or 3 fold? I "felt" that the augment of resistance will be in the range of 15 to 20%.

Thanks again.
Anoop

 

Well, it could be possible. Can you include a pic or a sketch of the hull configuration, just to let us understand better the shape, position and dimensions of these "protrusions"?

 

Please refer attached sketch.

 

Look, 'll make a very approximate calc, just to give you the first impression of the numbers you can expect...
I see that the protrusions are very well rounded, so I will approximate them with 4 semi-spheres (mean diameter is around D= 13 m) at the walls of your hull.
Numerically it will equal 2 full spheres (each one divided in two halves), so, I'll calculate the drag of one sphere and will multiply the result by 2.

The Reynolds number is of the order of 10^8, so fully turbulent flow and correspondant sphere drag coefficient Cd = 0.2 .
Frontal area of a 15 m sphere is S = 177 sq.meters.

The drag force for one sphere is given by D = 1/2 (rho V^2) S Cd.
Numericaly, it gives D = 480 kN for each sphere, 960 kN for 2 spheres (or 4 semi-spheres).

So, you have 750 kN by Holtrop, plus some 950 kN by 4 semi-spheres, which gives a total of 1700 kN.

The order of magnitude of this (extremely) simplistic analysis is in line with your CFD results, and explains you where the big increase in drag comes from.

 

I get between 1100kN and 1200kN based on a rough model of your hull. I have lifted the transom above water level. If it was fully submerged to the keel line then the drag would be much more. 1800kN might be close.

This boat will behave vastly differently to a slab sided vessel. Each side protrusion acts a separate hull. I have seen the before in a much smaller hull with just one protrusion on each side.

I see the drag rising rapidly between 8 and 9kts. This is the approximate hull speed of the four protrusions. Each of these will produce their own wave train. There will be a deep depression behind each one.

Your CFD guys should be able to see the same thing.

So you have created one large hull with lots of surface area and not much wave drag and four separate hulls attached with much less wetted surface but a huge amount of wave drag. The short hulls are extremely bluff shape as well. These attached hulls have very low length to beam ratio as well.

The boat will have a very complex wave train with a lot of interaction.

Rick W

 

Thanks for all the help Daiquiri and Rick.
Daiquiri, direct drag calculation is noted. But then again, does it mean that we need more power (950 kN) to propel these four protrusions than propelling a 120mx40m vessel (750 kN)?
Rick, which method/software did you use? We are getting same stern wave train just behind protrusions.

 

I used Michlet. It is based on thin ship theory so will have some error for a vessel of these proportions. I cannot say how much error.

If I had an exact model of the vessel (prefer .igs format) I could compare with your results. I have got reasonable results validated with Michlet down to L/B of 4.

I did a design review of a small 9m vessel that had been built and performed way below expectations. It had a large section change over a short length and actually behaved as two short hulls rather than one big hull.

The hull was improved a little by increasing the transition length but it was still a dog. It had a length to beam ratio of 4.5 and the improvement corresponded with my prediction.

I get the drag rising very steeply above 8 kts. If you were to suitably lengthen the side protrusions you may find the drag would remain low up to your target speed. Your CFD result should also show the same thing. The other possibility is to have a 12m wider vessel and avoid the bumps.

Rick W

 

anoop...

As you rightly concluded in #5....if this is a serious enquiry and real data needed, then tank test it. The CFD will help your sensitivity analysis, but beyond that, as daiquiri (& MJ) noted, all highly questionable for real hard data to place hard money behind the results!.

 

First let me underline again that the calculation I've provided is a very quick one and a very approximate one, as I stated before. The drag data for spheres comes from wind-tunnel testing, therefore is based on clean, non turbulent airflow (or waterflow, if you want). What you really have is a distorted flow-field due to the hull and also some wave drag involved, since the upper part of the protrusions (aparently some 25% ?) will be above water surface.

The fact that results of my simple calculation coincide so well with what you got from your CFD guys might indicate that the gross part of protrusions' drag will be viscous, not due to waves. Of course, a detailed tank testing or a reliable CFD (means "more-closely-fine-tuned-to-your-type-of-hull") might either confirm or deny that claim...
But if that conclusion is true, you then have a strategy for minimizing it - streamline your protrusions as much as you practically can. Prolongate the rear part of it and give it a correct shape (think of airfoils). It will make the surrounding pressure field vary more gently and will delay the flow separation.

Now, back to you question - you want to have numbers, but the only thing I can say without having other info about your ship is that most probably the additional drag due to 4 protrusions is of the same order of magnitude as the drag of the clean hull. It is pretty much possible, and the difference between your Holtrop and CFD preliminary calcs indicate that.

You now have here the gross value of forces involved (more or less confirmed by different parties), but only a tank test will give the final numerical answer to your question beyond reasonable doubt.

 

Gents,
We did a full CFD for this one again and got approx 1350 kN for 10 kts. This will give us very high thruster values, so we decided to change over to another (fairer) hull form. It will take a while for us to finalise new thrusters but I will let you know the final outcome of this newer hull form.
Daiquiri, I was on vacation and didnt see your mail. My apologies.

 

Can you give any dimensions for the new hull and what drag you are determining for it?

Rick W