-
Notifications
You must be signed in to change notification settings - Fork 464
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Linearization of the model #738
Comments
Dear @AbdElhamidNabeel, Here are my answers to your questions:
Best regards, |
Dear @jjonkman, Thanks for your reply, actually i have figured out some of what you said and there are still some questions rolling in my mind: 1- I apologies for my ignorance about some parts in the different offshore models as i'm an electrical engineer and i don't have enough knowledge about the the aerodynamics and hydrodynamics of the models, i used before FAST V7 offshore model and actually there weren't all that varieties of models to choose from, so i want to know what is the best model to mimic the uploaded files Linearized_Offshore_model_FASTV7 2- In primary.lin file of FASTV7 (included in the uploaded files), how to output linearization for different Azimuth angles in openfast? . Also previously, the States matrix includes only the Drivetrain rotational-flexibility and GenSpeed states (4x4 matrix), but in openFAST, based on the information you have provided i tried for simplicity the "5MW_OC4Semi_Linear" model and i can't figure out how to determine the number of I/Os and also the states (it gives out the ED States which i need and also the HD states which aren't of my interest so i want to discard the later ) 3- Can i perform linearization of already existent nonlinear offshore model i.e "5MW_OC4Jckt_DLL_WTurb_WavesIrr_MGrowth", as i have tried and turned on the linearization parameter value to TRUE But MATLAB gives me error in SFUNC when i try to run the model. so what are the parameters to be modified in that case?. So, please i want recommendation for the model choice and the parameters to be configured in the model to obtain similar results obtained from FASTV7. Thanks in advance and sorry for taking some of your time to help me. Best Regards, |
Dear @AbdElhamidNabeel I'm trying to put together some documentation and scripts for the linearization. I must say, Offshore cases might require a bit more work at the moment. Can you have a look at the following page (in particular the troubleshooting section, and the linearization sections). I'm hoping to make this part of the openfast documentation (see #308), after it has received some feedback from colleagues. Any feedback from your end would be great, in particular if any of this documentation does not cover your questions, I'll try to add to the documentation. Thanks, Emmanuel |
Dear @AbdElhamidNabeel, @ebranlard's work-in-progress documentation is very useful. Here are my answers to your direct questions:
Best regards, |
Thanks @ebranlard i wish it workout so we can use your valuable documents in the very soon ! Dear @jjonkman, Thank you for this valuable recommendations, actually i have tried the recommended model but upon running it in MATLAB it generates errors. here are the files of the model after making the following modifications Land_Modeshapes_Model Actually i have tried bunch of things of both models (5MW_Land_Modeshapes / 5MW_Land_BD_Linear) and i want to share what i have reached recently: 2- I discarded some outputs from the linearization output model from the input files and it doesn't affect the linearization process 3- I tried the settings (CalcSteady = True with NLinTimes = 36) in 5MW_Land_BD_Linear and actually the linearization process haven't done at all !!!, when i set the settings to (CalcSteady = True with NLinTimes = 1), the linearization begins and actually it doesn't compute at NAzimsteps = 36 like in FASTV7, but it do the linearization actually in random times / random angles which i couldn't figure out the pattern of the steps in azimuth angles !!!! here are the results and the settings 5MW_Land_BD_Linear_Model 4- when i tried to introduce the NRELOffshrBsline5MW_Onshore_AeroDyn15.dat file in the 5MW_Land_BD_Linear.fst input files model and tried to perform linearization to mimic the situation in the Previous FASTV7 Model , actually MATLAB throw error in the SFUNC. so is the AeroDyn is linearizable in OpenFAST ? and if it's linearizable, how can i modify in the 5MW_Land_BD_Linear model to enable the linearization feature of this part (I tried putting the setting of CompAero = 2 to enable AeroDynV15, and copied the NRELOffshrBsline5MW_Onshore_AeroDyn15.dat from 5MW_Land_BD_DLL_WTurb files to the 5MW_Land_BD_Linear directory ) 5- I want to perform linearization at Windspeed = 18 m/s, so i tried to add the NRELOffshrBsline5MW_InflowWind_Steady13mps.dat file in the 5MW_Land_BD_Linear.fst input files with CompInflow = 1, and changed the settings in the .dat file as following (WindType = 1 (steady) + HWindSpeed = 18 ). But also trying to perform linearization at this case, MATLAB throw error in SFUNC. So, is it linearizable or not ? and if it's how can i obtain this feature So, to summarize. How to enable the features of AeroDynV15 and InflowWind in linearized model and how to determine clearly the azimuth angles steps in the linearization process Best Regards, |
