[iman210]

Hi,can any one help me in finding 3D panel method (Boundary Element Method) code (Matlab or Fortran) to calculate wave resistance for ship hull considering free surface.
I have written a Fortran code for this case and my code has about 20% error in comparison with experiment. NowI need an exact and correct code to modify my code.
Thanks.

[adrin]

1) have you tried your code at higher panel density? That is, is your solution converged? (2) what makes you think that 20% error is poor, or that you can get better accuracy with other panel codes? do you have evidence of that in the literature? I don't know the details of your problem but it seems to me that expecting high accuracy (in what appears to be a problem with viscous effects) from a potential flow code is unrealistic.

adrin

[iman210]

1)No, I haven't. Because my laptop RAM don't allow. Yes, my solution is converged but until Fn=0.30, after this Froude number I couldn't get correct solution. 2)I see different paper used panel method and got relatively exact solution(error <10%).

[adrin]

Sorry, your answers to item (1) are inconsistent. You say your solution is converged though you also say you haven't done a grid convergence test! Perhaps you mean something else by "converged"?

As for others having obtained accurate solutions with panel methods, can you identify the fundamental differences between their approach and yours? The issue might be related to the order of the panel method, solution accuracy, modeling assumptions, implementation details, etc. I'm bringing up these points to recommend to you that there are a whole list of parameters for you to consider before delving into implementing something new/different.

adrin

[robo]

Just a quick note in reply to Adrin above, but for wave resistance, there should be minimal effect of viscosity on a conventional hull form. The free surface wave caused by a ship is a largely inviscid phenomenon. I would say a panel code should be able to do better then ~20%. Panel codes are still quite common for many marine applications involving waves, eg seakeeping. Total resistance however will have significant viscous effects.

[iman210]

Originally Posted by adrin

Sorry, your answers to item (1) are inconsistent. You say your solution is converged though you also say you haven't done a grid convergence test! Perhaps you mean something else by "converged"?

As for others having obtained accurate solutions with panel methods, can you identify the fundamental differences between their approach and yours? The issue might be related to the order of the panel method, solution accuracy, modeling assumptions, implementation details, etc. I'm bringing up these points to recommend to you that there are a whole list of parameters for you to consider before delving into implementing something new/different.

adrin

I used different number of panels for the code. The best solution was obtained when the size of panels in fore and aft of free surface was the same. So what's your meaning of "converged"?
One of the main articles that I have used is entitled 'Computation of wave-making resistance of a catamaran in deep water using a potential-based panel method' by Tarafder, There are some other papers by this auther for calculating wave resistance of a monohull ship. I implemented the formula used in the paper exactly but I couldn't get the results of the paper.

[iman210]

Originally Posted by robo

Just a quick note in reply to Adrin above, but for wave resistance, there should be minimal effect of viscosity on a conventional hull form. The free surface wave caused by a ship is a largely inviscid phenomenon. I would say a panel code should be able to do better then ~20%. Panel codes are still quite common for many marine applications involving waves, eg seakeeping. Total resistance however will have significant viscous effects.

You're right, but this is depend on the type of ship. When L/B is greater than about 7 the wave resistance is dominant.