3 Ways to Build a Model for Control System Design | Understanding PID Control, Part 5

Tuning a PID controller requires that you have a representation of the system you’re trying to control.

This could be the physical hardware or a mathematical representation of that hardware.
If you have physical hardware, you could guess at some PID gains, run a test to see how it performs, and then tweak the gains as necessary. This guess-and-check brute force tuning method might work, but you have other, more precise, options available if you have a mathematical model of the system. Therefore, this video presents three different ways to model your system so that you can take advantage of each of these methods when tuning your controller.

The first method uses a detailed understanding of the system to develop the model with first principles. The second method uses system identification, the known input, and resulting output signal to fit the data to the model structure of your choice. The third method creates a model by linearizing an existing nonlinear model around a given operating point.

With any of these models, you can start the process of developing and tuning a PID controller.

Other References:
--------------------------------------------------------------------------------------------------------------------------------------------------

© 2018 The MathWorks, Inc. MATLAB and Simulink are registered
trademarks of The MathWorks, Inc.

[BrianBDouglas]

Thanks for watching! As always, if you have any questions on this video leave them here and I'll try my best to answer them. Cheers!

[brajlecz]

This guy Brian is great, he really knows his stuff.

[asadsalim]

Wow. That's awesome man. It's getting really exciting and I'm looking forward to watching the next video.
Thank you very much and keep up the good work.

[fernando.liozzi.]

How much information have you given! I loved the video. Thank you very much.

[pnachtwey]

pretty good. In over 35 years of doing motion control I have NEVER come across a system with a transfer function. The first 15 years I tuned system by the seat of my pants. Then I got Mathcad and that opened up my eyes. The text books and most instructors do not teach controls correctly. First one must find a model. Brian doesn't say how but there are algorithms like Levenberg-Marquardt and BFGS that can find the coefficients for the differential equations for models. If you use differential equations, you can model non-linear systems fairly accurately.


There are algorithms like Ackermann's method that can be used for pole placement. The result is not always a PID. Sometimes it is a PI, PD, PID, PD with second derivative gain or even PID with second derivative gain. PID is not always the correct choice.



I use pole pole placement for tuning on I have done the system identification. It is easy to simply place the closed loop poles on the negative real axis in the s domain or on the positive z axis in the z domain.

[wiloberlies]

Yet another really well done video Brian, on a really important topic. Nicely illustrating the capabilities of MATLAB/Simulink as well. Thank you.

[blackboardradio]

Thank you Brian for these great videos. Best wishes and regards! Stay blessed

[muhannaddrakalsebai]

You can't imagine how useful it is. But please mention the analyzer should be ODE3 with fixed step 0.001. Because I was using ODE45 and the results was very disappointed. But after changing the solver I got a perfect results. Thank you Brian very much.

[Antonioqwert]

Hi, Brian your videos about control theory is great, I hope you will talk about MPC controller

[nabilandadamslaboratory]

Can you please make a full detailed video on how to control a motor's speed using PID? Thanks!

[hafiezamaniahmad]

good explanation. nice job. thanks you

[abhijeet]

Hey Brian, The lecture series is really helping me proceed in my project! Even the reference links are awesome too.
Here you are using a sensor block to get the feedback from the motor, how can I model the Sensor Block that you have used here?
Thanks!

[venkateshnayak]

Well, that was an eye opener moment! Now I know how to model a system. Or at least on how to begin in the first place. That's where most of the budding control engineers get stuck. Great work man, loved it. However I had a concern regarding, modelling a system using physical system, although you mentioned it you still carried out the process using your non linear model. Let's assume I have the physical hardware, how do I go about it, Give different signal input and obtain the output! But isn't that risky since I know very less of the behaivour of the physical hardware!
Also I would love to dwell in to the intricasies on how matlab system identification toolbox finds a TF using input and output curve. Any resources for that ? Or Maybe a video just on that topic.
If that is sorted anybody could tackle any control problem!

[calvinceogara]

very detailed explanation video tutorials

[ChristianMego]

Excelent video Brian, Hope you can share all code including simulink, thanks.

[maxipixie]

Hi Brian thanks very much for this series .
I'm working on the example that you explained in simulink .
I need to know what the block (sensor) contain.

[IchrisiI]

Can I see the first principles model of your sensor ? Thank you for your commitment!

[shrinivasiyengar]

How is the input to the PID controller a speed variable, but the output is a voltage? Where and how does this transformation take place? You enter the error in the speed of the motor into the PID controller that it tries to minimise. How does that convert to voltage?

[luismeron]

@9:04 the yellow waves shows noise compared to the blue line(our model)

[BilalAhmad-lspd]

Hello Brian!
You must have to take a real motor and then using identify the parameters of the model. Because you are living in a developed country you can buy or someone sponser you. I live in a developing country and can't buy all the hardware required.