Mismatch between OpenFAST 3.3.0 and OC3 Generator power for steady wind

Wind & Water Rotor Aerodynamics
1

Hi everyone

I am trying to reproduce the results from the OC3 Phase I project, specifically load case 3.1. In this load case, the NREL 5MW wind turbine is mounted on a tower and a uniform wind of 8 m/s is applied. No hydrodynamic loads.

The results produced by NREL with FAST that I am trying to reproduce can be found here: SimulationResults_NREL_FAST.xls - Google Sheets

In that spreadsheet, for that load case, it appears that the generator power stabilises at around 1870 kW

I am now running OpenFAST 3.3.0, specifically the regression test 5MW_Land_BD_DLL_WTurb. The only significant change I have made is that I am using NRELOffshrBsline5MW_InflowWind_Steady8mps as the inflow file (I also changed the initial rotational speed and the total duration of the simulation). My generator power seems to be about 10% lower than in the OC3 results (see picture below).

OC3 vs OpenFAST
OC3 vs OpenFAST600×448 21.6 KB

Can anyone see what I am doing wrong? Obviously the OC3 results are about 15 years old, could it simply be that the code has been improved since then, or that the definition of the 5MW turbine has changed? If so, does anyone recall what changes that could have been?

Thanks,
Loup

2

Dear @Loup.Suja,

Given that the FAST results you are comparing to are now quite old, it is difficult to say. Certainly, AeroDyn evolved a lot since then and BeamDyn was not introduced at that point. I would first try simplifying the problem, e.g., run load case 2.1a and 2.1b from OC3 Phase I (rigid turbine within and without control); how do you the loads compare for those cases? Does using ElastoDyn instead of BeamDyn yield closer results?

Best regards,

3

Hi Jason, thanks for your tips.

I had a go at 2.1a with Elastodyn, here is what the comparison looks like for some of the output:

Figure_1
Figure_11800×900 146 KB

It seems like the OpenFAST results are consistently lower (also for the in-plane blade bending moment at the root, although it is harder to see in the graph). It could be that the code has changed, or simply a user error somewhere (maybe I was supposed to change something in OpenFAST to match OC3 and I didn’t see it).

I’d be curious to understand why the results are different, but for now I am happy with this.

Thanks again,
Loup

1 Like