I am simulating the dynamics of a control system that has been designed using the control partitioning technique (model-based control + servo control).

I am having a query for the case when an external disturbance is present in the control system.

So there are a couple of configurations that are possible:

Number 1: enter image description here

Number 2: enter image description here

The only thing which is different between the two architectures is the summation block where the disturbance force is introduced in the control system.

The thing which is confusing me is that all the literature and books I have referred to use the 1st architecture but formulate the closed-loop dynamics according to the second one.

In the presence of disturbances both the systems show tangibly different responses for the same controller properties, a few examples given below:

Impulse force applied on the system at 5 seconds enter image description here

Continuous Static Disturbance on the system: enter image description here

Along with the different responses, the actuator efforts will also be different.

This is making me confused as to which architecture will be a truer representation of the actual response of the designed control system in presence of disturbance forces.

I hope my question makes some sense and will provide further elaboration if asked for!

Thanks in advance.

  • $\begingroup$ They should simply differ by factor of α , see if that accounts for what you see graphs. You can introduce disturbances at any node, and they have different interpretation. Disturbance per "number 1" is a plant input, disturbance per "number 2" is noise on the controller output signal. You can similarly have noise on the controller input signal (ie sensor noise), as well as sensor dynamics (pole or delay), which is worth including. Also I don't see input anywhere. $\endgroup$
    – Pete W
    May 15 at 14:15
  • $\begingroup$ for the system where the literature refers to a closed loop gain (e.g. Y/U), which node receives the input (e.g. U)? $\endgroup$
    – Pete W
    May 15 at 14:24
  • $\begingroup$ @PeteW yes they are different by the factor of $\alpha$. For example, say we have an inverted cart pendulum system, where we want to regulate the pendulum. In order to introduce a disturbance, we flick the cart. Will that disturbance categorized as a controller output noise, and if not, what does controller output noise encompass? Lastly, there is no input to the closed-loop system because I have designed this for position regulation. $\endgroup$ May 15 at 14:31
  • $\begingroup$ @PeteW According to the literature, the leftmost node receives any closed-loop input (like desired states). $\endgroup$ May 15 at 14:36
  • $\begingroup$ I'd say it could be either a disturbance in force, if plant input is a force, or disturbance to the system state (position and velocity of pendulum and possibly the cart too, if its control mechanism can back-drive). You'd have to show more detail in system diagram. Regarding the command input for the whole loop, you could still draw one (same dimensions and units as system output), and just set it to zero. I believe most control papers will show it even if it is 0 $\endgroup$
    – Pete W
    May 15 at 14:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.