Integrator windup

We explain how saturation and integral action can combine to produce the unwanted integrator windup phenomenon.

[@utilizator1701]

Excellent tutorial.

[@soltys7011]

great video!

[@jesseberg5728]

very good explanation, thankyou sir

[@thuannv1988]

Thanks a lot for your tutorial. It is really informative.

[@richard_pates]

Glad it was helpful!

[@user-tx8ek1hp5r]

Windup didply flashing between hi windup and low windup why

[@rommelreyes2209]

would lowering the Ki solve the windup?

[@richard_pates]

Nice question! The answer is yes, but... If you push this to the extreme, a very effective way to remove integrator windup is to remove the integrator (make Ki=0)! But then you lose the benefits of integral action. This is the real trade-off that needs to be addressed when using anti-windup compensation. On the one hand you want Ki small to avoid windup (using less Ki can also improve other things like stability margins). On the other, integral action is very important to track set point changes and compensate for constant disturbances - and for these effects you want Ki big. Windup compensation allows you to mitigate some of the negative effects of windup, whilst still having enough Ki to get the other benefits

[@rommelreyes2209]

@@richard_pates any books you can recommend that discuss this? I have intro books on control systems but they don't to this level. Thanks for the video BTW.

[@richard_pates]

@@rommelreyes2209 You can find a bit more in the book "Feedback Systems: An Introduction for Scientists and Engineers" by Åstrom and Murray in the chapter on PID control. This is a great introductory book on control in general, and it's easy to find pdfs online. It may not be enough if you need something very sophisticated, but hopefully they also give references where you can find more. In terms of authors, Åstrom and Hägglund are world experts in PID control, so you can also search for their work and publications too. Good luck!