Linearization issue using Fast V7.02

Provide feedback, request enhancements, and get help with wind-turbine computer-aided engineering tools.

Moderators: Bonnie.Jonkman, Jason.Jonkman

Jason.Jonkman
Posts: 5057
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Linearization issue using Fast V7.02

Postby Jason.Jonkman » Wed Mar 18, 2020 10:18 am

Dear Zhe.Zhang,

I'm not seeing a very good match between the linear and nonlinear results in your plots. But it is difficult for me guess why the results not agreeing. I would suggest simplifying the model down to one DOF to see if you can get a better agreement. Once you get good agreement with one DOF, add addition DOFs/complexities in steps.

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

Zhe.Zhang
Posts: 26
Joined: Wed Apr 03, 2019 1:40 am
Organization: Harbin Engineering University
Location: China

Re: Linearization issue using Fast V7.02

Postby Zhe.Zhang » Wed Mar 18, 2020 8:38 pm

Dear Dr. Jonkman,
As you said, I only enabled one degree of freedom, but the differences are still large. Are these differences due to different versions? I only use fast7 for linear calculations, but I use fast8 for non-linear models. If it is not a version problem, why is there such a difference? I noticed that the pitch angle of the linear model changes slowly at the time of wind speed change. Is this related to the pitch rate setting in fast7? This value can be set in PitManRat in fast8, but how can I change this in Fast v7? Please help me find out why the model results are different.
Thanks
Best regards,
Zhe.Zhang
Attachments
fast7.zip
(12.85 KiB) Downloaded 26 times
fast8.zip
(82.63 KiB) Downloaded 29 times

Jason.Jonkman
Posts: 5057
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Linearization issue using Fast V7.02

Postby Jason.Jonkman » Thu Mar 19, 2020 7:47 am

Dear Zhe.Zhang,

I would not expect large differences only because you switched from FAST v7 to FAST v8, assuming you've set equivalent inputs between the two versions (I did not check all these, and you didn't send me your AeroDyn inputs to check anyway).

Looking at your results, it looks like the blade-pitch response is very different between the two solutions. Are you using the same controller in both the linear and nonlinear solutions? Are using the units that are consistent with the linear and nonlinear models in both pitch controller? You could see if the problem is in the pitch controller by disabling the pitch controller altogether and seeing if you get a similar change in rotor speed with the step wind you are applying.

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

Zhe.Zhang
Posts: 26
Joined: Wed Apr 03, 2019 1:40 am
Organization: Harbin Engineering University
Location: China

Re: Linearization issue using Fast V7.02

Postby Zhe.Zhang » Thu Mar 19, 2020 7:50 am

Dear Jason
I noticed that the pitch angle of the linear model changes slowly at the time of wind speed change. Is this related to the pitch rate setting in fast7? This value can be set in PitManRat in fast8, but how can I change this in Fast v7?

Jason.Jonkman
Posts: 5057
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Linearization issue using Fast V7.02

Postby Jason.Jonkman » Thu Mar 19, 2020 7:56 am

Dear Zhe.Zhang,

No. ServoDyn input PitManRat in FAST v8 is only used in the override pitch maneuvers, which you have not enabled. The override pitch maneuver also exists in FAST v7, but the input is slightly different. In FAST v8, the override pitch maneuver is defined by inputs TPitManS, PitManRat, and BlPitchF whereas in FAST v7 the override pitch maneuver is defined by inputs TPitManS, TPitManE, and BlPitchF (i.e., TPitManE is used in place of PitManRat). But regardless, you have not enabled the override pitch maneuver, so, this will not effect your result.

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

Zhe.Zhang
Posts: 26
Joined: Wed Apr 03, 2019 1:40 am
Organization: Harbin Engineering University
Location: China

Re: Linearization issue using Fast V7.02

Postby Zhe.Zhang » Fri Mar 20, 2020 1:44 am

Dear Jason
Thank you very much for your patience for answering, it is really because of the unit of the controller.
sincerely yours,
Zhe.Zhang

Zhe.Zhang
Posts: 26
Joined: Wed Apr 03, 2019 1:40 am
Organization: Harbin Engineering University
Location: China

Re: Linearization issue using Fast V7.02

Postby Zhe.Zhang » Fri Mar 20, 2020 10:58 pm

Dear Jason,
I modified the parameters of my controller and got the following comparison results. What I don't understand is why the non-linear model fluctuates sharply at the initial moment. Is this a normal phenomenon? What do you think of this comparison result?
sincerely yours,
Zhe.Zhang
Attachments
result.zip
(100.4 KiB) Downloaded 29 times

Jason.Jonkman
Posts: 5057
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Linearization issue using Fast V7.02

Postby Jason.Jonkman » Sat Mar 21, 2020 8:00 am

Dear Zhe.Zhang,

My guess is your controller is highly sensitive to the initial conditions that are set. I would suggest either running a longer simulation and ignoring the results during the start-up transient or setting better initial conditions. Normally I would recommend setting the rotor speed and blade-pitch angle to their expected mean values for the given condition (mean wind speed), but you could always set other initial conditions (e.g., blade, tower, and platform deflections) to their expected mean values as well.

