'Wind Turbine' simulation project by 1318980


#1

I created a new simulation project called 'Wind Turbine Optimisation':

More of my public projects can be found here.

Contents

  • Preface
  • Introduction
  • Conceptual design
  • Modelling in OnShape
  • Meshing in SimScale
  • CFD Analysis in SimScale
  • Post Processing and Analysis Results
  • Optimisation based on results
  • Structural design
  • Structural optimisation in ST Inspire
  • Structural analysis
  • Transient simulation of final design

Preface
I’ve been working on this project for a while and was originally a test for an aerofoil script I had written in OnShape. Designing the blade incorporate different aerofoil profiles into a single blade and used guides to loft between them. This wasn’t as easy as I had originally envisioned and decided it would be a shame to let the design stop there, I need to see if my calculations were valid after so much hard work.

My initial meshes failed and it took me a while to remember how to go about the setup, however with the help of Anna and a previous simulation from the drone workshop I was on my way. I hit a few hurdles with blade design (which is why the designs start with configuration 2) and so I thought it would be good to include a bit of an overview to modelling the design in OnShape.

I have written about the design and optimisation approach I took, It was a learning experience and I most definitely haven’t got it right first time. However it does show an approach I feel most students like myself would take and highlights why ‘the boring’ stage of initial research is so important before jumping in and making a design in CAD.

Introduction
In this project I shall outline the mathematical models, simulations and data used to create the initial conceptual design, I will then explain how I created a 3D model from the concept in OnShape and then I will bring the design into SimScale to analyse it. Using this analysis, I hope to identify areas in which the design can be improved and iterate the design analysis workflow to produce an optimal blade design.

Once the aerodynamic optimisation is done, a further simulation will be done to generate data ready for a structural design and hopefully a bit of a contemporary optimisation using structural optimisation software ST Inspire (I think it’s a really cool idea).

As a lap of honour I will perform a transient analysis of the final aerodynamic shape and process the data to hopefully show some really cool results.


Wind Turbine blade, fluid and structural optimisation using CAE
#2

Hi @AnnaFless,

I need to get some help (again I know), Ive created a practice mesh operation called “MRF Prac” as I was having trouble creating a cell region for a rotating zone, I left “create multi region” unchecked and followed instructions on simscale documentation to create new cell region. Ive played about with quite a few options and still no region (I get an empty zone within my mesh, and there is only one region in the tree). So I went looking at public projects and couldn’t find any that were after the “create multi region” option was added (I presume that was added when CHT was.

Is there any chance you could take a look and see if I’m being dopey or if there is an issue?

Kind regards and sorry to trouble you,
Darren Lynch


#3

Hi Darren,

Just a very very small fix here - you’ll need to select a volume for refined surfaces of the rotating zone (rather than a face set as you have now).

Just to clarify:

You can have a look at this drone project as well for a similar example: https://www.simscale.com/projects/AnnaFless/drone_aerodynamics_project_1/

Best,
Anna


#4

Oh, thank you! I think because it was a surface refinement I was thinking to much about faces :slight_smile:

Many Thanks,
Darren Lynch


#5

@1318980 - saw the meshing error, one of our CFD engineers will have a look


#6

Hi @AnnaFless,

I don’t think anything is going wrong really I got the results I just changed the settings to increase the accuracy of the mesh and failed (Its been a fine line between success and failure!).

Many Thanks,
Darren Lynch


#7

Saying that, after looking at the log. Ive never had a fatal error :-S


#8

Hi @AnnaFless,

I have now got it sorted and I believe I have a good mesh, Thanks for your help.

Darren Lynch


#9

Glad to you were able to get it figured out! I wasn’t liking the sound of that fatal error :dizzy_face:


#10

Hi @AnnaFless ,

in my last mesh I got an error and I don’t know why, the same mesh setup has been used on the same model at a different blade pitch. There is nothing in the mesh log that I can see would result in a failed mesh, any chance you could have a look?

Many thanks,
Darren Lynch


#11

Just got it to work with an extra mesh level, but Its now a million more cells than the others! ill see how it goes :slight_smile:


#12

@1318980,

Excellent - glad you got this figured out! Let me know how the simulation works out :slight_smile:


#13

Hi @AnnaFless,

I have some odd post process results.

Compare the top screen shot (Front view) and the bottom one, in the top I can see the area on the leading edge to be a deep red as shown by the bottom shot which is a slice taken near the tip. The rest is green or approximately 0.

Now compare the second image (Rear view), where I expect blue (low pressure) I see red? This had been confusing me whilst I was processing the data, and when I did the slice I was totally baffled.

Can you help? I don’t think its an error on my part but I could be wrong :cold_sweat:

Sorry for troubling you again,
Darren Lynch


#14

I did have to play around a little bit here - it seems like the filters are not working in all cases. I loaded the results - then switched to p[cell data] with the results and then back to p[point data] and it worked.

I will report this to the engineers so we can have a deeper look. ParaView should work in all cases if you have installed on your computer

Best,
Anna

Oh and it looks like you found the right section, thanks for the post-processor feedback :slight_smile:


#15

Ok ill have a go with the work around, Cheers @AnnaFless!


#16

Hi @AnnaFless,

I’ve been doing a lot of work on my next design and have been doing a lot aerofoil analysis in an aerofoil program, Ive noticed that Renaults number makes a big difference to my results so I’d like to get some ball park figures for how Re changes with distance from centre of rotation. Is there any way of doing this in SimScale using my current design? I have been using cord length and apparent velocity however the figures I’m getting are around x10^7 which I imagine are fairly inaccurate?

Ive got a bit of a chicken and the egg situation, I need cord length to get Re I need Re to find Cl and I need Cl distribution to get cord length! oh the joys :slight_smile:

Kind regards,
Darren Lynch


#17

ok I’ve re-calculated my remaulds numbers to range from around 4.8 million closest to the centre and 4.3 million at the tips, does this sound sensible? sounds high to me but I have previously never used Re for any practical applications, I used aerofoil Re=u*L/v where u=apparent wind velocity at each radius, L=approximate cord length, v=1.46e-5 m^2/s as the kinetic viscosity of air at SL. Any advise would be appreciated.

Kind regards,
Darren Lynch


#18

Hey Darren (@1318980), my colleague @Ali_Arafat is having a look into this right now and will get back to you soon :slight_smile:


#19

Thanks @AnnaFless :slight_smile: