I have a few questions regarding sampling time, dead-time, and output disturbance rejection.

Given a typical first-order plus dead-time model:

$$P = {Kexp^{-sL} \over T_ps+1}.$$

I would like to calculate an acceptable dead-time, L, and appropriate sampling time, T, to reject (attenuate) output disturbances $\le \omega$.

In my past projects, I have followed the rule of thumb outlined here, that the sampling time should be 10 times per process time constant or faster, and taken it for granted

$$T \le 0.1Tp$$

However, I don't know how (or if) faster sampling can improve disturbance rejection. Modeling this in Simulink shows a diminishing return.

Furthermore, I am convinced that if T is fast enough, disturbance rejection performance is limited to dead-time. In the past, I have modeled closed-loop systems in Simulink and iteratively increased L until disturbance rejection was unacceptable. Is there a mathematical relationship between L and $\omega$ that clearly shows this?

Lastly, how do you apply the same exercise to second-order systems where Tp is not explicitly available?

Thank you,

  • $\begingroup$ Is $P$ the disturbance attenuation transfer function? $\endgroup$ – fibonatic Apr 14 '18 at 16:01

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.