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

Babak.Gavgani
Posts: 29
Joined: Sat Mar 14, 2020 12:03 pm
Organization: ---
Location: Belgium

Re: Linearization issue using Fast V7.02

Postby Babak.Gavgani » Fri Apr 17, 2020 12:05 pm

Hi,

This is my first post and this is a honour to be a part of this forum.
I am working on the linearizing of wind turbine at different Maximum Power Points (MPPs) for different wind speed values. To do this, I have obtained the rotor speed at MPP for different MPPs according to the Power_Rotor Speed curve, Fig.1. Rotor speed for wind speed = 9m/s at MPP is 160.38 rpm (10kW wind turbine is selected). I have used FASTv7 for Linearization.
fig2.JPG
Power-Rotor speed
fig2.JPG (27.79 KiB) Viewed 760 times

I have set the rotor speed to 160.38 (rotor speed at MPP) and the linearization is done with only 1 DOF and the generator torque is selected as the control input. The linearizatoin output is shown in Fig2
fig3.JPG
Linearization Output
fig3.JPG (143.11 KiB) Viewed 760 times

I would be appreciated to let me know if I am on the right direction.

Thanks in advance.
Babak[/b][/b]

Jason.Jonkman
Posts: 5057
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Linearization issue using Fast V7.02

Postby Jason.Jonkman » Fri Apr 17, 2020 2:04 pm

Hi Babak,

I'm not really sure what you want commented on. And I'm not following everything in your post because of some inconsistencies, i.e., your power-speed curve shows the MPP at a rotor speed of 16 rpm, the text in your post refers to 160 rpm, and the linearization output file refers to 7.5288 rad/s = 71.89 rpm. Can you clarify your question?

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

Babak.Gavgani
Posts: 29
Joined: Sat Mar 14, 2020 12:03 pm
Organization: ---
Location: Belgium

Re: Linearization issue using Fast V7.02

Postby Babak.Gavgani » Fri Apr 17, 2020 2:35 pm

Dear Jason,

Thanks for your reply and sorry for the inconsistencies.
I want to obtain multiple linear models of a wind turbine at different MPPs for different wind speeds and then design controllers for each of these models.
Here, I have attached the Power-Rotor Speed curve for wind speed = 9m/s, Fig.1. The rotation speed at MPP is 160.38 rpm.
Powercurve.JPG
Fig1
Powercurve.JPG (28.63 KiB) Viewed 753 times

Then, I have linearized the turbine around this point, by setting RotSpeed to 160.38 (rotor speed at the MPP). I have attached the linearzation output, Fig2. Linearization is done with 1 DOF and the control input is generator torque.
Linearization.JPG
Fig2
Linearization.JPG (160.22 KiB) Viewed 753 times

The question is if I am doing it in a right way? Is linearization at MPP?
Thanks for your time and considerations.

Best regards,
Babak

Jason.Jonkman
Posts: 5057
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Linearization issue using Fast V7.02

Postby Jason.Jonkman » Fri Apr 17, 2020 2:52 pm

Dear Babak,

OK, thanks for fixing the inconsistencies. The rotor speed in the linearization file, text, and plot now all match!

I don't really see anything problematic with what you are doing, but I did notice that the power at the linearization operating point does not match your plot, i.e., ( rotor speed )*( generator torque ) = ( 16.7948 rad/s ) *( 221.9 N*m ) = 3.73 kW, not 2.2 kW, from your plot. That said I'm not sure what your gearbox ratio or generator efficiency is (I assumed both are 1.0 here).

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

Babak.Gavgani
Posts: 29
Joined: Sat Mar 14, 2020 12:03 pm
Organization: ---
Location: Belgium

Re: Linearization issue using Fast V7.02

Postby Babak.Gavgani » Fri Apr 17, 2020 3:12 pm

Dear Jason,

Thanks for your reply.
I checked the primary input file and I changed the generator efficiency from 83% to 100% but the torque at the linearization operating point did not change. Please advice with this matter. what could be the problem that power values do not match? My linearized model is wrong or ...?

Best regards,
Babak

Jason.Jonkman
Posts: 5057
Joined: Thu Nov 03, 2005 4:38 pm
Location: Boulder, CO
Contact:

Re: Linearization issue using Fast V7.02

Postby Jason.Jonkman » Fri Apr 17, 2020 3:23 pm

Dear Babak,

The generator torque and speed are not directly influenced by the generator efficiency.

mechanical power = torque*speed
electrical power = torque*speed*efficiency

Unless the electrical power is fed back and used in the controller, it has no impact on the FAST solution.

Was your power versus speed curve derived from FAST and is the power mechanical or electrical?

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

Babak.Gavgani
Posts: 29
Joined: Sat Mar 14, 2020 12:03 pm
Organization: ---
Location: Belgium

Re: Linearization issue using Fast V7.02

Postby Babak.Gavgani » Fri Apr 17, 2020 3:58 pm

Dear Jason,

The power curve has been derived from FASTv8. I have used electrical power. How much difference is reasonable between the obtained power from curve and the power from linearization output? Can it be because of using different versions of FAST?

Regards,
Babak


Return to “Computer-Aided Engineering Software Tools”

Who is online

Users browsing this forum: No registered users and 1 guest