Dear @AbdElhamidNabeel, Here are my answers to your questions:
When linearizing, the azimuth angle is written near the top of the linearization file (line 10). Best regards, |
Dear Prof @jjonkman, Thank you very much for your help, actually i have successfully run the linearization. for answering the questions 4 and 5 ---> I solved it by the following: 1- in AeroDyn15.dat file, by modifying the value to be AFAeroMod = 1 and enabling the FrozenWake = True 2- exchanged NRELOffshrBsline5MW_InflowWind_Steady13mps.dat with NRELOffshrBsline5MW_InflowWind_Steady8mps.dat of the input inflow_wind file in .fst file and modified the required WindType = 1 (Steady) and changed the HWindSpeed = 18, Actually, i have a last very tiny question. Upon performing linearization for NLinTimes = 36 and CalcSteady = True, MATLAB doesn't stop at the 36 linearized model but it continue to perform the 37,38,... inf linearization (like the first 36 linearized model and the next 36 (up to 72 totally now) have the same results and so on up to inf !!!! is there any explanation for this why it generates the same 36 linearized models without stoping ! As i expect after performing the 36 linearization it has to stop !! |
The problem of the linearization continuing after NLinTimes is a bug in the FAST_Library. We have identified the issue and I'll be submitting a fix very soon. I'll report back here when it's ready. |
You can see if MATLAB stops after finding the trim solution in https://github.com/bjonkman/openfast/tree/b/Simulink_TrimLinear |
Thanks prof @bjonkman, Actually i have some questions regarding the latest updates in the program: 1- i have tried to build the openfast version that you have recommended and actually i receive errors during building in SysMatlabWindows.f90 file, and i have read the note mentioned in the openfast v2.6 regarding this point but i didn't actually understand how to overcome this issue. 2- i tried to use the ready-made files uploaded (MAP_x64.dll, FAST_SFunc.mexw64, and OpenFAST-Simulink_x64.dll), and the linearization files contain some creepy outputs of aerodynamics. These creepy outputs place error while using the MBC3 toolbox The error : Error in fx_getMats (line 223) Error in Linearization (line 99) Here are the settings and the outputs of the linearization Linearization output and linearization settings 3- Is there any method to include certain inputs during the linearization .... i want to include only the (ED Extended input: collective blade-pitch command, rad input) in the linearization process, i'm not interested in other inputs. Thanks |
If you don't want to change/build the OpenFAST code, you can otherwise change the |
@bjonkman I actually had tried the recommended changes and i still receive errors during compilation regarding FAST_Library.f90 and FAST_Subs.f90 files (I have chosen release_matlab - x64 options for building). It seems that SCDataEx_Types.h file or SCDataEx_Types.f90 doesn't exist !!! .Finally, i had to remove the aerodynamics inputs from .lin manually and change no. of inputs as well to overcome the problem. But, i perform the linearization for like 8 times for different windspeeds and for 36 azimuth angle (8*36 .lin files) !!! that would be such great effort for me to do that manually !!! . So, i hope you can find a programmable solution in the very soon. Thank you for your efforts Here is the error i received while compiling: |
It looks like your build did not recognize the supercontroller data exchange types. This may be because the OpenFAST Registry needs to be run again, or you may just need to build it again (Visual Studio doesn't always build the *_Types.f90 files after they are generated by the Registry, but will build them if you build the solution again.) The other errors seem to be caused by that first issue. Hopefully you can get that to build now. If you do need to change those .lin files by hand, I'd recommend writing a script to do it. |
Hello,
I have tried many OpenFAST CertTests and it worked well, actually i have few questions :
1- Where to find the offshore models/tests, as all i see is 5MW Onshore tests ?
2- How can i perform linearization on the model, is there any manual guide for this in OpenFAST ?
3- How to select certain input/output for the state-space output linearized model, i want to select the collective pitch angle as input and the GenSpeed only as output ? as i found only three cases to choose from : (0) none, (1) standard (6 inputs/60 outputs) or all inputs
4- How to determine certain states of the linearized output model ?
Thanks
The text was updated successfully, but these errors were encountered: