Excellent explanation. Very clear. I was always told windup typically exhibited itself when feedback signals are saturated for prolonged periods, which is no doubt true. However, I've seen the effect described in the video, whereby a real system experiences an unexpected overshoot with a relatively minor change in command. I have suspected integrator windup as a cause, but could never put the theory to it myself or via research (admittedly my attempts were feeble). My applications rely most on experimental (material testing) application of PID concepts, not control design through modeling. This is an excellent explanation tying together the more typical "ideal" systems undergrads are exposed to (my background) , and the possible phenomena seen in application.
Excellent analysis. Perhaps it is better to use the word saturation, in the first phase of the step-input, when maximum voltage is applied to the motor. Only when out of saturation does the system become a mathematical linear control system.
What does it mean that the controller output at steady-state is equal to the controller integral part output? In absence of a disturbance, it would be equal to zero too...