Tower natural frequencies using ElastoDyn and SubDyn

Jason, apologies if this has been answered before, I couldn’t find it.

Recently we simulated the NREL 5 MW mounted on a monopile in 20 m of water. The tower is modeled in ElastoDyn, with 2 FA and 2 SS modes, and the monopile is modeled in SubDyn. We ran a ULS case that includes breaking waves, and then analyzed the frequency response of the tower FA bending moment (TwrBsMyt). The response shows modes at approximately 0.25 Hz and 1.8 Hz, as expected, as well as at 2.2 Hz and 4.4 Hz (all approximate). We are assuming these higher modes are a result of the monopile flexibility modeled in SubDyn, but are unsure how to assess whether or not they represent realistic physical behavior. Do you have any thoughts on if these modes are believable?

An alternative approach it seems is to model the entire tower in SubDyb, and obtain higher modes this way. Is there a model of the monopile and tower in SubDyn you could share?



Dear Matt,

I don’t have a FAST model of the NREL 5-MW turbine atop the OC3-monopile whereby the entire support structure (tower + monopile) is modeled in SubDyn, but it would probably not be too hard to make one.

I’m not exactly sure about the 2.2- and 4.4-Hz response as I don’t run this model myself often, but they could be from SubDyn, or perhaps from the rotor.

Best regards,

Thanks Jason. In regards to the source of those two modes, whether SubDyn or the rotor, I should also mention that when we tune a TMD to the 4th mode, we get substantial reduction in the frequency response at that frequency. Also we’re in a parked condition, so it’s not related to the rotational frequency of the rotor, though I guess maybe the blade modes? Not sure if that helps figure out where they are coming from. More important though is deciding if they are physically representative, or a weird numerical result. One other point, I believe if you look at the SubDyn output, the first eigenfrequency for the monopile is quite high, like 28 Hz or something. Maybe this is before a very large mass is placed on it, which would presumably lower the frequency?

Anyway, would you recommend making the entire support structure in SubDyn?

Dear Matt,

Because the rotor is not spinning, you’re right that these frequencies may be a numerical artifact induced by the ElastoDyn-SubDyn coupling e.g. platform surge/pitch coupled with tower bending. Unfortunately, without full-system linearization functionality that includes SubDyn, it is hard to interpret the full system modes, natural frequencies, and damping of these coupled modes. However, we compared the results of the coupled ElastoDyn-SubDyn model of the OC3-monopile in FAST v8 against the old FAST v7 model without SubDyn and I don’t recall see major differences for the test cases we were comparing.

The 28 Hz reported in the SubDyn summary file for the OC3-monopile is the fixed-free mode of the isolated monopile (without tower and turbine atop). This is not very informative for your problem.

Modeling the entire support structure in SubDyn would allow you to choose how many support structure modes you want to model, but you will still have an ElastoDyn-SubDyn coupling at the tower top and you will lose all geometric nonlinearities in the support structure (ElastoDyn considers geometric nonlinearities, but SubDyn does not). We usually model the ElastoDyn-SubDyn coupling at the tower base because the geometric nonlinearities tend to much more important for the tower than the substructure.

Best regards,