Time Domain analysis by ANSYS AQWA

Dear all

I am trying to model OC3-Hywind in the ANSYS-AQWA, but unfortunately I am facing some problems in regard with Time-Domain analysis of the system.
I have verified the hydro-dynamical characteristics (i.e., added mass and damping) through the Diffraction module and also verified the natural frequencies of the system by running decay tests for all six DOFs. However, when I add a wave load case as is presented by Dr. Jankman research studies for the same model in the time-domain and calculate the motion responses, some of them such as surge and pitch are less than what is presented by Dr. Jonkman with NREL-FAST code. This is also clear when I transfer data to frequency domain, in each motion induced peak frequencies occur at the right frequencies, but the magnitude of motion induced peaks are less than NREL-FAST.
I think the difference is because of center of gravity which is at the sea water level in NREL-FAST, but I have taken care of this by transferring the CoG to the mean sea water level and added md^2 to the inertia based on the parallel axis theory, and still could not get the right answers.
I would really appreciate it if you could help me out in this regards.

Dear Fatemeh,

I’m not sure I can guess as to why your ANSYS-AQWA model of the OC3-Hywind system produces less wave-excitation than the FAST model. But let me clarify how the platform center of mass, and inertia are defined in FAST. In FAST, the platform center of mass is defined in a body-fixed coordinate system as the offset from point in the structure where the tower centerline intersects the mean sea level when the structure is undisplaced. The platform roll, pitch, and yaw inertias are defined about the platform center of mass. The OC3-Hywind spar has the following platform mass, center of mass, and inertia properties:

PtfmCMxt = 0
PtfmCMyt = 0
PtfmCMzt = -89.9155 m
PtfmMass = 7.46633E+06 kg
PtfmRIner = 4.22923E+09 kgm^2
PtfmPIner = 4.22923E+09 kg
m^2
PtfmYiner = 1.6423E+08 kg*m^2

Please note that these platform properties account for the structural mass and ballast of the spar, but neglect the mass of the tower, rotor-nacelle-assembly, and moorings.

Best regards,

Dear Dr. Jonkman,

Thank you for your quick reply.

Actually at first step, I used platform properties when I tried to verify added mass and damping coefficients with your paper “Definition of the Floating System for Phase IV of OC3” and it worked.

At next step I am going to verify structure position by another report of you "Offshore Code Comparison Collaboration (OC3) for IEA Task 23 Offshore Wind Technology and Deployment ". (Load case 4.1 & 4.2)

The question is that for this analysis I should consider only the platform properties or I have to use whole system properties which you mentioned it in this topic http://forums.nrel.gov/t/inertial-moments-of-oc3-hywind-components/610/9 .

Best regards,

Dear Fatemah,

The platform mass, center of mass, and inertia I mentioned in my prior post are only for the spar alone. If you’re modeling the full rotor + nacelle + tower + spar as a single rigid body, I agree that you should use the full-system mass, center of mass, and inertia provided in the forum topic you linked to. In FAST, though, we’ve modeled the spar + nacelle + hub as rigid bodies, and the tower + blades as flexible bodies.

Best regards,

Dear Dr, Jonkman,

Thanks for your information.

I added whole system properties and considered reference point at the CoG of system. But at frequency domain, magnitude of motion is less than FAST. These are my results from ANSYS AQWA compared with FAST.

Also I transferred CoG to the SWL and increased mass inertia based on parallel axis theory. But the result was totally different.
can you help me in this regard?

Best regards,
1.JPG
2.JPG
3.JPG

ANSYS AQWA results compared with FAST.
4.JPG
5.JPG
6.JPG

Dear Fatemeh,

The conversion of degrees of freedom from the center of gravity (CG) to the still water level (SWL) was discussed in the following forum topic: http://forums.nrel.gov/t/oc3-hywind-raos/1085/1. Are you applying the same transformations?

Best regards,

Dear Dr. Jonkman,

I calculated the transformation matrix and attached it. I would be appreciate you if you could take a look at it.

Best regards,
trm.pdf (23.3 KB)

Dear Fatemeh,

I agree that q_swl = TransMat^-1*q_cg, with the TransMat^-1 that you’ve shown. However, “q” in this equation is the RAO. I’m not sure why you are premultiplying some combination of inertias with TransMat^-1. Also, the inertias you list (taken from the following forum post I presume: http://forums.nrel.gov/t/inertial-moments-of-oc3-hywind-components/610/2) are the inertias about SWL, not about the CG.

Best regards,

Dear Dr. Jonkman,

I want to run the AQWA by transferring properties of the system to the surface so I can compare my results with FAST. That’s why I did it. I am not sure if it is OK?

You’re right. it was my mistake to use inertia about SWL. I calculated inertia about CoG as follow:
Ixx =1.90E+10 kg.m2
Iyy =1.90E+10 kg.m2
Izz =1.91E+8 kg.m2

But I don’t have other inertia. Is there any references for it?

Best regards,

Dear Fatemeh,

The forum topic I linked above (OC3-Hywind RAOs) shows how to transform the 6x6 mass/center of mass/inertia matrix from the SWL to CG i.e.:

M_cg = TransMat^TM_swlTransMat

where,

TransMat =
[ [ 1 0 0 0 -z_cg y_cg ];
[ 0 1 0 z_cg 0 -x_cg ];
[ 0 0 1 -y_cg x_cg 0 ];
[ 0 0 0 1 0 0 ];
[ 0 0 0 0 1 0 ];
[ 0 0 0 0 0 1 ] ]

and

TransMat^T =
[ [ 1 0 0 0 0 0 ];
[ 0 1 0 0 0 0 ];
[ 0 0 1 0 0 0 ];
[ 0 z_cg -y_cg 1 0 0 ];
[ -z_cg 0 x_cg 0 1 0 ];
[ y_cg -x_cg 0 0 0 0 1 ] ]

A general 6x6 mass matrix is written as shown in the image below, where the 3x3 inertia submatrix is defined about the origin of the coordinate system (not the CG).

Best regards,
MassMatrix.png

Dear Dr.Jonkman,

Thanks for your information. I will work on it.

Best regards.