Mode Shape Coefficients for modified IEA15MW tower

Wind & Water Structural Analysis
1

Dear all,

I’m currently working on calculating the mode shape coefficients of the tower for a modified design of the IEA15MW turbine installed on a semi-sub floater, to use as inputs for ElastoDyn. My first approach has involved using the version of BModes adapted for floating substructure (BModes_JJ: BModes - Google Drive) along with the Excel tool ModeShapePolyFitting.xls.

However I recently noticed that WISDEM’s TowerSE outputs also include floatingse.fore_aft_modes and floatingse.side_side_modes. Unfortunately, I couldn’t find documentation on these, which brings me to the following questions:

  • Are these outputs indeed the coefficients for the first three fore-aft and side-side mode shapes of the tower?
  • Which approach, BModes or WISDEM, would you recommend for deriving these coefficients?
  • If both can be used, what are the main differences between the tools?

Thanks in advance for your help!

Azélice

2

The BModes_JJ and ModeShapePolyFitting worksheet is a well established workflow here on the OpenFAST forums. However, BModes_JJ is also closed source and deprecated, in that there will be no future updates. There are also a number of ways that users can make mistakes in the workflow (I have made those mistakes myself).

The WISDEM outputs can be directly used in the ElastoDyn input files and were created for that purpose directly. Instead of BModes, WISDEM uses the Frame3DD solver to generate the mode shapes and NumPy’s polynomial fitting tools instead of the spreadsheet. We have tried to automate in code some of the key boundary condition requirements that can lead to mistakes in the original workflow. We acknowledge that it has been difficult to have robust mode identification for every use case, but code quality keeps improving incrementally.

The two workflows should give similar mode shapes for standard land-based towers and blades. The exact coefficients may look different as there are different normalization approaches.

3

Hi Garrett,

Thanks a lot for this detailed answer. It’s very useful.

Best regards,

Azélice

4

Dear Garrett,

Following your last response, I used WISDEM to derive the natural frequencies and mode shape coefficients. Since I’m analyzing a tower on a semi-sub floater, should I refer to floatingse.structural_frequencies or towerse.tower.structural_frequencies? I noticed a slight difference between the values provided by these two variables.

Thanks very much for your guidance!

Best regards