Ziegler & Nichols Tuning Rules 📉 PID Controller Design Examples! 💡

[@alimumcuoglu1625]

Truly the most understandable video in the world. You aaare very practical.

[@CANEDUX]

Thanks for your message :) Great to know that you liked the video 👍

[@Shurvb235]

God bless your soul man your videos have been seriously clutch

[@CANEDUX]

Thanks for your message. Great to know that you liked the video 👍

[@SMV1972]

Отличный урок. Отличные новые знания. Спасибо за Ваш труд!

[@CANEDUX]

Glad you liked it. You are welcome! Share the knowledge. Рад что вам понравилось. Пожалуйста! Поделитесь знаниями :)

[@arthursoares5832]

Great video. Very helpful.

[@CANEDUX]

Thanks for your message. Great to know that you liked the video 👍

[@dronzerdanks7163]

Great job. This was helpful.

[@CANEDUX]

You are welcome!

[@ArnabJoardar]

Hello, Thanks for the video. It was really helpful to know how to apply this. At 17:06, the reason why a root of omega is common between both the complex and real parts is because it is the pole of this closed-loop system, right? It is not just a coincidence for this problem in particular but it is an artifact of making it into a marginally-stable system.

[@CANEDUX]

Thanks for your comment! Glad you liked it! The fact that the critical frequency is sqrt(5) rad/s both from the equation of the imaginary and real part is not a coincidence. They should always give the same result. I wanted to emphasize this point in the discussion also, so you can use both forms to determine the critical frequency. In the end, we will go for the equation which will give us the results faster and easier. In this case, the real part setting to zero and solving is faster. I hope this clarifies the situation. Feel free to get back if you have further questions. Do not forget to like and share the knowledge! Thanks!

[@Ch40T1cX]

Wonderful video! However, could you explain how you rewrote the equation at 19:24 ? Ive skimmed through some of the previous videos but could not find an example

[@CANEDUX]

Thanks for your message! You can factorize a second-order expression into a squared expression of first-order expression. This is called completing the square. You can check this link for more info: www.khanacademy.org/math/algebra/x2f8bb11595b61c86:quadratic-functions-equations/x2f8bb11595b61c86:more-on-completing-square/a/completing-the-square-review

[@nothilexulu8667]

GREAT VIDEO AND HELPFULL. I WANT TO ASK HOW TO FIND Pct IN UNDAMPED FREQUENCY OF OSCILLATION

[@CANEDUX]

Thanks for your message! What do you mean by Pct?

[@nothilexulu8667]

@@CANEDUX Critical period

[@CANEDUX]

@@nothilexulu8667 You can determine this from the transient response. For example, read the time between two peaks in a pure oscillation.

[@AliAhmed-ye3rj]

well done.

[@CANEDUX]

Thanks!

[@ahmedzahran7115]

Hi, Great video..is there a way to draw the graph of example 2 in 20:30 without using simulation ? Also how to find zeta from the pid equation? Thank you

[@CANEDUX]

Thanks for your message. It is possible to draw this graph without a simulation program. For this, you need to determine the inverse Laplace transform of the closed-loop transfer function. This could very tedious and time consuming, though. Why do you want to do this? For your second question, what you mean by zeta from the PID controller?

[@henrihollebeke2496]

Hi Can Bijles Great Video. But I there is something I don't understand with confidence. When do you use the first or second method? What I think: The first method will be used when the delay time and lag time are experimentally determined. And the Second method will be used when the mathematical model is knwon. Is this the right way of thinking?

[@CANEDUX]

Hi Henri, thanks for your message! Your way of thinking is correct. The examples I worked out in this video also use the fact that the model of the plant is not known in the first method to determine the controller. In the second method, the plant model is indeed known and based on this, the controller is designed. Note that there are no specifications given in the examples in this video, like overshoot, settling time, and steady-state value, like we have done in the root locus design. See playlist with many examples: kzbin.info/aero/PLuUNUe8EVqlnY2zKWnx-6nyc6CqyPApDD A specific tuning method does not always have a strict guide and you may need to tune the design in the first round, as we have done in this video. The actual goal is to design to proper controller to meet the specifications, which might be not available always in numbers. However, in some cases, it is also required to know the mathematical model to a sufficient degree of accuracy. The final goal is that the closed-loop system is stable and has a reasonable transient and frequency response.

[@bankasravankumar6691]

Good morning sir, can you send transfer function of first method like 2d method transfer function.?

