Dynamic Mode Decomposition (Code)

In this video, we code up the dynamic mode decomposition (DMD) in Matlab and use it to analyze the fluid flow past a circular cylinder at low Reynolds number.

Code and data available at:

This video was produced at the University of Washington

[lukasf]

Looking forward to test it myself. Keep up the good work. Thanks!

[abhinavpandey]

This is a very good video, and I love your examples

[dweeder]

Thanks for these helpful videos. You explain DMD in a very simple way. I wonder what is the catch. haha

[amirpasagic]

Mr. Bruton, what is exactly the difference between POD and DMD speaking very generally, not in mathematical particularities.
Also, I think it would be super useful to see the whole process from modeling to control design e.g. taking actual data (perhaps even generating to include some basic CFD), doing some model reduction like DMD and using it on analysis and control design. I think it would be a great way to learn how to use it in practice (and when) and clearly see the benefits of applying such methods. In anycase, thank you for all the super useful material, its a true privilege to have such great lectures available.

[collinjoseph]

Thanks for the video. I would like to know the x and y coordinates of the 89350 pts whose velocity was inputted

[ahmedcelik]

Dear Brunton thanks for the video. The forward backward code in LDS_DMD_eig.m provided in chapter 8 of your DMD book does not generate the eigenvalues presented in figure 8.9. It seems to have something wrong in the code! Can you please check?

[ahmedgharieb]

I want ask how is the structure of time delay embedding of 1 dimension inside the DMD moldel does we use the H matrix or X time series multiple of U S V.?

[anshulsuri]

Where can one find the dmd function mentioned at the end of the video?

[FT-niyz]

Hi Sir, I have some picture captured by camera and I used the light intensity as a data of column vector of x matrix....i became confused alittle....what is the code of showing the modes in matlab.

[dunston]

Nice. But maybe I would have liked a precision about how to interprete Real and Imaginary parts that you are plotting separately... Someone knows ?

[hiddenhidden]

Does changing my last name to Brunton help me study this?

[zhangyuwen]

Hi Steve, just wondering why the eigenvalues of all the modes lie within the unit circle? I tried to apply the method to simulation results obtained from detonation wave, and i got all my dominant modes lie within the unit circle as well. Secondly, all my modes lie on the right-hand side of the imagenery plane, is it an indication that my system is inherently unstable ? Thanks.