Damping Questions

Q: What is the difference between damping, drag, drag caused by current, and drag caused by waves?
REV 7.04

A: First let's us agree that drag and damping are considered the same. Now, divide damping into radiation damping and viscous damping. The first of these is the dissipation of energy though waves being generated by the vessel moving. Everything else is viscous damping. Now, this stuff depends on the square of the relative velocity of the vessel with the relative velocity of the water particles. So, viscous damping depends on the vessel motion, the current velocity, and the wave velocity.

Q: Can you put the natural period in a report?
REV 7.04

A: No. This is because there is no natural period. Here the mass depends on frequency so classical natural frequecies do not exist. Also there is the issue of what these would be for a complex system. Unless the stiffness and mass matrices are diagonal the natural frequencies are not really for roll, pitch and heave.

Q: Why does the model of the BALDER have #TUBEs added for viscous damping when your answer to a FAQ says to use -CS_CURR 1 1 1?
REV 7.04

A: The #TUBEs are included in the model so that we can show the efects of using them. See Section XIII of the the document: http://bentley.ultramarine.com/hdesk/document/verify.pdf

Q: When I compute pressures on my vessel I find that the surge radiation damping is zero. Is there something wrong with my model?
REV 7.02

A: No, you model is ok, it is Strip Theory that is wrong. Strip Theory is an approximation of the true solution (3D Diffration) which is valid for long, slender vessels and part of the appoximation is neglecting surge radiation damping. If surge is important, then use 3D diffration.

Q: Can you provide me some write-ups or explanation on the Kato & Tanaka method (ie quadratic formulation, linearization method etc) for calculating viscous damping?
REV 7.00

A: The best source for information on the "Tanaka" method is a paper "Ship Sway, Roll, and Yaw Motions in Oblique Seas" by Schmidke in SNAME Transactions Vol 86, 1978. if I can get it scanned and will send it to you. As to the linearization method, MOSES has two of them. Click here and look in section III.G for the theory and Here for a more general discussion.

Q: How can I input a 6x6 damping matrix?
REV 7.00

A: In MOSES you model physical things. Thus, there is no need to input a 6x6 matrix. Consider for example, a single dashpot. This can be modeled as either damping added to a GSPR connector or one term on the diagonal of a #DAMP load attribute. Now, either of these act at a point, either the point specified for #DAMP or the ends for the GSPR. Also this acts in either the element system or the part system. Now, when the behavior is converted to the body system, a part of a 6x6 matrix results, but you only need to specify a convenient subset of the information.

Q: How does MOSES get low frequency damping?
REV 7.00

A: SRESPONSE does not include low frequency damping. The only damping here is due to viscous effect, so you can claim that this is conservative. In time domain, there is ample damping from all the real effects.

Q: How can I add a damping of some percentage of critical?
REV 6.01

A: A problem here is the definition of critical damping. Critical damping is defined in terms of the natural period, but there is no "natural period". See the question above. Thus the first part of the answer is to ask those asking for 6% Critical to define critical and then convert this into a real damping value.

The #DRAG command can then be used to add a diagonal damping matrix. Be careful to add it to the correct point to minimize coupling and you need to change the default to get it to be applied in the frequency domain.

Q: Should I alter the default damping values for really large structures, such as big TLPs or large floating jettys?
REV 5.10

A: You should turn off the Tanaka damping and use -cs_curr 1 1 1 on PGEN or &describe piece. This adds some viscous damping to the problem, which is needed for these types of vessels.

Q: How can I get statistics in the time domain of the force produced by a #DRAG load attribute?
REV 5.06

A: The basic idea behind MOSES is that one models "things" and these "things" always effect the behavior of the system. Even Tanaka roll damping effects roll in a time domain simulation. As a rule, unless we say it does not work explicitly, it works all the time.

Now, obtaining the variation of force with time is accomplished via post-processing. We do not store the forces on each item as a function of time since it would be highly inefficient. Instead, there are two commands, ELMFORCE and LDGFORCE which recompute the forces on beam elements and load groups and make these available in the Disposition Menu. Here, you can get the statistics.

Also, when I am in doubt as to how something works, I normally build a simple test to see for myself. For example:

The Data file

     &describe body box
        plane 0 100 -rect 0 20 50
     &describe load damp
     *cen 50 0 5
     #drag *cen .064*100*10*50 1

The Data file

     &instate -loc 0 0 -10
     &weight -compute 5 32 29 29
     &instate -vel 10
     &apply damp 0
     &loop i ( 1 2 )
        &status loadg_mult
        tdom -tstep .25 -maxt 10
               report loc vel
        &apply damp  100
would have told you quite a bit about how this works in the time domain. In particular, it does two time domains, one with and one without the #DRAG so you can easily see the difference.

Q: How can I account for the effect of bilge keels in MOSES?
REV 5.05

A: Use the #PLATE command. You will need to create the appropriate points along the hull for use in this command. There is a sample dedicated to this topic. Look a the b_keels in the "How to" section of the samples Click here.

Q: How can I remove all viscous damping terms in the damping matrix when computing RAOs?
REV 5.03

A: Do not include any Morrison elements, #PLATE or #AREA commands. Also, the Tanaka damping must be turned off using the A_TANAKA BNAME NONE command, and any panel damping needs to be turned off with the option -cs_cur 0 0 0 .