[@CANEDUX]

in the first method, you do not need the transfer function of the plant. From the step response, you will determine the delay time (L) and the time constant (T), and then you can use the parameters from the Ziegler & Nichols tuning method table.

[@joyernallifes]

hi, thank you so much for the video but I wanna ask bout the tutorial how to input that data to the simulink. I mean like step by step how to get that result the kurva you know. again thank you so mych and I really apreciate that if you want to do that

[@CANEDUX]

Thanks for your message. Glad to hear you liked the video. You can look at these links for more information about Simulink. www.mathworks.com/support/learn-with-matlab-tutorials.html kzbin.info/www/bejne/n4DQopqbrM9ojqs

[@h3h3podcastclipper]

Hello. I notice that the time constant is determined by tangent line at inflection point. However, other sources state that the time constant is the time needed for the output to reach 63% of the steady state value. Is there an explanation on the discrepancy?

[@CANEDUX]

Hi, thanks for your message. You can actually use both methods, depending on which one it is most accurate for a specific problem. Assuming a first-order system, you can determine its time constant using a slope at the origin and the final value of your system or you take the ~63.2% of the final value, but the exact value of 63.2% is 1-1/exp(1). I hope this is helpful.

[@riccardo4303]

Hi, thanks for the explanation. I have this request: "Perform the identification experiment described by the open-loop Ziegler and Nichols technique to determine the two FOPDT models describing the decoupled plant". The system is a MIMO and the only S shape that I get is with an open-loop configuration. In the closed-loop one or adding the decoupler the output has a different shape. So, is right and I have to apply the method on the open-loop without decoupling? Otherwise I can't understand the meaning of "decoupled plant" in the task. Thanks

[@CANEDUX]

Hi, thanks for your message. Do you have the system configuration and the actual problem description?

[@riccardo4303]

@@CANEDUX Yes, I send you the file by email if it’s not a problem. (I’ve found the email on your website)

[@CANEDUX]

@@riccardo4303 That is ok.

[@riccardo4303]

@@CANEDUX I've sent the email on tuesday. Is it arrived?

[@CANEDUX]

@@riccardo4303 I received your mail. I will try to respond to it coming weekend.

[@finfoil8544]

Why do you determine T at 80% of the final amplitude and not at 100%?

[@CANEDUX]

The tangent line was not drawn accurately here. You should indeed go up to 100% of the final value. The error in this case is not much.

[@osvaldoperez6979]

Hi how did you get your G(s) for the 1 method, I got to where you input the values but didn’t understand the final form .

[@osvaldoperez6979]

I would great appreciate the help a lot sir.

[@CANEDUX]

The First Method sets the plant in open-loop configuration. Applying an appropriate step input singal in the open-loop configuration, you can get the response. From this response (step response), you can determine the transfer function of the plant. Remember that not all systems can be set in open-loop configuration. Does this answer your question? Let me know if you have further questions.

[@osvaldoperez6979]

Oh okay thank you !

[@CANEDUX]

@@osvaldoperez6979 You're welcome!

[@ipanjohur3691]

22:46 if we make the zeros near to the 0 we decrease the overshoot. is that right?

[@CANEDUX]

It depends on the total effect of all the poles and zeros. It can be that the zero coming closer to the origin will be almost canceled by a pole at the origin. You can see this effect nicely using a simulation. Try it out!

[@ZawHtoo-p5u]

How should i know Kp=30 ? sir

[@CANEDUX]

I explained this in the video also. Check the steps after time 00:10:20 for the details.

[@chongmeilu9353]

Hello , can I know the link to find routh table ?

[@CANEDUX]

Hello, do you want the links to the videos about how to set up and use the Routh table? More information about stability using Routh-Hurwitz stability method, see the following playlist: kzbin.info/aero/PLuUNUe8EVqllsZiH66E09u6OAlNUdNpwQ

[@chongmeilu9353]

I mean the online calculator, to find the routh table

[@CANEDUX]

@@chongmeilu9353 You may use the following link for this: www.muchen.ca/RHCalc

[@abdullahmohamed8882]

20:30 how did u get the unit response

[@CANEDUX]

You can use the command 'step' in the MATLAB command window. In this case, use step(T).

[@vamshi9997]

Great video. I want to design in LabVIEW could you please help me

[@CANEDUX]

Thanks! I am not really familiar with LabVIEW. What is it that you want to design?

[@mohamedelaminenehar333]

Sir, do u have telegram?)

[@CANEDUX]

I do not have it.