Added mass coefficient in MoorDyn 4.0.2

Wind & Water Computer-Aided Engineering Software Tools
1

Dear OpenFAST developers team,

I have a question about the impact of normal added mass coefficient Ca of Rod objects in MoorDyn. We are looking to compare OpenFAST simulations with sub-scale experimental pitch decay tests and would like to adjust the added mass coefficient in OpenFAST version 4.0.2 to better approximate what we are seeing experimentally in terms of platform natural frequency and damping. A challenge with modelling our platform is its round torus geometry that sits partially submerged in the water. This platform model I’m testing is currently represented in MoorDyn version 4.0.2 as eight horizontal rods arranged in an octagon (under Rod Types in the MoorDyn input file) which are all connected in one coupled body. We’ve tried to model the platform in Hydrodyn/Subdyn in the past but have had complications with modeling both hydrostatic and hydrodynamic loads accurately with the OpenFAST 4.0.2 capabilities and so are sticking with the MoorDyn model for now. Currently I am trying to change the Ca of the 8 individual rods which make up the platform to understand its sensitivity in OpenFAST. However, it does not have the effect I am expecting. I have plotted pitch decay tests for Ca = 0,1, , 100 with an initial pitch angle of 10 degrees. I don’t see any noticeable change in the pitch response for small changes in Ca and when Ca is set to an extremely high value (blue line on the graph below), the pitch period decreases, where I would expect more added mass to increase pitch period. How is Ca for rods being used in MoorDyn and how can I capture the added mass effect of my experiment given the limitation of keeping the platform in MoorDyn?

Many thanks,
Amasa

Ca-100
Ca-100488×364 37.5 KB

2

Hi Amasa,

I had good experience with the rod hydrodynamics when first implementing it (after a lot of code iterations). Horizontal surface-piercing rods are the hardest case (with the most questionable modeling approximations), but they still worked in my experience. Though I don’t remember doing this specific kind of test.

Can you share the input file you’re using (or a sanitized alternative version that reproduces the same phenomenon)?

Ryan also shared that there’s a thesis here where the author kindly wrote up the theory of the implementation, in case it’s a helpful reference: https://github.com/FloatingArrayDesign/MoorDyn/issues/330

But I suspect we can start by looking for a bug in either the input file or the code.

Matt