Learning Forum

[enrjcutl]

Hello,

I am trying to simulate the model in the below picture, however, when I initiate time-domain simulation I get the following warning:

[enrjcutl]

This was the warning in full:
CONVOLUTION WARNING. An inaccurate convolution fit has been detected to the added mass/damping for structure#2/1 Surge(X)/Yaw(Rz) for frequency #51 . The fit quality (100% = exact) was -213%. This can be caused by insufficient frequencies or an unusual geometry of the structure. PLEASE - ENSURE THAT THERE ARE SUFFICIENT FREQUENCIES *BEFORE* contacting user support. Frequency range for structure 2/1 should be at least 0.149 to 1.965 rad/sec.

[enrjcutl]

Hello I have since discovered in the AQWA theory manual (pg 156-157 Radiation Force by Convolution Integration):
(1) the added mass matrix is used to obtain the impulse response function matrix.
(2) integrating this with an upper limit of 10 radians and a lower limit of 0.
(3) an extrapolation is carried out to estimate added damping values not in specified frequency range by user.
(4) this extended set of added damping coefficients is used in Hilbert transform to obtain fitted frequency-dependent added mass matrix
(5) a best fit is then obtained between the fitted added mass transformed from the damping coefficients and the directly-calculated added mass from AQWA.
This is where the problem above exists.
Based on my knowledge and applied to the above geometry, I presume AQWA takes into account the radiation effects from the large-volume structure on the small-volume structure to directly-calculate the added mass, whereas I presume for the fitted added mass it is not hence the difference.
My only question now, if it can be answered, is what are the consequences if any for an inaccurate convolution fit and continuing with simulation if the added-mass coefficients used are the ones directly-calculated by AQWA?
Josh