FAST8 Linearization

This forum if for discussing controls. Questions about how to implement controls in FAST are more appropriate to the CAE Tools forum.

Moderator: Bonnie.Jonkman

Taha.Fouda
Posts: 44
Joined: Thu Jun 22, 2017 1:51 am
Organization: DLR Institiue of Flight Systems
Location: Germany

Re: FAST8 Linearization

Postby Taha.Fouda » Tue Mar 27, 2018 5:14 am

Dear Jason,

I have a question regarding the linearization. I have linearized my wind turbine at region 3 ( wind speed above rated speed ). I have adjusted the initial pitch angel and the rotor speed at the primary input file. After I have done the linearization 36 times and I took the average, I found that the output rotor speed from the linear model is higher than the rated speed which I have put as initial speed. I don't know why ?

Sorry for my more questions.
Regards,
Taha Fouda

Jason.Jonkman
Posts: 3753
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: FAST8 Linearization

Postby Jason.Jonkman » Tue Mar 27, 2018 6:43 am

Dear Taha,

I'm not sure I really understand your question; are you referring to the operating point rotor speed or the rotor speed resulting from some time-domain integration of the linearized system? If the former, I suggest that you always review the time series before using the linearized solution to ensure that the time series is in a steady state or periodic steady state condition during the linearization process.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Taha.Fouda
Posts: 44
Joined: Thu Jun 22, 2017 1:51 am
Organization: DLR Institiue of Flight Systems
Location: Germany

Re: FAST8 Linearization

Postby Taha.Fouda » Tue Mar 27, 2018 7:39 am

Dear Jason,

I am doing the linearization of my turbine model at region 3 ( wind speed above rated speed ). I have defined the trim conditions at the primary input files.
the problem now is:-
- when I compare the linear response with the nonlinear response for the rotational speed . it shows different response at the first 10 sec before the change the pitch.
- why the linear response starts from 28.1 rpm and I have defined the trim rpm at 19 rpm.
- the blue line the linear and the yellow line is the nonliear.
- what do you mean by the time series?

Regards
Taha
Attachments
Untitled.png
Untitled.png (70.45 KiB) Viewed 1981 times

Jason.Jonkman
Posts: 3753
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: FAST8 Linearization

Postby Jason.Jonkman » Tue Mar 27, 2018 7:58 am

Dear Taha,

I'm sorry, but I don't understand enough about what you are doing to answer answer your question.

When I referred to "time series" in my prior post, I was referring to the time-domain simulation that FAST v8 is solving before linearizing at various LinTimes.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Taha.Fouda
Posts: 44
Joined: Thu Jun 22, 2017 1:51 am
Organization: DLR Institiue of Flight Systems
Location: Germany

Re: FAST8 Linearization

Postby Taha.Fouda » Thu Mar 29, 2018 3:01 am

Dear Jason,

Maybe I have found from where my Problem Comes. I have my own Generator model and control and I Need to do the linearization process. to do that I have to set GenModel 1 or 2. I can't set it to 3 to call my Generator Routine.

I have run the time-Domain Simulation with my DLL Controller and my Generator subroutine. it gives me the result which I expect.

So what should I do to linearize for my Generator model?

Best regards,
Taha

Jason.Jonkman
Posts: 3753
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: FAST8 Linearization

Postby Jason.Jonkman » Thu Mar 29, 2018 8:44 am

Dear Taha,

It is difficult for me to answer your question because I have no knowledge of your your generator model or controller. Is this model in a form suitable for linearization? Are there internal states? Are there more inputs than generator speed and more outputs than generator torque?

The linearization process in FAST v8 involves (1) linearizing each module to find the Jacobians of the state equations and output equations with respect to the states and inputs and (2) linearizing the FAST glue code to couple all of the linearized matrices together. See our TORQUE 2016 paper on the FAST v8 linearization process for more information: https://www.nrel.gov/docs/fy17osti/67015.pdf.

So in the end, you'd have to modify the FAST source code so that the linearized form of your model can be calculated and the FAST glue code can make use of new Jacobians. Depending on the complexity of your model, this may take a bit of work.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Taha.Fouda
Posts: 44
Joined: Thu Jun 22, 2017 1:51 am
Organization: DLR Institiue of Flight Systems
Location: Germany

Re: FAST8 Linearization

Postby Taha.Fouda » Wed Apr 04, 2018 3:15 am

Dear Jason,

I unterstand what do you mean. my Generator is a permanent Magnet.

I have changed the Generator subroutine in UserSubs.f90 and UserVSCont_KP.f90. when I set GenModel to 3 in the ServoDync Input file and I run the time series Simulation, it gives me the expected results.

In case of Linearization I cann't set GenModel to 3 only 1 or 2. so what do you mean is that I have to Change in the FAST source code in order to do the linearization?

Best Regards
Taha

Jason.Jonkman
Posts: 3753
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: FAST8 Linearization

Postby Jason.Jonkman » Wed Apr 04, 2018 4:00 am

Dear Taha,

I mean that if you want to linearize FAST with ServoDyn with GenModel different from 1 or 2 will require that you modify the source code and recompile FAST.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Taha.Fouda
Posts: 44
Joined: Thu Jun 22, 2017 1:51 am
Organization: DLR Institiue of Flight Systems
Location: Germany

Re: FAST8 Linearization

Postby Taha.Fouda » Fri Apr 06, 2018 2:17 am

Dear Jason,

Thank you for your Response, I have a question regarding running my Controller with bladded and with FAST. I have run my DLL PID Controller one time with FAST and one time with bladded. I have found that the rotational Speed with fast are more better ( smooth Response after the start up time ) but with bladded it shows small continious oscillations after the start up period.

