Hi,
I’ve made some modifications to my local version of WISDEM to allow me to get what appear to reasonable results out for floating platform designs. However, there’s a couple of things I’d like to be able to add in to improve this, and I wonder if anyone can help with these.
- I’d like to be able to run 2 load cases for the optimisation, essentially covering DLC1.6 (as is coded in as the standard case) being rated thrust and severe sea state, and I’d also like to consider DLC6.1 to cover an idling turbine with extreme wind and metocean conditions. I’m wondering if there might be a way to implement this using the existing workflows?
- For the above, I would consider that I should be able to set up the two DLCs as the loading conditions in the modelling inputs file, however, two points on doing this I am unsure of:
- The turbine RNA forces and moments - which outputs of the RotorSE (or other) modules would be representative of the inputs required here?
- Is it possible to include wave conditions within these as well - I guess it would be necessary to include additional inputs to cover this and account for these in the environment module?
- I’m also wondering if anyone has any experience of modelling concrete floating platforms successfully in WISDEM? I’ve tried this quickly, but seem to get results with masses of less than half of what I’d expect so clearly something isn’t working quite right with this!
I understand WEIS would be a better tool to use for this, but considering the computational effort I would prefer to get a solution working in WISDEM that can provide an initial level of insight from a reasonably quick calculation.
Thanks for any help you can offer!
Hello and thank you for your interest in WISDEM and its FloatingSE module. I think WISDEM can do some, but not quite all of what you’ve outlined.
- There are floating platform-only optimization examples included that use multiple “DLCs”. Look at the
spar_only or semi_only drivers in the 09_floating example that use this modeling options file. These DLCs just have to specify the loading condition at the tower top.
- You mentioned wanting to run a rated thrust and an idling case and maybe different sea states. WISDEM does not have an easy way of dialing in a series of operating points (rpm, blade pitch angles, environment state) and pulling out loading values. This sounds like a helpful feature that we’ll have to think about for future development. For now, you might have to create separate input files that force each condition.
- When running WISDEM to pull out the tower top properties, look for
towerse.rna_mass, towerse.rna_I, towerse.rna_cg, towerse.tower.rna_F, and towerse.tower.rna_M
- Regarding concrete for floating platforms or towers, we have discussed this occassionally. For specifying material properties, you could pattern concrete like the blade composite layers with anisotropic properties arising from the rebar. One major obstacle is the stress and failure analysis to form the optimization constraints uses DNV steel structure analysis, which do not apply to concrete structures.
Others may chime in with additional comments and suggestions.
Hi Garrett, thanks for the quick and detailed response to this!
- I’ll check these out and see if I can use the outputs you note from the towerse module!
- In terms of using the 2 loading conditions, I’ll have a try to get the outputs for idling and operational. I suppose in terms of the two different sea states this would have to be added to WISDEM then as currently it would just use the one wave height and period for both loading conditions that are entered as RNA loads?
- Yes, that makes sense. Maybe I’ll just leave the concrete for now.
Thanks!
Hi Garrett, I’m wondering if you can provide any insight into the NLBGS convergance aspect of the optimisation? When I’m running with the 15MW model this converges based on the inputs provided in the glue_code.py script typically by the absolute tolerance being below the threshold given, however, when I use the 22MW model that has been provided these absolute tolerances don’t seem to ever get any lower than 1 or 2 at a minimum - quite a way off the 1e-2 tolerance provided. I’m wondering what in this model might be causing this, and what I could try to resolve this?
Thanks!