[katakgoreng]

Ni Niels,

I want to initialized by simulation with a stokes 5th order theory over the whole domain. I want to shift the initial solution by specifying the phi value. However, it seems that I can't do it for the stokes 5th, the initialized solution is the same as when the phi is zero. It works without any problem for the 2nd order stokes theory though.

Part of the waveProperties.input dictionary is as follows:

relaxationNames     ( inlet outlet );

initializationName  inlet;

pName               p_rgh;

inletCoeffs
{
    waveType    stokesFifth;  
    Tsoft       0.00;
    depth       1.50;
    period      1.00;
    phi         3.1416;
    direction  (1.0 0.0 0.0);
    height      0.10;
    stokesDrift 0.00;

    relaxationZone
    {
        relaxationScheme    Spatial;
        relaxationShape     Rectangular;
        beachType           Empty;
        relaxType           INLET;
        startX              ( -5.0 -99.0 -99.0);
        endX                ( -4.0  99.0  99.0);
        orientation         (  1.0   0.0   0.0);
    }
};

Am I doing something wrong with the wave properties specification?

Kind regards,
katakgoreng

[ngj]

Good morning Katakgoreng,

Unfortunately I do not have access to the source code at the moment, but a bug is likely. You will get the quickest answer by looking into the source code yourself.

According to the wiki 'phi' is accepted as a variable (even required).

Kind regards,

Niels

[Olbap]

Hello Katakgoreng,
in StokesFifth.C, the cosine argument is defined for eta and U as

 scalar arg((k_ & x) - omega_*time);

instead of

 scalar arg((k_ & x) - omega_*time + phi_);

as it is for StokesSecond.C. So I guess that is part of the issue. Maybe you have found it already, did you solve your problem?

[katakgoreng]

Originally Posted by ngj

Good morning Katakgoreng,

Unfortunately I do not have access to the source code at the moment, but a bug is likely. You will get the quickest answer by looking into the source code yourself.

According to the wiki 'phi' is accepted as a variable (even required).

Kind regards,

Niels

Hi Niels,

I guess it is a bug as pointed out by Pablo.

Originally Posted by Olbap

Hello Katakgoreng,
in StokesFifth.C, the cosine argument is defined for eta and U as

 scalar arg((k_ & x) - omega_*time);

instead of

 scalar arg((k_ & x) - omega_*time + phi_);

as it is for StokesSecond.C. So I guess that is part of the issue. Maybe you have found it already, did you solve your problem?

Hi Pablo,

Many thanks for pointing this out. I added phi to both eta & U routine, recompile & it works now.

Kind regards,
katakgoreng

[ngj]

Good morning

Thank you for the bug-fixing to Pablo and Katakgoreng. I will put it on the list, so it is corrected, when I make the next update of the code.

Small question: Have you written it as "+ phi" or "- phi" and is it behaving as expected in terms of the shift in the free surface, when you give a small value of phi; e.g. phi = 0.1 = displacement in the propagation direction? Could you please check this for me, now that you have it all up and running?

Thank you and kind regards,

Niels

[katakgoreng]

Hi Niels,

My implementation of phi for Stokes Fifth is as follows:

scalar arg((k_ & x) - omega_*time + phi_);

I ran a few cases corresponding to phi=0, pi/4, pi/2, 3/4pi, pi.
The plot of eta vs x for all cases (Stokes Fifth):



I ran the same cases for Stokes 2nd Order and got the following result:



Hmmm..I guess the implementation for stokes fifth should be -phi then..

Kind regards,
katakgoreng

[ngj]

Thank you very much for this very clear analysis. I will adjust the stokesFifth theory with "-phi", when I am back at a computer with OF. This is a minor change, so the change will be communicated through the Wiki:

http://openfoamwiki.net/index.php/Co..._.28History.29

Kind regards,

Niels