Could you please give a reason behind this, if you have?

Best Regards

Jason.Jonkman
Posts: 3753
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: FAST8 Linearization

Postby Jason.Jonkman » Fri Apr 06, 2018 3:54 am

Dear Taha,

Sorry, but I can't comment on that.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Charlie.Sheng
Posts: 17
Joined: Wed Jul 25, 2018 3:00 pm
Organization: western university
Location: canada

Re: FAST8 Linearization

Postby Charlie.Sheng » Fri Oct 12, 2018 2:22 pm

Dear Dr. Jonkman,

My work is to perform the dynamic response analysis using FAST, but before that, I want to check the natural frequency of the built wind turbine in FAST. So I wanna use the Linearization module to do this analysis. I run the Test14 which is used for linearization analysis for WP 1.5MW using version of FAST v8. While I have read the user's manual and Readme for v8, also the output file after the linearization and post-processing MATLAB scripts, and some posts about linearization (not all of them as there are too much). I still have some questions for inquiring you:

1. After running Test14.m and MBC3.m MATLAB script, I get the results of DampedFrequencies, DampedFreqs_Hz, NaturalFrequencies, NaturalFreqs_Hz,DecrementRate,DampRatios,MagnitudeModes,PhaseModes_deg etc. But I don not know what these values mean. For example, I get the 16 frequency values in the MBC_DampedFrequencyHz, but it is not clear about which frequency is for blade or for tower. Can you explain a little for me? Or is there document about explaining its calculated results?

2. I find that in Test14.fst except Elast module is on, the other moulde such as inflow, Aero and Servo modules are off. And the initial conditions of wind turbine in ElastoDyn is still. So, my concern is if I would like to get the natural frequency of whole system in still condition, I can also only contain ElastoDyn moudle for the linearization analysis, right?

Looking forward to your reply.

Best regards,
Charlie

Jason.Jonkman
Posts: 3753
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: FAST8 Linearization

Postby Jason.Jonkman » Fri Oct 12, 2018 2:37 pm

Dear Charlie,

The CampbellDiagram.xls workbook sorts the eigenvalues and eigenvectors by frequency and highlights the dominant components of each mode to aid in interpretation of the modes. This has been discussed several times on this forum e.g. see: viewtopic.php?f=4&t=475 and viewtopic.php?f=4&t=833&p=3352.

I'm not sure I understand your second question, but you only need to enable the ElastoDyn module if all you want to do is calculate the wind turbine natural frequencies in standstill without aerodynamics.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Charlie.Sheng
Posts: 17
Joined: Wed Jul 25, 2018 3:00 pm
Organization: western university
Location: canada

Re: FAST8 Linearization

Postby Charlie.Sheng » Mon Oct 22, 2018 5:28 pm

Dear Dr. Jonkman,

Very appreciate your helpful info.
Now I can get the natural frequencies of the blade, tower etc. based on the posts you introduce me. Then I come with another question. For instance, now I can get the natural frequencies of my tower from the pre-processor "Mode" with the value of f1tower=0.3483(Hz) and f2tower=3.245 (Hz). Right now I can also obtain these values from the MBC post-processor with the value of f1tower=0.344(Hz) and f2tower=3.056 (Hz). Among these, the values from two results for f1st are pretty close, but the f2nd show some difference. So, my question is the result from which tool(Mode or MBC) is right?

Best regards,
Charlie

Jason.Jonkman
Posts: 3753
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: FAST8 Linearization

Postby Jason.Jonkman » Tue Oct 23, 2018 7:41 am

Dear Charlie,

The natural frequencies calculated from the FAST linearization analysis + MBC3 should be consistent with the time-domain solution of FAST. The Modes preprocessor does a reasonable job at calculating the mode shapes and natural frequencies for the blades, but tends to be inaccurate for the tower because of the importance of the rotational inertia of the rotor-nacelle-assembly on the tower mode shapes, which Modes does not account for. This is why we recommend that BModes be used in place of Modes for calculating tower mode shapes.

Best regards,
Jason Jonkman, Ph.D.
Senior Engineer | National Wind Technology Center (NWTC)

National Renewable Energy Laboratory (NREL)
15013 Denver West Parkway | Golden, CO 80401
+1 (303) 384 – 7026 | Fax: +1 (303) 384 – 6901
nwtc.nrel.gov

Charlie.Sheng
Posts: 17
Joined: Wed Jul 25, 2018 3:00 pm
Organization: western university
Location: canada

Re: FAST8 Linearization

Postby Charlie.Sheng » Tue Oct 23, 2018 12:58 pm

Dear Dr. Jonkman,

Thanks your prompt reply and answer which are quite helpful for me. Following your reply I further have one concern: How can I get the mode shapes of blade and tower, to be exactly, if I wanna plot the mode shapes for blade and tower, how should I do?

Best regards,
Charlie


Return to “Controls”

Who is online

Users browsing this forum: No registered users and 1 guest